TWiT 1017: Yellow-Bellied Marmots - Return of Pebble, O3-Mini, Apple AI
03:16:25 Share

It's time for Tweet This Week in Tech. Great panel for you. Christina Warren is here. She's got a new job. Maybe we can find out what she's up to. Shoshana Weissman from R Street. She just wrote a piece on age verification and why it's a terrible idea. And Dan Patterson, whose company specializes in

disinforming disinformation. Does that make sense? We will talk about an amazing week in AI, not just deep seek, but new models from open AI, the breakdown of government websites, 8,000 websites and counting disappearing from the internet and

And 23andMe is about to sell your spit. It's all coming up next on TWIT. Podcasts you love. From people you trust. This is TWIT. This is TWIT. This Week in Tech. Episode 1017. Recorded Sunday, February 2nd, 2025. Yellow-Bellied Marmots.

It's time for Twit This Week in Tech, the show we cover the week's tech news, which is mostly AI this week. But that's all right, because we've got a panel that can handle the tough stuff.

Shoshana Weissman is here from our street. Great to see the chairman of the sloth committee in person. That's not a sloth, that's a marmot. Yeah, yeah, yeah. Thank you for having me and my marmot. Oh, are you now the chairman of the marmot committee? No, but I've taken them up as an interest. Look how fun it is. Oh, marmots are wonderful. They move faster than sloths, but not much.

I mean, not when they're this chonky, but they can move fast other times. They're all chonky. You know they have iron butts. Did you know that? They have iron butts. They're a problem in Australia because they roll up when they're scared and they roll up on the highway and you hit them and it's really like hitting iron. Marmots? Yeah. They dive into their holes and they leave their butt. This is not a nature show, but thank you.

I have some uh I have some interest in marmots as well apparently Dan Patterson is also here he's director of content at Blackbird AI a regular writer at ZDNet it's great to see you Dr Dan uh YouTube but keep going about the marmots I mean this is like the best start of a show ever I don't really know if that's I thought we'd be talking AI I didn't expect marmots sarcastic yeah no they're amazing animals yeah

But that's for another time. I do the same thing when I'm scared. Yeah, roll up and stick your head in your hole and stick your butt out. And that way you'll never get hurt, I guess. Don't we all? Yeah.

Christina Warren's here with no job. Hi, Christina, but she is living in the Simpsons house, so that's okay. I am. I'm in Springfield. You know, almost every state has a Springfield. I know. I know. I think that's one of the brilliant things about them choosing that as the name of the town because for so long, people are like, well, where does this take place? It's like, well, it could be anywhere. It could be anywhere.

There is also, I mean, this is incidental, but there's also a Laporte in most states. Did you know that? No. I think it's because Mr. Laporte got around. Many, many years ago, my ancestors, I think he was a trapper, a fur trapper, but he really, yeah, he got around. Did he trap marmots? Yes.

I certainly hope not. I don't think the marmot fur really makes for much of a coat. Yeah, I wouldn't think so. Yeah. I have a picture of a marmot that I took. I was, I know. What are we doing here? Why are we here? What is going on? I was in Tasmania on a photo shoot expedition with some really good photographers. And for some reason, I decided to lie in the grass and

And a marmot came trundling toward me. And I didn't really know if they were dangerous or not. So I said, help. And they just laughed. I don't know if I could find it. It's buried somewhere in my pictures. But yeah, before the show's over, I will share my, this is a little tease to keep you all watching. I will share my marmot photo with you.

okay enough let's talk ai much more interesting six takeaways from a monumental week four ai from the wall street journal guess what all six of them were about deep seek the chinese ai um which has really rocked kind of rocked the stock market rocked nvidia

DeepSeq claims, we don't know, actually I'd be curious what you know, Dan, covering AI and all this. They claim that they trained their model for a mere $6 million with an M, and it's very efficient, and they, because of the export restrictions, weren't using the latest NVIDIA cards, didn't need to. And this, of course, tanked NVIDIA's stock. It was the world's most valuable company the week before NVIDIA.

went down 17 Monday and has been going down all week one trillion dollars worth of stock market value gone uh did start to recover towards the end of the week but we don't really know uh what deep seek cost I mean we just know what they say what do you think Dan well I think you know as a reporter and as somebody who works in AI I don't know anything more than anyone else yeah we don't I do think opacity is the story here right like this

This launched onto the market so quickly. It was so disruptive, and we know so little about it, but it had so many great knock-on effects that it's almost impossible to look away. Yeah. Then, of course, OpenAI...

in the funniest uh reaction ever said they stole our they stole our stuff from our model they copied us which of course is funny because open AI basically copied everybody else you know they copied the internet in order to train their model uh deep seek apparently used a technique called distillation which is training your AI model by asking another AI model questions and learning from its answers

And of course, you could do that at great speed if you're both AIs. And DeepSeek said it distilled from an open source model from Meta, Lama 2, and from one of its own. But OpenAI said, we're looking at indications DeepSeek used us for distillation as well. So what? It's my reaction. Like, well, that's fine. That's how it works. I think it's...

It's so interesting that we didn't know anything. I mean, we didn't really anticipate this product happening when, but we could have certainly anticipated this event happening and the AI bubble taking a hit like this. But there are so many fantastic questions that this opens. I mean, was it trained for so little? If that is the case, then what are the implications for all of these models and companies seeking billions, if not trillions of dollars? Yeah.

I'm also not sure it hurts us. In fact, Sam Altman had an interesting tweet. He kind of said, "Open AI has been on the wrong side of technology." It's really, it was an odd thing to say. I think, he says, "I personally think we need to figure out a different open source strategy."

Open AI has been on the wrong side of history when it comes to open sourcing its technologies. It's hysterical, well, not hysterical, but it's interesting to note that Open AI was founded specifically to be open, hence the name. Hence the name. And that eventually, I think they realized it was going to cost a lot more to train these models, and they created kind of a for-profit. And they hid the weights, which so it basically became a proprietary technology.

closed development. He said, "Not everybody at OpenAI shares my view, but it's also, and it's also not our current highest priority being open. We will produce better models going forward, but we will maintain less of a lead than we did in previous years." I guess it's funny because there really isn't much to say. It's like, okay,

It's interesting messaging. I mean, they could have been defensive and they could have been opaque themselves. They could have been aggressive, but instead it seemed to be a fairly conciliatory. And yes, I mean, I would call it a smart comms response. Yeah. I mean, I think it's a good consequence response or not, but it's a smart response. Yeah, I would agree with that. I do wonder, though, it does seem to go a little bit against the other reporting, which is that they think.

that, you know, they stole their research, their data, right? So it's like, so which is it, right? Like, I think that it's a conciliatory response. It's a good response. I think that being defensive about this wouldn't be good. But at the same time, you know, you say this right after, you know, their report's going, oh, well, we're going to look into how do they do this? Because this was clearly using, you know, our inference or, you know, data or whatever the case may be, which, you know,

how anyone's ever going to prove this. And how you would, even if you could prove it, how that would change anything at this point, I think is a very open question. But it's not a bad response. I think that it would look a lot worse if you came out and you said,

you either try to diminish what the technology actually was and what the breakthroughs have been and and we don't know how much this cost we only know what they say it costs and and that could absolutely be false but we do know that this is something that seemed to come out of nowhere that you know even compared with what we saw with r1 versus what we saw at christmas you know a massive um improvement and and definitely you know from a cost perspective um

is already making the other companies alter some of their strategies, which is fine, right? This is what you want. You want competition. You want back and forth. You don't want to operate on disinformation. And of course, the Chinese government might indeed have incentive to destabilize our stock market and NVIDIA's stock price, given the sanctions against it.

Sure, sure. I mean, it's at that point, that's why my one kind of question about this and a lot of, you know, the most important kind of data we got from this was from scales kind of benchmarks, which, you know, just kind of showed how good it was in addition to just anybody using it themselves and seeing, hey, wow, we're getting really good results and we can do these things incredibly cheaply on local hardware even. But one of my questions is, okay, you know, they claim that we were, and this is a very common Chinese government policy,

tactic and from Chinese companies to say, oh, we did all this ourselves and we didn't use any of your stuff. And we know they did use some NVIDIA chips. We don't know how many. I have no basis of knowledge for this at all. I want to make that clear. This is just strictly speculation. But I have a feeling that that number is higher than what they are letting on. But we saw this with ZTE and especially with Huawei, where the Chinese government wanted to tout

what types of chips and breakthroughs that they have themselves to create without TSMC and others. And it turns out, you know, the reality once you get past those things isn't quite there. And so I have a feeling that that's probably also the case in terms of, you know, how many NVIDIA GPUs they were using and how much this costs and how much time this took.

some of those things might not be exactly as they appear to be. But even if that is the case, this is still incredibly impressive and shouldn't be something I think that people should be angry about. You know, it's like this is progress regardless of who came up with it. Of course, David Sachs politicized it. He's the AI and Bitcoin czar in the new administration. And his contention is that he said, well, he was on Fox on Tuesday. It's clear. There's substantial evidence that DeepSeq used open AI's models.

Yes, and? Right. And what are you going to do? You're going to sanction them? Right, and what can you do about it? You're already putting those things into place. It's like this is not, they are not beholden to you. I hate to break it to you, David Sachs, but... I want to, by the way, just parenthetically, apologize to our friends from Australia that it was not a marmot. It was a wombat, right? Is that a wombat, Shoshana? Yeah, it's a wombat. Was that a wombat on your...

This is my wombat picture. But that's not what you're talking about. You're talking about... Different things. Yeah, I confused the wombat and the marmot. The wombat started coming for me. This is... Oh, they're so fast. It's walking closer and closer. And this is where I...

A little afraid. I said, do they eat humans? But no, they don't. They don't. They are really cute. I didn't remember Australia having marmots, but I was excited that maybe they do. So are you in marmots or are you in wombats? I'm sorry. What is your investment? Marmots. Okay. I'm a wombat guy myself. Huh?

shoshana is that your photo no that's not mine i have a lot of really good marmot photos but none were really suited for a background you can see why i'm confused doesn't that look just like very similar they look very similar little babies and i love them like they're so cute it is the wombat that has the hard butt though not the marmot so please don't run over marmots

I was going to be surprised if they had a hard butt. I feel like I would have known that, you know? They're squishy. I think they're squishy. Yeah, a little squishy. Yeah. Aww. So just to show you that... Huh?

i was just gonna say there's a tick tock of a guy in australia who has like a pet wombat with his with his daughter and it's the cutest thing it's like it's like they go around and like like the little girl she's probably two years old three years old and she and the wombat have basically grown up together and it's the sweetest thing you'll ever see in your life it is it is great it is like the best most pure content i love wombats but marmots look pretty good too anyway i just wanted to show you that humans are capable of hallucinating as well

You know, they get mad at AIs. Like if AI made this mistake, people would go, you see? You see?

but if a human makes it it's just like well he's just an old man and he's Dottie well you know what's kind of funny with with uh with everything too like I saw a lot of people trying to trick the model to saying anti-China stuff or say or just make China oh you can't get Tienanmen Square out of it that's for sure but I actually thought it raised a really good point that like open AI has it's like legal limits but our legal stuff is different than their legal stuff and people kind of overall they kind of get it but

when it comes to China's legal stuff, it just makes it less useful for people. It's like it can actually infringe on just the usability of the platform. And I think that kind of sets it up for failure. Like, I know people have issues with TikTok, and I do too. I have some issues with TikTok, but they don't censor stuff the way that DeepSeek is. Like, it's a very, very different kind of thing. And I think when you have that, like, state censorship at that level, it actually sets it behind and makes it, like,

less usable overall, which is kind of an interesting, like almost like diplomacy problem within the AI world. There is an issue with safety and deep seek notch. You know, I have mixed feelings about this whole AI safety thing. I feel like you're telling the AI, you know, don't talk about say Tiananmen square or how to make atomic weapons or whatever it is, is, is really artificial. You can't, but anyway, deep seek, you know,

Researchers at Wiz, which is, I guess, an AI security firm, found a major security lapse in the Chinese AI deep seek. Actually, it's a pretty bad one, but it's happened to many companies all over the world. They left a database exposed.

a click house database with a million log entries containing chat histories secret keys back end details this happens you know you you forget that you have an s3 bucket that has no controls on it uh the database allowed full administrative control without authentication um such a comforting breach i mean

I don't mean, I mean, I'm more worried about breaches, frankly, than safety, than AI safety. I think that's more of a concern. Oh, for sure. And it's, it's like, oh, if all of these fancy, super intelligent, very forward leaning systems, it's like you had a perimeter breach of a database. Yeah. Usually, usually what happens is the poor programmer makes a git commitment to

a git commit that has you know the secret keys in it this one they didn't have any secret keys it was just unlocked one one way one way you save on cost is if you just don't have an it if you just don't do it yeah i was gonna say right it's like it's like it's like um i have to remember the from a year ago the abbott the rabbit r1 the other r1 yeah the other r1 i did yes of course i did a thousand percent she's got it

Right here. It's the cutest thing ever. It's so cute. It's so cute. I'm there with you. Yes. Heck yeah. Heck yeah. But no, it makes me think. I'm a lot of teenage engineering. I mean, like. Hey, as did I. That's exactly how I justified it. I was like, this is the dumbest thing and this is fine. And then it turned out to not live up at all to its hype. And then the thing was, is that we realized the security was basically non-existent and that they were asking you to like log in with your Uber credentials and stuff so that it could, you know, have their, their,

like what they called Asians, but really it was just macros do stuff for you. And, but the security was like so bad. And, and I, I saw how the, um, the login works that it was just like a VM thing, like in the browser. And some other people who are smarter than me were able to even go deeper about how they were able to get doom running, um,

on like the VMs and it was just there's a good use for that though that's a good thing you know if you can get your R1 to run Doom I buy every stupid AI product out there these are the brilliant glasses nice which don't do anything I also just bought this was shown at CES this is the B

AI device it's recording everything all the time yeah I got one of those in June um not the final form factor I got one when it was still like a pendant that would go over your neck and uh and I immediately turn it off yeah no I'm loving it because it says kind of it's kind of silly but we're going to get the creators of it on our new show about AI but let me just read you some of the daily memories

This is from yesterday. Piano melodies, leaky decks, and Sinatra's palm springs filled the day.

but actually that's all true it sounds weird leo had a productive and enjoyable day he made significant progress on a new piano piece true mastering the challenge of playing chords with one hand and melody with the other he shared a laugh with lisa over funny video and received encouragement to record his piano playing later when leo went grocery shopping he and lisa discussed dinner plans i mean this thing is listening to everything

And I gave him access to my Gmail and my calendar and my contacts. What could possibly go wrong? I mean...

Why are you laughing at me, Shoshana? Shoshana's laughing. That's so stressful to me. Like, that's so stressful. It's funny, Christina, what you're saying about the Uber credential for the login. So I'm not sure if you caught this, but it's just funny because it's like another layer of the same problem. Last year, it turned out that the age and identity verifier for like

Uber for every like big Twitter like Bumble was breached for a year and a half. And I've been into this stuff. I know, I know, I know. I've been into this stuff from the age verification stuff because they're like, it's secure. You don't have to worry about anything. You're being silly. So like when I saw that, I'm like, oh my gosh. So it's so funny when you mentioned it that like that it was happening through like your Uber credential. And I'm like, guess what? You have to like be careful of that stuff already because everyone knows it because nothing's secure. Yeah.

so i i mean i have my attitudes towards privacy are probably not typical of most

I don't really care because I'm on the air all the time, so it's nothing really private. They give you a, this thing generates facts and then says, are these true? So this is one of the facts that just generated. Leo has issues with ankle socks not staying up. That's true. Leo's concerned about the freshness of grapes. That's true. Leo considers fish a light meal option. That's true. Leo has meatball mix in the freezer from Whole Foods. That's true.

Leo has a task related to getting breweries tickets. It adds this test to my to-do list by the way, which is great. And that is also true. Let's see, I'm showing you these

Leo is comfortable using voice commands to interact with technology. That's true. So it now knows so many facts about me. What's your issue with ankle socks? They fall down. So like they keep slipping under, right? Yeah. So you want the ones with a little tab on the back that like changes. It changes life. Like I'm like loyal to little tab ankle socks. You see already AI is changing my life. It's making it better.

I'll get the tabs. Thank you. You're welcome. Yeah. I always thought ankle socks were dopey. I used to wear socks all the way up at my, I don't know, it's a long story. That's stylish now. Higher socks are stylish. I was going to say, stylish socks are now in vogue. And in fact, like if you have, like I used to always wear no show and then no show last year became like passe. And so now you have to wear the high socks. I don't want to see men's ankles. Women, it's okay. But men's ankles, I don't want to see those.

But you shouldn't wear those with sandals, am I right? The high socks? - Right, right, correct. Well, actually, if you wear them with the sandals that go across, not the toe sandals, that's like sporty and stylish, you know? - They're saying, "It's not the AI that has changed my life, it's Shoshana." That's true, that's a good point, that's a good point. AI systems with unacceptable risk are now banned as of today,

in the EU don't know what an unacceptable risk is. Apparently the EU isn't really clear either. It's designed to cover use cases where AI might appear to interact with individuals. Okay. Minimal risk is for instance, spam filters will face no regulatory oversight. Okay. Limited risk includes customer service chat bots. They'll have a light touch regulatory oversight.

High risk AI for healthcare, that's a good example, yeah, will face heavy regulatory oversight. And then the unacceptable risk applications, which is what these new laws or new regulations cover, will be prohibited entirely. AI used for individual social scoring, like in Black Mirror, building risk profiles based on a person's behavior. AI that manipulates a person's decisions subliminally or deceptively, like convincing me to put tabs on my...

Ankle socks. AI that exploits vulnerabilities like age, disability, or socioeconomic status. Now, I know you are not a fan, and R Street in general is not a fan of over-regulation, Shoshana. Does this seem over-regulation or appropriate?

So they don't know what they're doing. I mean, like they, the Europe's just been trying to like kill technology for years now and it's absurd and it's why they don't have a lot of tech companies. I think like, like with mental health and health services in general, I think it's worth having oversight of AI, but there's,

I mean, the way they're going to go about it is just fine, everyone, and be really sick over carrot. I've been talking with Utah, actually, and they're doing some really cool stuff in how they're trying to regulate AI. They're just trying to figure out what's the framework that's going to work best in each situation because they want AI to be something that can help people

for health services, but obviously they see the risks. So they're working like really cooperatively with a bunch of different companies to figure out, okay, what's going to work best here. What's, you know, they're taking a really collaborative approach. And I think it makes sense to do, especially with emerging tech that's happening fast, that like, you don't want to stop this from,

you know, from growing, but you also want to make sure that it's safe, especially in certain circumstances. But Europe has just tried to crush everything and like get lots of positive attention for crushing things. So even here, like I'm not, and I don't think even we're against, right,

regulating a little bit more here, if it makes sense. It's just about figuring, do regulations already cover this? Like discrimination is already, you know, regulated in a lot of cases. Providing crappy health services is already regulated in a lot of cases. But if there's additional layers, that's, you know, it's something we're open to, but Europe is a terrible model for that. You know, I feel like their GDPR is giving us in the States some degree of privacy, whereas our Congress seems unwilling to do that.

So there's some things that they're doing

that are good, right? You think GDPR goes too far? - I think from a cybersecurity perspective at least when it comes to AI and regulation, we've found, or I shouldn't say we, but I have found and seen that Europe has been pretty amenable to a lot of types of innovation, but that is at least in that limited experience and from a cyber perspective. But I mean, right now it kind of seems like a very good place to look at cybersecurity. - Yeah.

AI that uses biometrics to infer a person's characteristics, like their sexual orientation is banned. AI that attempts to, this is interesting, this is something we are doing, I think, in the United States, attempts to predict crime based on people's appearances. Oh no, that's just what the cops do. AI that collects real-time biometric data in public places for the purposes of law enforcement.

Terrible. Have you guys read the book AI 2041 by Kai-Fu Lee? No. Oh, Kai-Fu Lee is great. Yeah, he is. Leo, as you're reading this, it kind of feels like almost a chapter by chapter breakdown of this. It's really a book of short stories, but they're all thematically connected. But we talk a lot about the future and we love science fiction because it kind of

It helps us imagine the future, if not predict it directly. But this is a very down-to-earth, because of Kai-Fu Lee's background. This is a very down, not down-to-earth, but a very practical forecast of where we could be with AI by mid-century. And again, as you're reading that list, Leo, it really feels like this book come to life. Yeah. Yeah.

And he's warning us against this, I presume. I don't know that it's... There are certainly some warnings in the book, but there are also... There's more nuance. It is from a Chinese perspective, but there's... Especially when it comes to...

the relationships we will form with artificial intelligence and the way it could augment our life and become a type of relationship. If you know, right now we look at like, well, maybe there are human like avatars, this kind of short term thinking. But as we, as we progress, we might find these agents are part of our life and express themselves because they have such access to data in really, really unique ways. Yeah.

We in California had a similar anti-AI or, I'm sorry, not anti-AI, but regulating AI law that the governor declined to sign. He vetoed that required, among other things, that companies of a certain size would have to have a kill switch for their AI, unlike RoboCop, I guess. It seemed, I don't know where I come down on this. I feel like

Just as in the early days of the internet, you could make a mistake by rushing to regulate it too quickly. And we didn't do that with the internet. Congress was very careful to kind of have a light touch on the internet because we didn't know what it was going to be.

In hindsight, maybe you could say, well, they should have done something more to protect us. I don't know. But I worry that by, and maybe the EU is doing this, by over-regulating AI, we're preventing it from growing in a way that might be very useful. Agreed or disagreed? Am I wrong on this?

Yeah, that's where I come down basically and our street worries about too. We're always down to figure out in certain use cases, in certain if there seems to be more of a risk in one area than another. There can be like regulations that attack this that we're not opposed to, but so much of this just seems to be

broad and tons of requirements that don't really have a real focused end. It's just like, oh, you know, these systems should have to tell the government everything, which I don't think is great either because people, you know, you want to be careful from the privacy

when it comes to the government itself. We're also for a national privacy law, but that's a whole other debate. And we're also really interested in the cybersecurity side of stuff. So one thing on that side we're doing is trying to get the federal government to align its cyber standards and reporting standards across agencies so that companies know what they have to do. And it's a little easier to hold them accountable as well. But when it comes to AI, it's so unfocused and it's so...

I mean, so much of government is, I think, afraid of experimentation here. But then you even see other parts of government really embracing it. A couple of states have used AI to try to review their regulatory code and make it in human language and simpler and something that they would pass through. Perfect use for an AI. There's so many good uses of it to make government better. Another one that I found out about is using it for permitting applications in, I think, California, actually, and that it got

through permitting requests a lot faster. And I think you can use that for occupational licenses and other stuff as well. But it's always a little bit frustrating because

You know, government is so focused on stopping things just in case rather than focusing on fostering the good potential and figuring out where the problems are and how to narrowly address them. And I just I worry about that because it's sexy to be against tech. I think that's starting to go away a little bit. I hope so. We're definitely in an era where it's just sexy if you're a regulator cracking down on evil big tech, you know? Yep. You agree, Dan? Yeah.

That it's sexy to be against? No, no. The earlier premise that we shouldn't over-regulate AI until we understand better what it's doing and what it is. Yeah, for sure. I mean, I'm of...

I'm not a policy person, and I think Shoshana is a million times smarter than I am about this. But I do think that we need an environment that safely fosters innovation. I have some trepidation if we look back at how social media played out. I think that perhaps we could have introduced regulation that –

encouraged innovation but slowed things down a bit. So I mean I have some trepidations. But that's in hindsight right? It's in hindsight right. We didn't know. Right we didn't know and maybe the benefits exceed the cost I don't know. Right now I'm not a big social media person but it could the same anyway I think I agree with Shoshana mostly but I opt for safety over speed. Okay safety over speed. Christina your vote.

Yeah, I mean, I would definitely say safety over speed. And I would go further to say this. I think that I would feel differently about how some of the regulation frameworks have come into place if I felt more confident that any of the people regulating this industry actually knew what they were talking about. Yeah, for sure. And so this is an area which is difficult, right? Especially when we're talking about emerging technologies where your experts are going to be your researchers or your people in industry and not people necessarily making policy.

but there has been, and I think I do agree with Shoshana, it has become popular and sexy to be against big tech. And that is for some good reasons. I think big tech has made a lot of decisions and made a lot of moves that have pushed...

The pendulum was in one very positive direction for a very long time. And then because of things, frankly, around privacy, around, you know, user protections, around things that now regulations, I think, understandably want to curtail people's perception and, you know, the attitude from people in making policy has shifted into the other direction. And so, but because of that, there is, and hopefully this gets better, there is now this friction

between the people doing the research and creating the products and the people setting policy. And friction isn't a bad thing, but I think where it is a bad thing is when you have the people who are setting policy who genuinely do not understand tech. And this has been a problem for decades, but it's only getting worse. And I do sometimes think, especially in the EU, they almost, in the United States, to be clear, but in the EU, it almost seems to be a point of pride.

Right? Like, we don't understand this and we don't care. Do you think that they, I mean, some of the, Mistral, for instance, comes out of France. Some of the best AI models are European. Yes.

Do you think, though, that this is kind of as we feel about China, we're xenophobic about Chinese AI, they're xenophobic about American AI? Maybe or maybe just their perceptions about how tech works are different. Right. I mean, like they have been much less risk averse and they have much stronger labor standards and different things are

you know the way that that the eu operates the way the united states operate are just different and um i think that um i think that having caution and having skepticism is good i think sometimes setting policy before you even understand whatever things is doing can be both a bad thing because it prevents positive use cases from taking place and it can also be bad because you might be you know setting to motion you know policies that

prevent you from preventing the really bad stuff that could happen, right? So I don't know. I feel like I would feel better about this if I actually had any confidence that the people setting policy knew anything and weren't just either listening to the experts that are saying what they want to hear, which oftentimes is that the sky is falling prognosis, or had any amount of digital

sensibilities at all. And at least in the United States, I can't speak for the EU, but in Congress, you know, the number of our, you know, people both in the House and the Senate who have any sort of digital literacy is almost non-existent. And that's very scary to me when we look at what's going to take place from a policy perspective.

- I'll cop to this. I am an AI accelerationist. I've kind of come to the point of view. My attitude is I'm so blown away when I, and I have a folder of AI tools that I pay for on my phone, literally, you know, probably 100, 200 bucks a month on all of these things. And I'm blown away by what it can do, by coding, by just answering questions.

I talked to a chat GPT. I have my action button on my phone. It opens up chat GPT-01. Actually, I think it's the new one. 03-mini? Yeah, 03-mini, exactly. And I ask it questions all... It's my research assistant. I am blown away. I think maybe we've gotten complacent about what's happening, but we are in the middle of a revolution, right?

It seems to me, now this is just my attitude and you can tell me I'm full of it and many have. I feel like, so what if it's smarter than us or we hit the singularity, let's see what happens. This is amazing and I think being cautious is gonna be, put a damper on something that could be truly revolutionary. And it already is. It blows me away what we can do in a year.

I agree with that with a couple of caveats. One area that I do actually, I'm in favor of more regulation for, and that bothers me very much, is when we see these systems being used

in cases like trying to identify, you know, people who might have committed a crime. But all right. But to be used to judge, you know, sentencing guidelines or to be used. When you went to see the Eros tour, there was face recognition as you went in. And because Taylor Swift has had stalker problems and it was looking for her harassers to stop them. Right. That's OK. And

And look, that's fine. And if you have those systems that are fed with data that can maybe show things, fine. What I'm talking about is like there are startups that are basically saying we want to basically ride along with police officers and take their body cam footage and I guess transcribe things and write their reports for them. Now, if you want to help with transcribing what is on the body cam, that's fine. But I'm not at all

comfortable from like so many, you know, like basic like like, you know, like constitutional amendment perspectives. I am not OK at all. Fourth Amendment, Fifth Amendment, like both of them. I'm not confident with saying that the cops should just be able to have AI generating those police reports. So many problems will happen there. Right. And to say, oh, well, we're going to describe what happened in the video and we're going to to write these things out in a better method. No, I'm not OK with that at all. And so I feel like there are some

And because we've seen as good as these tools are, we've seen how they're not perfect, how they can hallucinate things that aren't there. And frankly, because it is your job, that is your job as a police officer. If you want to help like go through a transcript and make sure that you didn't miss something, fine. But your job is to actually fill out the police paperwork.

and to do that. And that if I'm being charged with a crime, you better have actually done the work and not just trusted a computer to generate a report based on what was on the body cam footage and what it picked up based on what voices it heard. I'm not okay with that. One of the things the EU law prohibits is

I guess, profiling people pre-crime, trying to say, well, that person looks like they're about to commit a crime. And we know that face recognition often makes mistakes when it comes to people of color, does false positives. People have been arrested, they've been jailed. They've actually been convicted from, speciously from false reports by AI face recognition. So that's a problem, I agree. I think we should work on that

but i don't want to prohibit face recognition because there's some real valuable and honestly i don't if uh i mean if you can use an ai to generate your report

I think you should have to check it and you should have to sign it and say, I verified this. But I think that, I mean, we're using AI to generate show notes. I understand that's not exactly a mission. No, I, but I feel like those are completely different things, right? Like show notes for your show. If you get something wrong,

It's not the end of the world. Nobody goes to jail for that. If it's part of a police report where that is going to be used as the basis to charge someone with a crime and then take them through the process, I think that's completely different. At the very least, there should be massive red flags so that lawyers know this was aided with this instance of AI. But you know they won't do that. You know they won't. And I just feel like this opens up lots and lots of, you know, like...

I just see lots of false convictions happening because of the basis of these tools. And that does scare me. And so on the one hand, like, I think we should experiment. On the other hand, like...

I'm very, very nervous about how this could be used in law enforcement just because I don't have a lot of trust in a lot of law enforcement anyway. And these sorts of systems could make that even more complicated. That's just me. I think a good example here is like from the bias that they found in AI system, just like how police bias can get into anything, just like any other bias can get into something. So if they're like working with a police force that just for whatever reason, there tends to be a lot of biased members and like

They could have data that shows more aggressive reports about black people, just for example. And if that goes into the data, then that could be dangerous. I'm open to testing it, but I definitely wouldn't want that to be the final thing, at least for a long time. There should be a lot of work to make sure that bias doesn't get in there, which is a hard thing in general to do. Yeah, bias gets in from humans. Yeah, of course.

There's a differentiation here between what we, when we talk about safety and AI and, and they're both important, but there are two different things. What Christina is talking about is maybe the, the expression of, or the potential dangers to society, which could happen in a lot of different ways. And then there's the AI training and the AI, uh, acceleration safety conversations, which are more, a little more technical and, uh,

about how quickly we should develop the actual capabilities of these models. Two very different, very valid conversations, but different and distinct. Yeah.

i want to take a little break we have lots more to talk about and of course uh mentioned somebody was saying in the chat you guys should have an ai show we do have an ai show it starts wednesday we're rebranding this week in google to intelligent machines uh the guy who coined the term intelligent machines ray kurzweil will join us next month on the show and we're going to bring in ai experts because i freely admit i i don't understand it well enough to have a strong opinion i just i'm excited about what it's doing and what it can do

And I think we need to talk about it. I think it's an important part of our mission. So tune in Wednesday for the first episode of Intelligent Machines, 2 p.m. Pacific, 5 p.m. Eastern, 2200 UTC on TWIT.

uh great panel so nice to have christina warren you're going to tell us what your new job will be sometime tomorrow maybe yes yeah yeah it should be sometime tomorrow or sometime next week uh starting next week um i it will be uh it'll be when you start i start tomorrow um uh technically i don't know when i will have access to all of my work systems and whatnot but yeah i probably will make the announcement tomorrow are you gonna move

No, no, at least not right now. No plans moving. I mean, that could always change. I'm sure GitHub was sad to lose you. That's all I can say. And I was sad to leave them. It's the most amazing place, the most amazing team. You're a great representative. I'm a huge fan, always forever, of GitHub. And GitHub Copilot, I think, is one of the best tools out there. There's an example, yeah, of something pretty amazing that AI can do.

fully agree yeah well good luck with the new career thank you and we will uh stay in touch of course we will yes unless they say and no more of that podcasting thing you know what i i i i'm very hopeful that that will not be the case but uh but it could happen it could happen it happened once with dan patterson uh when he left cbs uh but we we're glad to have you back director of content at blackbird ai

Dan Patterson.com still writes about technology for ZDNet. Always a pleasure. Yeah. It must be hard when the, when the company says, yeah, we don't want you talking on podcasts anymore. Is what it is. It was sad. We lost, we lost your input. We're glad to get you back. Appreciate it. Great to be here. Yeah.

and Shoshana Weissman our street is enlightened and says talk as much as you want head of digital media at our street I guess when you're the head of digital media and you hang in with marmots uh you know not a wombat

No, no, not a wombat. Wombat heads are bigger, I've been told. So that's how you can tell. Yeah, that's a really good point. Also, they're gray. There's a variety of colorings in marmots. This is the yellow-bellied marmot. That's like common to the West of America. There's a bunch of variations, though. There's got to be at least seven subtypes. Like there's a Himalayan marmot, totally different. I now have a new insult. You yellow-bellied marmot, you! Great.

Great to have you Shoshana. Thank you all for being here. Our show today brought to you this week by US Cloud, the number one Microsoft unified support replacement. Why would you want to replace Microsoft? We've been talking for a few months now about US Cloud, the global leader in third party Microsoft support for enterprises. I can give you three reasons.

Number one, you're going to save switching the US cloud can save your business 30 to 50% over Microsoft Unified and Premier support. Maybe that explains why 50 of the Fortune 500 now use US cloud for their Microsoft support. But it wouldn't be any good if it was just less expensive. It's also got to be as good, right? It's better than Microsoft support faster, twice as fast average to time to resolution.

Twice as fast average time to resolution versus Microsoft. We're talking minutes. But now US Cloud is excited to tell you about a new offering. They do a lot to help you save money, not just by cutting the cost of your Microsoft Premiere support. For instance, their Azure cost optimization service. This is very cool.

When was the last time you evaluated your Azure usage? There's a little bit of Azure creep in there, right? If it's been a while, you probably have some Azure sprawl, little spend creep going on. Good news, it's now easier than ever to save on Azure. US Cloud offers an eight-week Azure engagement powered by VBox.

that identifies key opportunities to reduce costs across your entire azure environment really be honest do you know exactly where all your azure dollars are going with their expert guidance you'll get access to us cloud senior engineers by the way this is the third reason they're better they're lower cost they're faster and they're smarter their senior engineers have an average of over 16 years

16 years with Microsoft products, they are the experts. I talked to the folks at US Cloud, I said, "How do you attract these people?" They said, "Well, they're all US-based. We offer them great salary, great benefits. We make it a great job for them, and that's how we attract the best engineering." So anyway, back to this Azure engagement. At the end of the eight-week engagement, your interactive dashboard will identify rebuild and downscale opportunities and unused resources.

it will actually surface where you're spending that you don't have to allowing you to reallocate those precious i.t dollars towards needed resources and may i make a suggestion invest your azure savings in us clouds microsoft support actually that's what a few of us cloud's other customers have done completely and eliminate your unified spend i'll give you an example sam he's the technical operations manager at bead gaming

He rated US Cloud five stars out of five saying, quote, "We found some things that had been running," this is the Azure engagement thing, "We found some things that had been running for three years, which no one was checking." These VMs were, I don't know, 10 grand a month? Not a massive chunk in the grand scheme of how much we spent on Azure, but once you get to 40 or 50,000 a month, it really starts to add up. Yeah, no kidding.

It's simple. Stop overpaying for Azure, identify and eliminate Azure creep, and boost your performance all in eight weeks with US Cloud. Just one of many reasons people are switching to US Cloud for their Microsoft Unified and Premier support. Visit uscloud.com. Book a call today. Find out how much your team can save. I think you'll be impressed with how good these guys are. US Cloud.

Go there right now, book a call, get faster Microsoft support, better Microsoft support for less. And by the way, if they ask you to say it, make sure you tell them, oh yeah, we saw it on Twitter. That helps us. UScloud.com, they're there to help you. All right, we're not quite done with AI. We'll move on. But that was most of the news this week was DeepSeek and AI and OpenAI.

They said that they are going to partner with the U.S. National Laboratories on scientific research and nuclear weapons security. Maybe this is the kind of thing that might have made you get nervous. They're going to deploy an open AI model on Venado. That's the supercomputer at Los Alamos National Labs. Los Alamos is where, of course, they...

our nuclear program, our nuclear weapons get developed. 15,000 scientists working at the national labs will be able to access O1, the new reasoning series. OpenAI will also work with Microsoft to deploy one of its models on Venado. Venado is powered by NVIDIA and HPE. Working on nuclear weapons, quote, focused on reducing the risk of nuclear war.

Shall we play a game? And securing nuclear material and weapons worldwide. I think if they work on security, that's probably a good thing. Yeah? Yeah. Go ahead. My strong informed opinion is just securing nuclear weapons as opposed to leaving them unsecured is... Yeah. I'm for it. In general. Yeah. Yeah. But if we give...

the engineer has access to 4.0 doesn't 4.0 also have access to the engineers uh that's a good question no because I guess it's running on the computer inside the labs right so in theory it wouldn't exfiltrate stuff out of the labs that's always a concern like my little AI friend here who's recording this entire conversation they don't even I don't even know where it's going it could be going to China I don't know I don't care listen all you want uh

if it if it wants to hear me play uh the saints come marching in over and over and over again on the piano go ahead listen in okay so let me ask you a question though how would you feel if everything that you said to this thing was released somewhere and then emailed to like your top 10 like most contacted people in your address book

Well, it's not going to do that, but even if it did, so what? I'm just saying, because I think that that's the real kind of worry that a lot of people have about this stuff. Isn't so much like, oh, these things have my system, but like this could get out there. I am not knocking anybody's desire for privacy. That is absolutely your choice. I think I gain an advantage. I think, by the way, this is why. So one of the things people are talking about is Apple intelligence and how stupid it is.

And it's bad because it's private. It's on device. It's limited in the size of the model, but also it's limited in what it knows. And honestly, I think if you're willing to give the AI everything, even if it emails it to my 10 top contacts, you're going to get more from the AI, right? And I think that's what Apple's finding out.

Maybe. I think there's like a broad spectrum between the two though of uses. I wear this to my doc... somebody's saying, "What if your doctor was wearing it?" My doctor does record our conversation. He has a sign that says, "I'm going to use AI to transcribe this and we're going to delete it." But I also am recording him. And that way I, you know, when I get notes back, I don't remember everything I get. I think it's just, it isn't obviously a complete trade-off between privacy and intelligence.

But I do think that the more willing you are to give up on privacy, the more valuable the AI will be. They said the same thing about social media.

Yeah, and I fell for that too. But I've deleted all my tweets since. Me too. I still, by the way, I still check X, but that's kind of part of my job, right? I have to see what's going on out there. It's interesting, the national, what is it, the NTSB, National Transportation Safety Board, has announced...

that it is not going to update the press in the way it used to with emails, press releases. It's all going to be on X.com. This, to me, seems like a bad idea. It seems like it's really kind of a little, well, backsheesh to Elon Musk, right? Yes, clearly. And putting aside who controls X, or Twitter, whatever we're going to call it, because that, to me, is the least...

most bothersome part about this. It's like, these are third party services and you're talking about like official communications that you're relying on a platform that might change its, you know, mail might change like what it wants to do. Like what if they decide that they don't want to, you know, display things the same way as like if there's, gosh. - What if the entire US government decided to use X or Worth Truth Social for all of its public announcements?

It seems like that would be a misuse of our taxpayers dollars. National Transportation Safety Board will only update the press about plane crashes on X, not over email. They announced on Saturday they will use the @NTSB_Newsroom account to share news conferences or other investigative information. They said this is to help them better manage their incoming emails.

NTSB later said reporters should email [email protected] for all other inquiries. Oh, it's specifically about those two incidents, the two plane crashes. Oh, okay. Well, I'm not against using X in addition to the normal, but if you stop having press conferences and media briefings, you just use X. That's not good.

Yeah, also, like, I mean, the government was kind of slow to adopt social media in helpful ways. And I'm all for having stuff in more places, like hook up the automations, have everything go everywhere. And I'm all for that. But yeah, it doesn't...

Doesn't make sense to do it this way. Even if they only want to open up inboxes in certain places, like if you can only inbox them via like the regular email plus Twitter DMs, like I get it because most people use their Twitter DMs as like a secondary inbox if they're tweeting a lot. So I get that.

but yeah this is i mean like that's bizarre even i mean even if they wanted to do this for these two incidents just host it like make it a page on the site where like follow these threads you know like make it easy to find for yeah no i have nothing against them announcing stuff on x yeah just not exclusively x that's what's weird thing yeah and requiring a login well right well oh that's right you have to have a login don't you you can't just look at a tweet

I mean, and look, not that this would matter with this administration, but I think that that lockdown thing that you bring up, Dan, which fantastic point, that might violate some sort of either accessibility or some sort of other like actual like government guidelines that exist in terms of getting people access to these sorts of things. So, yeah. One of the replies on X is NTSB is breaking the law. Government agency is not permitted to restrict its obligation to inform the public by making it available only on a privately run platform.

Precisely. Right. Like, I think I think how it on X is great. Like, that's actually a great use case. And to Shoshana's point, I think government was slow to adopt some of those things and and and, you know, maybe too slow. But like have an RSS feed, have like a separate page that's up where you can put your updates that are cross posted across those things. Set up the automations, you know, even their own government mastodon and sort of cut you off, Christine. Oh, yeah. No, exactly.

is this so as somebody on on x said a new level of transparency i don't think so i mean it you certainly redefine transparency um okay well just just curious um by the way that's the uh same platform that the naked video of the current uh new uh secretary of transportation uh a video of him dancing naked on the on the

big brother but that's an it's another story the real world the real world I mean yeah yeah yeah yeah yeah it's nice to know that he's got he can let his hair down have a good time there are no videos of me dancing naked on real world I just want to say even if me neither

Even if this records everything. All right, let's talk about O3 Mini. This was released. It's interesting. OpenAI responded to DeepSeek, and this is why competition is a good thing. Responded to DeepSeek by saying, oh, oh, oh, let's release O3 Mini, its latest version of O3, the reasoning model. I like these reasoning models. DeepSeek is a reasoning model, too. They take a longer time, but you can watch them as they kind of come up with the answer. I think it's kind of an interesting model.

uh thing sometimes it takes 15 20 seconds a minute but often they say the answers are better open ai is saying the new model is both powerful and affordable because it is mini which means it's a smaller reasoning models fact check their answers before giving out results uh oh three mini fine-tuned for stem programming math and science i immediately gave it

a uh i cut and pasted from the advent of code day seven problem that has been stymieing me since december and literally in 30 seconds it came up with the correct answer and all the code in common lisp no less a uh an ancient language when apple intelligence i'm just looking at my phone right now when apple intelligence um opts to use gpt plus do we know if if we are able to choose the model

Yes, you can. You can. Well, I pay for it. Yeah, me too. I'm just looking at my settings. If you have a paid account, you go to your ChatGPT app. There's a dropdown and you can choose. Choose. Yeah.

Yeah, so if under model, I can choose GPT-40, 40 with scheduled tasks, 01, 03 mini, and 03 mini high. And this is what it responds to. This would correlate then to using the Siri button on your iPhone, right? Oh, I see what you're saying. You're asking if I use Apple intelligence. Yeah, right, right. Can I choose? No, I don't think so. I think Apple made a deal with the...

open AI to use GPT for period. It's just for, for that's kind of what I thought. But, uh, so I, I do almost, I thought you meant in the, in the chat GPT client, but no, I mean, that's, that's a good tip too, because I, I didn't look to change that. I do something very similar with the action button, but I put it on perplexity. Um, same thing, the same thing. Right. But I, but it's a little easier to choose the model. And I, that's why I do it because it's in the interface. It's just fast and easy to choose the model that I'm interacting with. Uh,

I had perplexity until 03 Mini came out, then I changed it. I go back and forth. Every once in a while, I change it to Claude as well. Yeah, same, same. They're all different, but they're all kind of, I feel like, remarkable. Am I just being suckered? No, I'm there with you. I am too. Yeah. It's very easy for us to kind of take it for granted, but look at what we were seeing two years ago.

compared to what we're seeing today. It's we're making amazing progress. Look, everything on our site at Blackbird, not to log roll for us, but everything we partnered with this great firm Punch and every design on the site is AI generated. And my role is I work with AI engineers on one hand, and on the other hand, I work with intelligence analysts. And often they both send me this incredibly intelligent

buckets of notes, but I use AI every day to synthesize that. Otherwise it would take me forever to do my job. And, and using a combination of the cloud API with GPT, uh, uh, four Oh, and, uh,

perplexity that really chaining those things together well chaining those things together I can also see the the power and potential of agents but that's like that is my day is throwing throwing things

between different AI systems. Agentic AI does stuff for you, right? It books an Uber or it writes a little script to get you to automatically download every article mentioning Blackbird AI, that kind of thing. Yeah, right. I think you might hear a lot of jargon. I mean, I know you're going to hear a lot of jargon about this and a lot of often, and I really dislike jargon because it's off-putting and it doesn't bring people in. It, it, it,

kind of creates a shield and a wall. So really, agentic AI is automated tasks, and that is one certain capability that's a consumer capability. But if we look at agents at large scale, especially running in the enterprise, there's a lot of potential because you could chain very complex tasks together and run them automatically. But what we...

We're also seeing this, you know, one of the really disruptive and very interesting knock-on effects of, I keep saying knock-on effects, sorry for that cliche, but one of the knock-on effects of DeepSeek is, well, if this really did cost just $5 million to train, then maybe we could see a crazy democratization of agents and of different types of agents.

AIs, what really might happen, the potential future that I'm excited about with artificial intelligence is if the cost to produce it and the energy cost to produce it and to produce high quality generative results has sunk dramatically. I mean, that's from an environmental perspective, that's really incredible. It also makes these things far more accessible to broad numbers of people, far more than it is now. Yeah.

What, Christina, with Copilot, of course, you used to work for GitHub, but I'm sure you only want to say good things about your former employers. But in general, the people that you talked to, you were in developer relations. Were they using Copilot as an assistant or to actually write code?

It was sort of very, right? So I think primarily assistant, but the models have gotten better a lot faster. So, you know, Claude 3.5 Sonnet for code especially is really, really good. In my opinion, it's the best one of them. And Copilot right now offers, I think they just added O3 Mini into Copilot chat and O1 Preview and O1 Mini was available at the end of last year. Gemini 1.5.

is available for some users and that has obviously a really big context window, which is awesome. And, you know, of course, like the original GPT-4.0, but like in my experience, Claude has been really fantastic at code. And so the primary way to use it right now is

to kind of, you know, be kind of auto-complete and in some cases maybe generate, you know, big blocks of code or functions depending on what you want to do. But what's starting to shift and we've seen this, especially with less technical people who maybe don't have...

like maybe like the the fear about like how to interact with these systems and hear me out because this that that sounds antithetical but but but kind of makes sense where you know they will go into chat gpt and be like help me write a program for this and and get fairly far along whereas people who might have a little bit of experience might be more hesitant to chat with um the the assistants the same way but what we've seen especially with some of the newer models and and

with reasoning, you can start to actually get some really good solutions with this, with things like O1, and I'm sure with O3 and R1 and things like that, would be to solve and kind of debug more difficult problems. And in that case, I

I do think that we've started to see and you can see more code generated from end to end that's AI. Now, how much stuff is going to be committed to production that is AI generated? I think that that's going to be more minimal, right? But in terms of running tests, sure. And maybe in doing certain sorts of reviews or helping do debugging,

that's already happening. And then you have tools like Spark, which the GitHub Next team is working on, which will kind of allow people to kind of enter it in plain text. Well, I want to do this and help kind of create smaller, you know, mini apps to do those things completely in natural language. GitHub Workspace, which is similar, but also can bring, you know, give the user a lot of, a higher level of control involved with those systems too, can do that. And so,

I think that the primary thing right now, to answer your question in a very long-winded way, is still kind of completion, but it's even in the last three months, the amount of code that can be generated and generated well has exponentially increased.

Yeah, Mark Zuckerberg said that he thought this year they would be able to replace senior engineers with an AI. Yeah, I don't know about that. I mean, I think that maybe with the right instruction and with the right kind of babysitting, you might be able to have things that says, okay, you know, create these things. I don't know, right? And he obviously has access to way more research than I do. And that's...

That certainly sounds good to investors. Well, he also, and this is something we have to remember about Sam and Mark and the whole gang is they have a vested interest in selling this. I was going to say, they do. They're marketers. Yeah. Exactly. And that's why I'm saying, I'm trying to like put some realism there. I think that some types of code you could generate, yeah. And in some ways, like I even know just from using, you know, systems like Copilot Chat or even like,

Claude to say help me write you know a regex that might be like a really difficult one like I did one oh god yes yeah there was

I stumped it actually, and it took me a few back and forth to go with it. But the reasoning models were really interesting because they would run tests against it and then say, well, that didn't quite work. Let me try this again. Well, that didn't quite work. Isn't that fun to watch it think? Yeah. It's kind of amazing. It is kind of amazing. But that was the sort of thing where I had to kind of go back and forth with a few prompts to get the correct regex. Now, to be clear, it still took me a fraction of the time it would have if I had to write that manually. Right.

But these models are getting better and better. And so I think that with the right prompts and the right kinds of instructions, you might be able to get solutions that would match whatever engineering level they're wanting to claim might do. But in terms of

how much hand-holding will be involved in that. I think that the idea that you could just let these systems loose and just give them a very simple prompt and they would give everything back exactly as it's needed to be, I think that is not going to be accurate. I think that what is more likely is that these might just be more assistance that could work alongside senior engineers who know what they're doing and they could go back through and code review and go, "Okay, this was correct. Now I can move on to another task."

but that's not the same thing as just being able to take anybody off the street and say, oh, well, you can be, you know, the person who just writes the instructions now. And we have this senior AI engineer who can write it all. And we don't need to worry about you being able to check if it's good or not, because you don't think that's just around the corner. You don't think I don't know how far that is. I mean, I don't know if it's in a year.

uh five years i wouldn't say i wouldn't bet against that but in a year i i don't know i still feel like i think it's gonna happen this year thanks i think it's gonna happen this year i'm with zuck i i know that he you know that could well be hype but i actually think we're very very close remember instructing a computer what to do see i don't i don't i think ai creating art writing music

pictures, that's one thing. That's a human skill that requires humanity. Coding is very much the opposite. It's telling a computer what to do. What better agent to do that than another computer?

You're not wrong, but I guess what I'm pushing back on is the idea- And I don't mind. I mean, it's not like, I mean, as much as I like to code and it is an art form for me, there isn't much future for artisanal coding in the world.

guess what i'm pushing back on is is how much someone who doesn't already have an understanding of what it what is being generated is correct or not like if you're still going to need someone who can who can help do the review and make sure that things are being done the most efficient way that you don't have lapses insecurity and whatnot because i guess you could write tests you could write test harnesses of course you can of course and you can run those tests but i'm saying you still need someone who's going to know what a test is you

yeah just another just get the get claude to write the code and then uh oh three mini to write the tests and if they agree everything's fine yeah that won't that won't lead to anything humans are notoriously bad at writing code i should point out yeah right we are we are but but i feel like because it's not like we're the paragons of perfection no i just feel like you know that's why you try to have like

actual as many eyes on something reviewing it as possible before things go out. And obviously we could always have more of that. And I think that these systems, especially in automated natures, if they can look over stuff and go, hey, we saw this string and this seems malformed and this could be a security issue alerting things in advance. I think that's an awesome opportunity and a use case we've already seen with some AI systems. I'm just saying,

I think that this idea that we can just have these things running autonomously without having to have any human interaction or checks at all, at least in the next year, I don't think that that is going to be the case. Why did you buy your R1? I want to ask both you and Shoshana, what you thought that little orange thing was going to do for you? Was it just a curiosity, Shoshana? Oh, I didn't buy one. I don't have that. Oh, you have one. I'm sorry.

Shoshana and I are the only ones who didn't. I tried, by the way, but my credit card was refused. So I did not get one. But I did try right when it was first announced. So, all right. Dan, why did you buy the R1? Was it just an experiment? Was it like me buying my little B computer? I mean, first of all, just fun. It wasn't a crap ton of money. I like teenage engineering. Yeah, it was cheap. Still overpriced, as it turned out, but...

It seemed cheap at the time. Right. We just had a conversation about like how we put chat GPT into our phone and talk to it. And, you know, I can take a picture of stuff. And I mean, that's really what this represented in a fun design. And it, I mean, really the selling point was it came with a year of perplexity. So it was like, well, if I'm going to pay for a year of perplexity anyway, then that's

That's a good point. I didn't get it, but I did pay for the year of complexity. Did I know how fun this story of Rabbit would turn out to be? No, Leo, I didn't anticipate how incredibly entertaining this story would be. However, I mean, like...

it does feel a little obnoxious to my broke punk rock youth that I would have an extra 200 and some bucks to like blow on a thing that I don't care about. But like beyond that, whatever. Well, these meta glasses cost me that much and they're equally fantastic, but those are awesome. Those are awesome. Are you kidding me? Those are the greatest things. I think,

i think the meta i think the meta glasses are great what do you use your metaglasses for in a million years just listening to music win the ai race well apple we're going to take a break but when we come back apple has decided not to compete and we'll talk about that in just a little bit good do you think they look good i look a little nerdy in these no but i am listening to uh to rap music while we're talking right now yeah you can't hear it but i can't listen to the boys of summer

with your way the boys yeah you got wayfarers on yeah yeah gotcha oh yeah oh hey good yeah good good catch let me see if i can get that to uh play boys of summer christina i'll find out why you bought that little orange doohickey in just a minute you're watching it's playing uh i don't know what it's playing is it doing it it's playing something can you hear it even if you have a monitor in yeah yeah can you hear it sounds like miles davis it's actually quite nice

Anyway, that's rude of me to listen to music while I'm talking. We'll have more with Shoshana Weissman, rstreet.org, Dan Patterson, blackbird.ai, and Christina Warren, an employer to be named later as we continue this. Unappointed for one more day. One more day.

I'm glad you took a cruise. That's a good use of your time. Yeah. I'm a little nervous about leaving the country. I'm afraid they might put a tariff on me when I come back. I don't know. I'm just nervous about leaving the country right now. I don't know if we'd be very welcome as we travel around. But the Bahamas, that's nice. Everybody loves you in the Bahamas, right?

I mean, I went to the cruise lines. Private Island. Yeah. And that was it. No ferners there. I didn't get off of NASA. I've seen NASA before. I was fine with that. I was like, I just want to be in a hot tub. I'm good. The best time on a cruise ship is when everybody's gotten off NASA and then you have the ship to yourself. Exactly. It's very nice. Very pleasant.

Our show today brought to you by a great sponsor. We've been with them for I think seven or eight years now. Thinkst Canary. It's a little, it looks about like a external USB drive. Plugs into the power and it has an ethernet jack and plugs into your network. And what is it? It's a honeypot. Oh my, my, this Thinkst Canary is amazing. It's a honeypot that can be almost anything.

And it is functionally, it looks, to a bad guy, it looks exactly like the real deal. Mine's a Synology NAS. And it's even the MAC address is a Synology MAC address. And the user interface is DSM-7. It's exactly, you wouldn't know it's not real.

Easy to deploy. I can in seconds make that be an IIS server or I could make it be a Linux box. I could make it a SCADA device. Seriously, you can make it almost be anything and it will always look like the real deal. It doesn't look vulnerable. It looks valuable to a bad person. So if someone is accessing your network, if they've gotten in through your perimeter defenses or it's a malicious insider snooping around,

you're going to get the things canary will immediately alert you and by the way let you know in whatever fashion you like text message email syslog it's got it supports webhooks slack you can use an api i mean just any way you want

But the thing is, when you get an alert from your ThinkScanary, pay attention. These are only the alerts that matter. I mean, they really they also can make files, lures that you spread around that look like PDF files, Excel spreadsheets, whatever you call them. And I minor name things like employee information, something subtle that would make a bad guy sit up and take notice. But the minute they try to open it, same thing.

your Thinks Canary will let you know. So you choose a profile for your Thinks Canary device, sprinkle some Thinks Canary lure files around, register that with a hosted console for monitoring and notifications, and then you sit back and wait. An attacker who breaches your network or malicious insiders cannot help but make themselves known. They see this and they go, "Yeah, bingo.

They tell their hacker buddies, we're in, right? It's not what they think it is. It's a honeypot and it's drawn them and now you got them.

I think these are the best thing. Look, everybody's got the perimeter defenses, I do too, and all of that stuff going, but you also need something inside your network to let you know if you've been breached because honestly, companies on average do not figure out they've been breached for 91 days, three months. And that means three months, a bad guy can be wandering around exfiltrating data, looking for hiding places, doing anything they want to your network.

you need to know when they're inside visit canary.tools/twit i'll give you a pricing example you know how many you need really depends on the size of your operation big banks might have hundreds of things canary spread all around a small operation just a handful like us i'll give you an example five things canaries okay that's that's a pretty good number for your network 7 500 bucks a year

you get your own hosted console you get upgrades you get support you get maintenance actually if you use the offer code twit in the how did you hear about us box you're going to get 10 off the price forever if you're at all nervous though and i understand maybe you you know you've been thinking about it but you you've heard the ads but you're not sure you should know there is a 60 day two month money back guarantee for a full refund so there's really no risk i should tell you that in the eight years

that Thinks Canary has been advertising on this show, no one has ever asked for a refund. Nothing but happy customers. You can actually go to their website, canary.tools/love and see all the love for Thinks Canaries. Some of the smartest, top names in security just love this thing. It's hard to do a honeypot on your own is non-trivial.

I know because I've talked to the guys who wrote the first honeypot, Steve Benelvin, and we had Bruce Cheswick on, and he wrote the first honeypot. It was hard. No, it's trivial. It's easy. Canary.tools.twit. Don't forget the offer code TWIT in the how did you hear about us box. We just think the world of these things, and we're really glad to be able to tell you about them. If you have a network, if you want to protect it, you need a Thinks Canary. Canary.tools.twit.

What did you buy your R1 for, Christina Warren? It's very similar to Dan. I mean, the main thing was I didn't even know about the free perplexity subscription when I bought it. I don't think that that had been announced yet. So for me, it was just like they had announced it at CES and I was like, this looks dumb, but it's teenage engineering and...

They designed the box. They weren't the company. Oh, and I knew that. I knew that. But I was just like, hey, it'll look cute next to my Playdate, which is also, you know, teenage engineering and a much more useful little device, honestly. But yeah, the Playdate's fantastic. Are they still selling it? Yeah, the Playdate? No, no. I know the Playdate. I have a Playdate. I have no idea. Yeah, I guess they are. Look, look. Yeah. Buy now.

You can ask it questions, really? Can you? You can, but I mean, you know, it works fine. It's just, I think they got a little ahead of themselves in terms of how automated the stuff would be. But I was just like, okay, whatever.

whatever we'll do it and then they were like oh and you'll get a free year perplexing i was like well that would cost as much as the device did and then it became to dan's point like funny um also as you know leo i like to collect like f'd you know company or or or or potentially f'd company things and i was like look this is going to be great this is going to come in favor someday and look what other defunct tech have you acquired lately

Well, lately, not much, but you know, what was interesting is that I, I found, and I have it around here somewhere. I don't know if I'll be able to find it or not. Um, but I had a, uh, I had a pebble. I have a number of pebbles. Awesome. They're back, baby. That's what I was going to say. That's what I was going to say. And so I was so excited about that because I was like, Oh, I have like a brand new inbox pebble steel that I got at a conference, like a monogrammed one. I have a bunch of pebbles cause, cause I, I interviewed Eric a lot back in the day and I was a big,

I was an early backer of theirs. I was a big fan of what they've done. And I was so excited when I saw the news this week, when I was on my cruise that, that pebble is coming back and that Google open sourced the, the old OS. And I was like, that's fantastic. I'm going to, yeah, actually Micah interviewed Eric, uh, for, um,

tech news weekly this week awesome and they're talking about the fact that he opens i think this is so great i don't know why more companies don't do this open sourced the pebble uh software and eric uh mijakovsky who founded it said we're gonna even make pebbles again because google i guess google which bought pebble

And incorporated some of it, I guess, into the Google Watch. Well, Fitbit bought Pebble. Google then bought Fitbit. You're right. I left out of stage. Okay. So it took from... And Mike interviewed him, so he would know more than me, but from...

based on the things I read, like it took, you know, maybe finding the right people at Google and asking, would you be willing to open source this? And they did, which I think is amazing. And now I think they're all working to see, okay, can we get these things to build? Can we get the firmware to build? Because, you know, a lot of these systems might be, you know, eight or nine years old. And, but then they can like go from there. And yeah, Eric's talking about, you know, making hardware again. And he told Mike, he said, I decided to start making new, a new pebble watch hardware and,

Hardware that will run the Pebble OS. By the way, I wish more companies did what Google did, which was, okay, we're not going to use it. Let's open source it. Let's let it free. I think that's exactly what you should do. I wish we saw that more often instead of just killing products. Mijakovsky said, we're going to do this a little bit differently this time. The name of the game is

I'm not sure what that means. I think that means not trying to go immediately from like you have your Kickstarter kind of backers and people who are interested in it to being in Best Buy stores to try to take on Apple Watch, right? Yeah. I think that what he's written before about kind of like why Pebble failed was, you know, it was this perfect kind of moment when it launched and it was a very good product. I thought the design was very good and I thought, especially for the price point, like it was a very good smartwatch and

but then you immediately had competition from like the the android um stuff and especially the apple watch and they tried just didn't seem to have that on yeah google just didn't i don't know they bought fitbit and then they incorporated bits of it but they didn't really continue on with the mission i guess that happens

Big companies buy little companies. Well, I'm glad that Eric was able to get the code back. Me too. I'm thrilled. It was an e-ink. Wasn't it an e-ink display? Yeah, it's an e-ink display. And so it had great battery life and a very, very robust developer community from the beginning who were building things around it. And they even hired some of the people who...

kind of worked on the, you know, kind of the community side to work for Pebble, the company when they were starting up because like I remember the early days when, you know, the first version came out, it was just on Kickstarter and then it broke a lot of records and

the number of people who would just create their own little watch faces and little apps and even really robust iOS apps. It always worked a little bit better on Android, but you had some really good iOS apps and integrations too. And the community of hackers around the community was just fantastic. And I'm excited to see that ethos kind of return because

I feel like there is a place for it. It doesn't have to be something that sells millions and millions of units. It can have kind of a small community of users who likes it and hopefully sustains it, which is exciting. Apparently there are a lot of people in our community who have one. Dominic says, I just took my Pebble Time out of the drawer after hearing the news. It still works great. He's in our YouTube chat. In our Discord, Grammages is still managing to keep a Pebble Steel alive. Only smartwatch I've ever liked. Yeah.

It was great. And what was funny, what's interesting is that the original idea for it started as a companion for a BlackBerry because Eric... Oh my God. It is retro. Yeah. Because Eric, well, Eric is from the university, he's Canadian. He went to the University of Waterloo and he had kind of an idea of having it be a BlackBerry companion sort of thing and kind of hacked up a kind of proof of concept, I think, with like a BlackBerry Bold. And then a few years later, they came up with the concept, put it on Kickstarter, and then it went from there.

I'm trying to remember if I had a Pebble watch. It's long gone. It feels kind of like a retro, it's like a Zippo lighter or something. Yeah, kind of. What was cool is the original one, like the steel looked a little more professional and the round had a good look, but like the original one kind of looked like a swatch watch. It was just kind of plastic.

but you could get them in different colors and, and people would even make little wraps for them and whatnot. And Eric was kind enough to give me a pink one. So I had, I had like, I backed one on Kickstarter and then I received a number of others over the years. And he gave me a pink one once, which was, which was very kind of him. I hope I have that somewhere. Have you been schlepping all this stuff all over the country?

Yeah, I'm a hoarder. What can I say? See, I give it all away and now I'm kind of regretting it. I used to remember in the old studio, I had a museum. Yeah, you had all the stuff. Yeah. And I gave it all away. This is all that's left.

A PDP-11? An ImSign? I was going to say that's an Altair, but that's an Altair, isn't it? It's an Altair. Yeah, it's a fake one, but it's a pretty good recreation. That's a real Mac, original Mac, 128K Mac, though. Oh, my gosh. I was going to say, is that the Berkeley Systems After Dark or whatever? After Dark still works. Still works.

still got the flying toasters amazing I am not going to throw these uh these these glasses away although Apple according to uh Bloomberg according to Mark German on Bloomberg has decided to can its AR glasses project which surprises me they've yeah me too is it I mean clearly they don't think the Vision Pro is the future or do they

They can't. Come on, they can't. Come on, guys. So it seems to me that Apple's, by killing their, remember, these were going to be glasses like the Orion glasses that Meta's developing that look like regular glasses, but have a heads-up display and so forth. The one that they were working on tied to a Mac, which is a little bit of a disadvantage. You're not walking down the street with it, but it's the first step.

It seems to be that if they give that up, that must be kind of tacit admission that there's not much future

With VR, AR either, right? With mixed reality at all. Oh, it could also be directly tied into or very related to their position in the AI race. I mean, those Meta Ray-Bans are only, it is only AI that makes them do the cool stuff that they do. And I mean, Meta is a leader with Lama. It's undeniable that Lama is one of the best models and Meta has a ton of exclusive data.

exclusive data that nobody else has between all of their social apps. So, I mean, if I'm Apple, I maybe sit here and I think, well, you know, the Vision Pro is a boondoggle and we don't have enough data

to at least, I mean everything, like you said Leo, Apple intelligence is running on device. That's great, but on device, on your glasses is hard. - Doesn't do much. Yeah, well these, I mean the meta glasses do tie to an app on the phone. - Yeah, well that's what I'm saying, like Apple maybe-- - And your phone is a lot of computing power and a lot of storage and a lot of RAM. I mean these things are really mini computers, super computers in some ways in your pocket.

And they're always connected. I mean, it seems to be this is a perfect match. Apple is going to announce new, well, it's not, it's not, it's Apple, but it's the, it's the, what do you call them? Headphones. Not the, they're not AirPods. Sorry. No, the other ones.

I've forgotten the name. Dr. Dre's Beats. Thank you. This week, they're going to announce an update to the Beats that will do heart rate monitoring, which the AirPods do not do yet. So that's interesting. It sounds like maybe Apple's thinking that the future is not in vision, but in the ears. Or wearables in general. But that's the thing. They're killing the glasses wearable, which surprises me, frankly. Yeah.

I mean, it's also possible that there could be two things, right? It's possible that they could be killing maybe that particular version, but there might still be dedicated to doing some sort of glasses or some sort of AR VR thing that maybe that particular implementation, they just project, they just aren't committed to, but they might be having something else, right? The fact that it was tethered to the Mac might be a deal breaker. And maybe they are looking at like ultimately having something that can

either sync with a phone or do something else wirelessly where you don't have to have that kind of accessory component. I don't know. I mean, you start by connecting it to a Mac. That's obviously not the long-term solution.

future. But it makes sense to develop it from that point forward. There's basically a little Mac in your Vision Pro. That's why they're $3,500. Right. And you have to have a battery pack. I was going to say you have to have the attachment. So I don't know. I'm just saying I could see them maybe

And they could always revisit that idea later on too. So I don't know. They've invested a lot in Vision Pro and in this AR VR thing. It would be odd for them to give up so quickly. I wouldn't necessarily say it would be a bad thing because I don't think Vision Pro is it, but it would be odd for them to give up that quickly. But at the same time, I think that it's undoubtable like AirPods,

are one of the best things that have happened. You know, I think they're one of the best products that Apple's introduced, period. And, you know, the AirPod Pro 2s, my mom and my dad are both using the hearing aid mode. Really? Do they like it? Yeah. I mean, honestly, my mom really enjoys it. My dad, I'm still having a hard time getting him to use

Did they have regular hearing aids ever or was this in lieu of a hearing aid? This was in lieu of, I've been trying to get my dad to get hearing aids for a long time and I've been, I've faced a lot of pushback on it. Yep.

My mom isn't quite at the point where she would need them, but even having a fairly decent hearing test built in just to kind of show what sort of loss there is was helpful. And she loves her AirPods and she wears them all the time anyway because she's always listening to podcasts around the house. And so having that mode on has been really great. But for my dad, we were even noticing, I got them for him for Christmas. It's like the third time I've bought him AirPods and he finally has tried them. Yeah.

For more than like 30 minutes. You just do not give up, do you? I mean, at this point, I was just like, this could really help you, Dad. And I really wanted him to get hearing aids for a long time. How deaf is he? Is it pretty clearly? I mean, it's not at the point where I think it's like, it's not like the point where they are basically like, we will be of no help to you. But it's, you know, I guess whatever the level higher than moderate is. Yeah.

So it's definitely, you know, could be, could be useful. So how about your mom? She, she, does she have, she's like, she's like, she's like moderate or, or, or light. I think it's in some, and she thinks the AirPods are good, are good enough to be hearing aids for. Yeah. Yeah. For her. I think so. Um,

I mean, I think what's great about it, I mean, honestly, is that there are a lot of people who will never take a hearing test and might not even be aware of like what they're- Oh, I agree 100%. How things can be improved and it could be a boon, right? And they're over the counter. Could you get better ones? Of course you could. But for a lot of people, it might be better than nothing. And if anything- They're not bad. I wear $7,000 Oticon hearing aids.

And that's a lot of money for people to pay. And I tried the AirPods Pro and they do pretty much the same thing. There's really one difference, which is normal hearing aids don't block your hearing. They don't seal your ears. They just sit little speaker in your ears to amplify voices. The rest of the stuff still comes in. When you're wearing the AirPods Pro, you're kind of sealed out, even with transparency mode. You know, you're relying on them for all the audio. And that's kind of a different, and I don't like it quite as much, but...

But it does the job and it's for 250 bucks instead of 7,000 bucks. Well, that's the thing, right? And then, and it already has the Bluetooth built in. I'm sure yours have your, your, Oh yeah. I can use them as all of that stuff. Yeah. Yeah. But sometimes you have to get like the, you know, attachments to your TV, the, the, the, you know, the voice coil and all that stuff. Yeah. Right. Exactly. And, and so I, I feel like for a lot of people, especially since, you know, uh, uh,

Medicaid or whatever doesn't often cover hearing aids. And so you have to pay out of pocket. And I think that having- Yeah, very little insurance coverage. You'd have to have really spectacular coverage. Medicare does not cover that. Exactly. Yeah.

Yeah, Medicare, excuse me. I always get them confused. But yeah, I mean, I think it's great to bring more of these sensors to things. What's interesting is I guess they don't have the same, I guess, patent issue with doing the heart rate and monitoring and whatnot. Oh yeah, Massimo, well, that was blood oxygen. These are the Beats. These will come out February 11th. These new Beats, Power Beats Pro 2.

and they have the same chip that's in the AirPods, but they do something the AirPods do not do. They have the W2 chip or H2 chip, sorry. But they do also have heart rate monitoring, which Apple has yet to do in the AirPods. So I'm thinking of buying a pair just to see how they...

and i like having the over the ear thing i was going to say the power beats have always been really really good um for you know anybody who wants to work out a little bit more right because right you wouldn't like i've got airpods in right now i would not want to i mean like if i were on the the treadmill like shawshan i would probably be okay with them but like i would not want to be like going shawna you're on a treadmill right now no she's not right now

Not that you know of. I was impressed. I wanted to get a little bit of cool down. I thought I could get another thousand, you know. But, you know, that would probably be okay. But like, I would be concerned if I were going for something more strenuous. Yeah. Right. Whereas the Powerbeats are great. And I guess that's why they're introducing that. It's interesting. 250 bucks. So they're the same price as the Pros, basically. Right. February 11th.

cool new colors including orange in case you want everyone to know orange impressive orange that's what Mark called it and I don't know if that's the name impressive orange although if you could color name that's a Samsung name impressive orange um yeah I just I we are almost exactly one year in on the Vision Pro and I'm just wondering if Apple's killing the AR glasses

but you know what mark says is that executives didn't want to have them attached to a mac because that's not portable but they found it drained the iphone battery if you have it attached to the iphone so they needed the mac for the processing and the and the power but that's i mean look the vision pro wasn't exactly a consumer product either you have to start somewhere if they're killing the ar glasses i wonder what their commitment to the vision pro is a year in yeah

I just think also like the glasses didn't have a great like regular person appeal. I think I might be wrong on this, but I think it was Meta that did a commercial where you have some model just like swirling around with her glasses on to show, look at me, I have glasses. She's like very like high-end wearing clothes that look terrible.

but like in the way that it's like high end and like she like translates some Spanish that says there's a party and like she goes to a freaking party but like none of that is like relatable to a person like they're not going to be going to parties in other countries that often like that's not a regular through your day thing.

And I feel like there's other ways to get people interested. But I just think all of them have done pretty crappy jobs at marketing. Because it's, again, you're not a model. You're not wearing crappy clothes. And you're not going to parties in other countries. Like, when I'm watching that, I'm like, why the hell do I need this? I don't get it. I'm sorry to interrupt you, Shoshana. That was exactly, you reminded me of that early iPod commercial that was relatability, human, like,

making the iPod look human and relatable with more than human. Cool, right? Right. Cool. It was like, could you see it like the iPod commercial for the Vision Pro? You don't look cool wearing a nerd helmet. I'm sorry.

No, you don't. Well, not only that, but like it's heavy on your face. Like my biggest concern, I didn't buy one for a lot of reasons and almost none of them were the price. I mean, the price was part of it, but it was mostly because I couldn't live with myself looking at it not being used and knowing how much I spent on it. It really didn't have anything to do with like, oh, because I'll spend and

$200 and never use it again. Fine. But like 4,000, that's a little much, but like, I didn't want it on my face because I'm like, this is going to leave marks. This is going to leave marks. This is going to like be bad for like, you know, my, my circulation. Like this is going to be a problem for wrinkles, for all kinds of things. Your posture. If it's like that heavy on your head, like, Oh my God. I mean, somebody, a scooter X in our discord is saying, well, they just killed that one product.

I don't know. I mean, that product is a stepping stone to continued AR glasses development. It's hard to tell. It's Mark Gurman. It's a rumor. Apple's not announced it, so it's hard to tell. But he says, this is a quote from the Mark Bloomberg Power on Newsletter, a company just canceled plans for a pair of AR glasses that would pair with a Mac. The ultimate goal of making standalone spectacles with augmented reality is probably at least three to five years away. So he implies that they're still working on it.

It's hard to tell. This is a leak. It's not an announcement from Apple. And Apple never talks about unreleased products. So we will never know. - Thank goodness they didn't hire a swirling model that they'll have to fire now. Like, that would feel really bad for her if she had to go like find another job where she could wear bad clothes and swirl, you know? - Key raises a larger issue, which is that Apple is really failing at AI.

and that Apple is not impressing people with Apple intelligence. They're advertising capabilities that aren't even in the phone yet that are coming. He says the AI and machine learning division, which acronym is AIML, is mocked by employees as aimless. The group, Gurman writes, has missed several deadlines and its large language models are less powerful than those of rivals.

And it is true. If you play with Apple intelligence as it stands today, it is unimpressive to say the least. You know what? It's actually really interesting to me that I don't think Apple intelligence, I don't have an iPhone, but I don't think Apple intelligence does this. And I don't think Google phones are doing this either. It's something super obvious for AI to help with scam texts. Why isn't it help identifying a scam text and other scam interactions on your phone? You're right. They're not. That's like,

such a perfect use of ai something like everyone would love i think you know and even say you know this is experimental take it with a grain of salt but like we want and you can turn this off whatever we think this is spam we think this is this is uh incorrect this yeah and you're exactly right because they already have signals of who's in your contact list and you know people you frequently communicate with even if they're not in your address book they should know presumably

the language when you report, you know, that this is incorrect, that they could train their model based on those types of things. Instead, they're doing, they're doing notification summaries that are basically useless. Here's the one from my ring. There was a person at your front door multiple times.

Oh my gosh. You know, it's wild. I saw a commercial for ring where they're like, Oh, what? It was like the committee. I don't know pop culture, but it was the two famous, like Amy Poehler and someone else. And Tina Fey. Yeah. And they're like, Oh, what was the tacos we got the other night? And they're like, we can search ring. I'm like, that's creepy. I don't want a full record of like everyone. The doorbell knows. The doorbell knows. Oh, I can zoom in. I saw that commercial.

commercial I have the same response because I saw that commercial too and they're like oh I'm gonna zoom in on what type of bag it was I'm like just go through your DoorDash history oh my gosh yeah like this isn't do you really want your doorbell to know all this

Yeah. No, no, I don't. I'm already bothered that my neighbors have Ring. And so that means that they can give my stuff to the cops anyway. Like that already bothers me, right? Like I'm already very bothered by that and I can't do anything to stop it. But like now Ring is like just fully leaning into it. Oh yeah, we are absolutely a surveillance device. Zoom in and see what brand of tacos...

Again, just open up your DoorDash history. It's easy. Their other commercials were bad, like where they're like, oh, it's what happens in your life. See, when you guys got excited over your daughter getting into college, like that was creepy enough. But I'm like, this isn't, this isn't, you have this elsewhere. Like you have a checking account, you have a credit card, you can go online, worst case, and check out where you paid for tacos. Like this isn't, like Ring is not helping you. It's just surveilling you.

No, if anything, I would be like, that would be a much better use case for the bee thing that Leo was wearing, right? Yeah. Where I would be like, well, I opted into this, and I knew that this is what this is doing. So if I could be like...

Where did I order tacos from before? Right. Like, see, to me, that's one of the advantages of this is I can say, hey, Lisa and I were talking about something the other day. What was the name of that? Or yeah, that's valuable. I think that is. But again, I think the difference is, is it's like you go into it without expectation and you go into it. I gave it permission doing this. That's right. Right. Whereas this is my doorbell.

I have a ring doorbell, I have to say. But actually, I have like two cameras on my front door. I also have a Ubiquiti camera. So I have lots of information. In fact, I just saw Amazon drop off a package. So I'll be right back. No, I'm going to stay here. Just have to think about what it might be. How exciting.

Apple did have the quarterly results that looked pretty good on the surface. They made a ridiculous amount of money. Most money ever. Yeah, most. There always is. Every quarter is the most money ever. They made, what was it? It was...

The equivalent of something like $8 billion a month. They were making $2 billion a week. Well, that's revenue. Profit, I'm sorry, profit was only $1 billion a week. Only. That's kind of mind-boggling. So it's not like they're hurting, but there were times of trouble in paradise. China, iPhone sales down 11%.

which is probably going to get worse, I would imagine, with tariffs and so forth. Actually, Apple's got to be a little bit worried. The last time, in the last Trump administration, he did announce tariffs on Chinese goods, and Tim Cook went to him and said, if you tariff the iPhone, people are just going to buy Samsung phones. You're punishing an American company and rewarding a South Korean company. Do you want to do that? Trump backed down.

But he's talking again about tariffs for everything coming out of China. He's also talking about apparently tariffs against Taiwanese chip makers, which is specifically, I think, targeted at TSMC, which makes the chips in all in the iPhone and the Mac and the iPad, the primary processing system on a chip in those devices could raise the cost of iPhones significantly. There's some headwinds for Apple, I guess.

Look, I'm not making stock recommendations. I neither buy nor sell tech stocks, and I certainly don't recommend you listen to me when it comes to stock, the stock market. Just as a business, I think Apple is, they're feeling like they're getting a little behind. Now, they've been there before and come back. Is Apple in trouble? And to find trouble. Yeah. They are a $3.5 trillion company.

So they can afford a little trouble. I mean, I don't think they're in trouble. I don't think that anyone would argue that they are leading the conversation when it comes to AI. Or virtual reality or any of the next generation of technologies, right? Sure. No, right. But I don't think anyone would argue that. But they are still the most valuable company in the world. And they still make, like you said, a billion dollars in profit every week and $80 billion in revenue. Yeah.

I think that it was interesting to see, based on conversations with people that I know,

I think that a lot of the push for Apple intelligence was relatively quick. And that's one of the reasons why it's taken a long time for some of the things to show up and that things weren't maybe finalized. And I think that they were kind of, you know, reacting almost defensively. So I think that, but Apple has been there before, as you mentioned, you know, they've definitely been seen as being, you know, behind in certain scenarios and come out okay. But,

I don't know if their own models and what they're doing with AI is... I don't know if it is going to be able to be competitive, but that doesn't mean that it couldn't be. It might not be right now, but we're at the beginning of this era. I don't think that anybody is finalized yet. I mean, we...

people were, the whole market freaked out because of R1. Apple was relatively insulated from that, which is a good thing. That's true. That's a good point. It hurt NVIDIA, it hurt Microsoft, it hurt all of the AI companies, but Apple, because they're not a player...

Didn't hurt them. They're kind of like not part of it. But also they're building their own hardware, which I think is important, right? So they also have a little bit of insulation, whereas they kind of have this cold war with NVIDIA, although they have made some conversations about working together, I guess, in some small context. But NVIDIA hardware is never going to be in Apple machines, right? And so I think that...

Like I said, I don't think they're leading, but I don't know if I would say that they're in trouble. It doesn't necessarily matter. They, you know, they're famous not for not being the first to market with anything. No, absolutely. I would be more bothered if, let me put it this way. I want to wait until we can see what they're going to kind of show off at WWDC and what statements they say in the future. If there aren't looking like they are, you know, kind of making investments in

in some ways, at the very least for some of these other models to interact on their systems, you know, if the user chooses to allow them to, then I would be bothered by some of those directions. But I don't...

i think it's too early to be let me let me correct the numbers i gave out apple revenue 124 billion dollars uh for uh the year is that wait a minute no for the quarter so that is a billion dollars a week but that's revenue profit i'm sorry it was only 36 billion dollars so that's a billion a week so it's still a lot of money

46% margin is a pretty good margin on your business.

Let's go back to that iPod example. Think of those dancing commercials. The iPod was iconic because it was simple. It was easy to understand. The same with the iPhone. Even the iPad, people mocked, but it was pretty easy to understand. When we get into VR and AR territory, we've said this a million times, but who's the market? What is this? Why? Who cares? Right.

I don't mean that in a flippant way, but I mean, who cares? If I were in a meeting, I would say like, who cares? Who is the market for this? And what is the plan? What is the goal? And it's not like you can't run all of the AI models on an iPhone. I mean, I have chat GPT. I have Claude. I have everything on here. I mean, that's why they're saying the M4 is so great because local processing, right? Yep.

Yeah, but I'm not running it locally. When I run DeepSeek, it's going to China. It's going straight to China. Right, but you could do some of those things. No, but I think that's a great point, Dan. Like who are these things for and how is it being positioned? And I don't think they nailed that. Whereas the other things, they had a very clear target. With AI, I feel like,

because the first what's but come out has just hasn't been very polished and hasn't had a great experience people have kind of a negative you know line through it and and siri already has a bad reputation deservedly so oh she's just gotten stupid how did they let that happen i mean well i'm tell i'm my it's my attitude that it's because apple's decided that they want to be the privacy company and that as long as you are unwilling to uh you know release that information in the world

You're never going to have AIs that's as smart. So they've hobbled themselves by deciding to protect the privacy of their users. Now, users will get to decide which they want. Do they want, you know, are they on the one hand going to be putting a device that listens to everything and sends it to China? Or on the other hand, are they going to keep it all on their phone and keep it private and have less intelligence? That's where...

That's a choice. Right. And I think that's exactly where Apple is positioning themselves right now, which is really interesting in the middle. You can make that choice and you will continue to make that choice. They're pushing the M4 and you can process locally or they integrate with open AI. Anyway, I'm just saying it's an interesting position for them. It's new. It is. And I don't think it's, I mean, it certainly hasn't hurt them yet. Yeah.

I think it's kind of like in the do everything social media era where Snapchat is like, hey, politicians, come onto our platform and reach kids. And it's like, what?

No, no one is on Snapchat to like check on Hillary Clinton, make sure she's chill. Like that wasn't a thing. And I think you're getting that with like some AI stuff that it's like, like there was a meme of a place and I think it was a real place called, but what was it? Like coffee GPT or cafe GPT or something like that. It was like ridiculous, but we're kind of, we're going to be in that era for a little bit of like,

AI for like your hair or like just the trash yeah it's just really stupid stuff and I think like I think some companies are flailing because this is a understandably like a bigger buy-in because it's like social media has wide applications but AI is even far wider applications for all kinds of things so like they want to be in it but like they don't know why it's like a

It's like how Pokemon was so popular and parents didn't understand it. It's like a trend that they can't get their hands around for the trend part, but they're forgetting the trend's not why you sell your stuff. The purpose of how it's used for people is why you sell yourself. iPod was a thousand songs in your pocket, and this is like AI assistant in your pocket. Yeah. One is cool. The other is like AI. Yeah.

But if you said all of human knowledge in your pocket. Yeah. Okay, there you go. Right? All right, much better. You win. Fine. Yeah. An expert in your pocket. Ask them anything and they know and they can answer. Wait, no. A thousand songs is still cool.

Less useful, but cooler. It was cooler. Well, it was cool, but I really do think if you sold it as all of human knowledge in your pocket or get information what you want, I think that that would be cool. I still think the music is cooler, but you could...

But I think that's how you'd have to position it. And instead, it's kind of like this weird thing like, oh, we're going to help you with your writing, but it doesn't really. And we're going to summarize stuff, but in kind of an annoying way. And the summaries take up more space than just the subject line would on the emails. And I'm like, how do I turn this off?

But I don't want to turn it off because I need to know how all this stuff works. And yeah, yeah. I have to, unfortunately, I take a break. I don't want to interrupt because it's a great conversation. Hold that thought, put a pin in your head, and we will continue in just a moment. You're watching This Week in Tech, our weekly tech news roundtable.

You're going to hear a lot of AI in all of our shows, of course, but we have now an AI show starting Wednesday. This week in Google, mutates into intelligent machines. Jeff Jarvis, Paris Martineau, and I will talk about AI with AI experts, and it's going to be a lot of fun. So I hope we'll see you on Wednesday. This week's tweet is brought to you by ZipRecruiter. According to research, a major challenge many employers face is

is the pressure to hire quickly. It's so funny because when you put it that way, you know, it's very sterile. But I can tell you as an owner of a small business, when somebody leaves your company and it's a small company, that's a crisis because you have now two weeks to find somebody to replace them or you're going to have to do their job. It is, and while you're trying to do their job, you're scrambling to hire, it's very difficult.

It's a tough hurdle to overcome because it's so time consuming to search for great candidates to search through applications. That's why when people leave our employee, and it doesn't happen very often, but you know, sometimes somebody says, well, I got a job closer to home or whatever. We go to ZipRecruiter. If you're an employer, have you ever tried ZipRecruiter? It is incredible. They have figured out

how to solve that crisis where you've got to fill that position. In fact, four out of five employers who post on ZipRecruiter get a quality candidate within the first day. It's been our experience that it's within just a few hours. We'll post, you know, Lisa will post the job at breakfast.

Now what happens immediately? A couple of things that make ZipRecruiter great. First of all, that post goes everywhere. More than 100 job boards automatically, social everywhere. So you're casting the widest net, which means you're going to get a lot of applicants.

And by the way, they don't call you and they don't go to your inbox. They go to the ZipRecruiter interface where they actually reformat resumes so you can scan them quickly. They have screening questions. They have a lot of tools to make it very quick and easy to rank. But then something else happens that's really cool.

Zip Recruiter, people come looking for work to Zip Recruiter, so they have more than a million current resumes on file. They use AI then to look at those resumes, look at your requirements, and make matches. And they will actually, within an hour or two, give you a list of people who are looking for work who fit your requirements. You can then rank them, rate them, and invite the best ones to apply. Now, that does a couple of things. First of all,

and you probably know this if you've ever looked for a job if a company comes to you and says hey we like you we would like you to come to work for us you go right to the top of the pile right i'm going to apply to that company because they every wants to be liked so in a competitive environment it is very competitive now to hire the best people you're going to go right to the top you're also going to get great candidates fast by lunch

Lisa's going, "Oh, hey, here's a great one." I mean, we're worried, right? How are we going to fill this position? "Oh, here's a great one. Oh, here's another great one." We would literally by noon have two or three really excellent candidates thanks to ZipRecruiter. Right now you could try it for free too.

Give it a shot. ZipRecruiter.com slash twit. G2 says, ZipRecruiter is the hiring site employers prefer the most. And I would agree. That's what we use. How fast does ZipRecruiter smart technology start showing your job to qualified candidates right away?

to the largest pool as possible. That way you're most likely to reach the person, the one perfect person and fill that job fast. So take it easy, relax employers. Let ZipRecruiter take a load off, speed up your hiring. See for yourself, go to ziprecruiter.com/tryitforfree. That's the same price as a genuine smile from a stranger.

or a picture-perfect sunset or a cute dog running up to you and licking your hand, ziprecruiter.com slash twit. ZipRecruiter is the smartest way to hire. And I know from personal experience, ziprecruiter.com slash twit. We thank them so much for supporting This Week in Tech.

uh surveys over right Benito we this is it they stopped it on Friday so thank you for everybody who submitted uh surveys we appreciate it there's still a way you can help us and that's by joining the club uh we have advertisers we have great advertisers but they don't cover the full cost of what we do

For that, we need your support. And I am very pleased to say we have 12,000 club members. It really is great. We keep it affordable, seven bucks a month. You get ad-free versions of the shows. You don't have to listen to ads. Even this pitch, you don't have to hear if you don't want to.

Although I know a lot of people still listen to the ad-supported versions for some reason. You also get special events that we do. You get access to our Club Twit Discord with really smart, fun, interesting people talking about all the nerd topics. I think it's a great club to be part of, and it really helps us keep the lights on, pay our staff. It is not cheap to do what we do.

And I feel pretty strongly, I mean, I'm ready to retire. I have saved up. I could retire. But I feel like more than ever, the next few years are going to see some remarkable developments in tech. And I think it's really important that we as consumers, as users, as people impacted by tech understand it. And that's what I think our job is.

And I want to do that job as best we possibly can. So please join the club and help us do that. twit.tv slash club twit. And I thank you in advance. And I thank all of our fantastic club members. Is Taylor Swift going to win a Grammy tonight, by the way? I don't think so. No. She doesn't have a new album.

Well, no, she did have a new album, The Torture of Poets Department. It was a good album, but the problem is she won last year for a not great album, in my opinion, which she won it for the Eros Tour. And the album she did release, there's a lot of competition. And she already won last year. Who's got record of the year and album of the year? Who's going to win? Who are the big winners? I don't know any of these people's names.

I mean, there's a big push for Beyonce to win one of the big ones, but I don't know if that's going to happen. I...

i think chaperone had a really good year she had a great year she had a great year that would be kind of my my pick for some of the for some of the stuff but um well we're going to get this show over before the grammy starts so you can you can watch we are missing though apparently uh right now uh open ai is doing a stream of some kind

that they announced at the last minute. I don't know what they're announcing, but if anybody's watching that at the same time as you're watching the show, you can, you can fill us in. Um,

They're not streaming it on X. We are. We got 400 people watching us on X. Hey, hey, everybody. We stream. Did I mention? I didn't mention this. We stream in eight different venues now as we do the show live. The show's never been really, it's not a TV show. We don't expect you to watch live, but if you want to, you can. 2 p.m. Pacific on Sundays on YouTube, Twitch, X.com, Kik, LinkedIn, Facebook, Twitter.

TikTok, we're on TikTok. And I'm missing one, there's eight of them. What else am I missing? Twitch, YouTube, anyway. You can watch us live. We are live, as I said, 400 people watching us on Twitter, but I don't see the OpenAI thing. So I don't know where they're doing that stream. Let me see if it's on their website. They've been, yeah, there it is. They've been really active in the last, oh, Deep Research is introducing Deep Research, okay.

So isn't that what deep seek is? Is deep research that kind of reasoning? Yeah. All right. Well, we'll keep you filled in and we'll talk about it on Wednesday, won't we, kids? Oh, what was I going to... I had something for you that I was going to ask you about, but now I've... Oh, here it is. We were talking about X. Elon's done a deal with Visa. They're going to have X money.

Elon's always said he wanted to turn X into the everything app, right? And the first thing would be financial services. He said this week on Tuesday that he struck a deal with Visa to be the first partner for what they're calling the X money account. Visa will let X users move funds between traditional bank accounts and their digital wallet, make instant peer-to-peer payments, kind of like Venmo or Zelle.

Do you trust Elon and X to make your financial worries go away? I kind of trust Visa. I trust Visa. I was surprised that Visa said yes. I don't think they're going to mess it up. The world has changed. No, it's true. Since January 20th.

Visa is probably going to keep stuff like in some, I don't, I'm not that good at this kind of coding at all. I can like HTML it a little bit, but I'm sure they're going to keep some stuff in a container away from like where it can be harmed because they know that that would really mess with their brand reputation. Plus like, I forgot which freaking agency was going after them. And it was for something totally stupid, but they still. He's suing all the other advertisers. Yeah.

For boycotting him. I don't know if you can sue him. I don't think I could sue people who don't advertise on Twit. Could I? I mean, maybe you just aren't trying hard enough, Leo. I don't even understand the rationale. How do you sue people for not buying ads?

I guess the argument is that they have been pushing others not to advertise with them, right? So it might not be enough that like Nestle and Lego and whoever being like, we don't want to give you money, but by being part of some consortium, they're encouraging others not to advertise. I still feel like that's a speech issue. Yeah. I don't know. I still feel like that's a speech issue. Can you believe he's suing? I can understand suing Nestle, but suing Lego? That's like, I don't know. That's like suing mom or apple pie.

Of course, they are Danish. Maybe it's part of the overall anti-Danish movement in the United States. That makes sense. The suit targeted the World Federation of Advertisers, which immediately dissolved, by the way. I respect that. I do. Like, find us. I've actually seen that in other cases, in residential cases where there's an entity, like an HOA or whatever, that's like, bye! We don't exist. Sue us.

He started this in federal court last year. In CVS, he sued Twitch. Early on Saturday, they amended the complaint yesterday to include Lego, Nestle, Tyson Foods, the chicken maker. Well, they don't make chickens. God makes chickens, but Tyson slaughters them and brings them to your door. Abbott Laboratories, Colgate Palmolive, Pinterest, and Shell.

They're going after everybody. I mean, like Pinterest is such a weird one. Like Shell, okay, want to really get into the war with the oil companies. All right. I think Elon's feeling a little empowered right now. It's become apparently clear that Elon has moved in to the Office of Personnel Management, that the letter, the email that was sent out to government employees saying, you could, you could, you know,

uh quit in september we'll pay you through then you don't have to do anything was at was actually elon it was almost identical to the letter that he sent to ex-employees uh that it was from elon's people um i feel i think he feels empowered at this point not sure that uh uh that's a good thing he's empowered i think he's empowered i i don't i i yeah i i think he's got some he's got the president on his side he does he does now how long this bromance will last before

the egos i thought they would have broken up by now but i have we did have a pool and uh i went back and checked in my i said by june so i have a few more months you have a few more months you've six more months you you've got some time but like i i don't think he feels empowered i mean i think that we should just call a spade a spade i think he is empowered he is yeah he is he's the uh

uh first uh first friend or something I just want to point out that Goldman Sachs had claimed is trying to get desperately out of the apple card deal they made a deal of course to they they're the bank behind the apple card they say they've lost a billion dollars I don't know how you lose a billion dollars on a credit card yeah I don't that's it takes a unique ability

Did they not have a credit limit? Like, were they just like spending on that card? Cause that's the only way. I don't understand how you lose a billion dollars with a credit card, but they did. In fact, well, Goldman had never done consumer banking. That was the problem. That was the problem. And, and then from what I understand, like American express, I mean, this is from reporting the wall street journal and others have done over the past few years. The,

the terms that Apple wanted were very advantageous for Apple, not necessarily advantageous for Goldman Sachs and Goldman Sachs, who had never done consumer before and has now exited consumer entirely except for the Apple car deal, which they desperately want to get out of. The problem is, is that like if you go to Chase or if you go to, you know, Citi or you go to American Express, which is the obvious partner for Apple, they are not going to give Apple the same terms that Goldman Sachs did. Yeah, I'd like to point out the layers of irony in that.

that they didn't understand the disfavorable terms of doing a deal with a larger and more powerful entity. Shocking. I wonder what the deal is with Visa and X, you know?

You got to think X really wanted this. But on the other hand, maybe because of the political climate, it was Visa said, you know, probably wouldn't be a bad thing. I've been working on that for like over a year. So there have been reportings about that. So I don't know. I mean, and who knows what the terms are there? I mean, I don't know how demonstrably different it is than like when Jack Dorsey owned Twitter. And, you know, like there was kind of a square integration and then square, obviously, you know, has some ways.

Well, now the good news is Elon can debank people. He has some more power if you bank with X. Are you really debanked if you just have like a visa pass through? Yeah, really. I mean, oh, I can't use X to send my buddy money. Oh, well. Oh, well. Linda Iaccarino, the CEO of Twitter.

xing what is she posting she's eating another milestone for the everything app Visa is our first partner for the x money account which will debut later this year um first of many big announcements about x money this year lfg which is uh code for let's freaking get it x money does have an account it has 192 000 followers for all your money moves powered by x

I'm so tired. All right. I don't, you know, it's hard. I don't want to get into politics. And yet Elon is in there is in, you know, is a tech mogul and he's certainly embedded in the government at this point. Doge, the department of governmental efficiency. We had the story last week that was really just taking over for USDS, which is United States digital service.

which was uh started by Silicon Valley people including our friend Matt cuts who ran it for a while he was the director to uh help the government after the failure of uh the ACA website to help the government modernize its technology goodness knows that's a good thing but now I'm thinking it's more than just a doge thousands of U.S government web pages have been taken down

Since Friday, 8,000, according to the New York Times, across more than a dozen U.S. government websites have gone down, disappeared. In fact, you can go to GitHub and watch them disappear in some cases. I think it's just confusing for people here, too. It's just like, even if you don't like the stuff said on it, it's confusing, just in general. I think that's the point.

Yeah, and I think one thing, too, though, is I was surprised that Vivek lasted less long than Elon at Doge. I would not have bet money on that at all. Agreed. That's like this whole weird dynamic there. Oh, I totally would have. You could see this 10 miles away. He had no power whatsoever.

yeah i think elon in the long run because he's the richest man in the world friends yeah they wanted him out like they never wanted it look i mean just like the inside baseball dc they didn't want him for two seconds yeah get how he i know you're in dc i don't mean to talk down to you no no it was a weird one because i didn't get how he got there in the first place and then all right i guess he's there so and then when he went away it's cronyism yeah but cronyism only goes so far right

So somebody's saying, oh, all the pages that were removed were DEI pages. Well, more than 3,000 pages from the Centers for Disease Control, including 1,000 research articles filed under preventing chronic disease, STD treatment, guidelines about Alzheimer's warning signs, overdose prevention training, vaccine guidelines for pregnant people.

The Times hypothesizes that perhaps the words pregnant people could have contributed to its removal. Oh, is that bad? Really? Really? I mean, maybe they just did like a find and delete and then just it went broader. Like, that's kind of like what this is striking me. It feels a little overbroad, perhaps.

Yeah, I know it is not uncommon when new administrations take over that like the old website will go down and that you might have 404s and whatnot. I think it's just the breadth of the number of pages, like because typically what happens, I mean, I think what the Biden administration did was that they would create like old.whitehouse.gov and then that would

you know, live on for government transparency reasons, which is a good thing. And instead, it seems like you were now just left with a bunch of 404s and no real idea if any of these pages will return, you know, whether pregnant people has been replaced with the language that they prefer or not. I don't know. I mean, I can go on and on. It's read the New York Times article. It's 8000 pages. Maybe it's an accident. I don't know.

The Musk aides have locked workers out of the Office of Personnel Management computer systems. Musk's aides, some of whom are as young as 19 years old, have access to the database which contains personal information for millions of federal employees, including social security numbers, dates of birth, home addresses, pay grades, length of service,

This is a quote from Reuters who talked to one of the OPM officials. They talked to two OPM officials who spoke to Reuters on a condition of anonymity, said some career employees at OPM have had their access removed. We have no visibility into what they're doing with the computer and data systems. That's creating great concern. There's no oversight. It creates real cybersecurity and hacking implications. Remember, OPM was hacked. And millions of government records were released some time ago.

Some years ago. And the government's just hacked all the time. When I did my age verification thing, I'm like, hey, guys, careful. Do you really want to give your personal information to the government? Yeah. Yeah. Because it can't manage it. It is hacked all the time. All the time. So much. Yeah. So in that regard, I would love it if Doge came in with some smart technology people and helped them lock the systems down and so forth.

I have learned that CISA employees are exempt from the request to resign. So that's good news. Oh, interesting. Yeah. I wouldn't have expected that because he's had a lot of issues with CISA and the stuff they've been doing. That actually really surprises me. Yeah. Well...

It's hard to get any information about any of this stuff, by the way. This comes to me from a friend who works for the DOD who said, yes, CISA was exempt. Many government officials are locked out of the Office of Personnel Management. Meanwhile, Doge employees who were basically not elected, just chosen by Musk, have full access. This is from...

yahoo tech from mashable i guess uh it's in yahoo tech uh that uh musk employees have embedded themselves deep within the agency since january 20th doge has more from a non-government advisory panel to the rebranded tech unit inside the white house according to the report the team set up sofa beds in the opm director's office to work around the clock

securing access to the enterprise human resources integration database that's got every why he likes sleeping in the office so much like i just loves that it's a thing he likes i just don't get it i'm like it sucks musk's jet has not left dc since inauguration he's he's there he's embedded um

The highest ranking career official at the Treasury Department on Friday resigned after refusing to grant Musk's operatives access to the government's entire payment system, responsible for $6 trillion in payroll annually. So behind the scenes, there is this move into government with access to these systems

Mashable rights must could bypass legal hurdles and potentially shut off funding for social programs opposed by him and Trump further expanding his grip on federal operations. So I guess, you know, this is how you cut $2 billion from the federal budget more than I mean, the pages, the number of pages, the variety of pages.

They even pulled down a video entitled, "Here's how to avoid IRS penalties and interest." And the form private schools have to submit to certify they've not engaged in all racially discriminating behavior. There you go. Two dozen research notifications from the Nuclear Regulatory Commission. But they were all just from the year 2000, so. It seems kind of random. I guess there's nothing more to say. It's just happening, right? We're watching it happen in real time.

Google has decided to rename the Gulf of Mexico to the Gulf of America. It's also going to rename Denali to Mount McKinley, its old name. These are two things the president requested in an executive order. And, of course, the executive order changes this with the, what is it? It's called the Commission on Geography. Let me find the name. But there is a official United States Geographic Commission, and they name the names of

But Google says it's only going to change the name for people in the United States looking at the maps. If you're in Mexico, it'll still say the Gulf of Mexico. Google says we have a longstanding practice of applying name changes when they've been updated in official government sources. So the Geographic Names Information System, or GENUS, has changed it to the Gulf of America. So Google says, okay, that's fine. Then we'll change our maps. Apple has not yet changed its map. Good for them.

I mean, you can name it whatever you want, I guess. I mean...

i'm used to the gulf of mexico i don't think it's an insult to the united states as long as you put in parentheses right so we all know like what we're searching for yeah it shows how inter how how interrelated our chains of information and services are um and whether google wants to or doesn't want to make this chain if they're saying it is because of a policy change

that they rely on, it just demonstrates that one policy change upstream can have some pretty significant effects. By the way, somebody in our YouTube says, I love the guy in the attic critiquing the wealthiest man in the world. Yeah, I should. I got nothing to say. He's the wealthiest man. He must be the best man in the world if he's the wealthiest, right?

Yeah, I think that's how it works. That's how it works. Of course, we reward him for being the best. Right. It wasn't a Nazi. So I, for years, have of course gotten hate mail from people on the right saying you're just a libtard and stuff like that, which is fine. I understand that. I am now getting hate mail from people on the left saying, Oh, that's funny. Why didn't you call out Elon's Nazi salute?

You can't, I don't know. I know it's not, I don't, you know, whatever. It's always been the same for me for years, maybe at least for the last 10 years, people always say I'm a right wing nut job or I'm super liberal. I get both. And I'm just like, you get both. Yeah. Yeah. I'm like, whoever's being stupid, I'll say they're being stupid. We might disagree on what the stupid is, but like, I know some stuff. Our street is typically thought of as a conservative thing, not a right wing, but conservative think tank.

Yeah, more center right than anything. Center right, yeah. Yeah, we do right stuff. We do center stuff, whatever makes sense. But we're about being free market. So markets guide our thinking. Like wherever, you know, when government does stuff where it doesn't create the right incentives, we just remind people like, hey, markets are good incentives. And when you can create markets for something and encourage competition, it usually works out better, stuff like that.

you you write about age verification you actually wrote the article a couple of years ago about why age verification systems are inevitably a bad idea any update on that because we're starting to see it certainly the uk now requires it oh my gosh there's a there's a new model for app store age verification that the porn companies are like cool with because they don't have to do it there's no porn in app stores and i'm like so all all these

All these companies are like, oh, you know what? If you want to stop kids from watching porn, verify an app store. There is no Pornhub app. Like, it's ridiculous. They just want to pass the buck, right? They say, well, you know, it's the app store's job. They do make the point the app store might have a better idea of what your age is.

Yeah, but it all comes down to the same problems. Because when you say, you know, you have to, these bills all say you have to have parental consent to download any app. Well, social security numbers don't even tie you together that way. Like you would probably need like a birth certificate to prove the parent's the parent and the child's the child. You know, parents don't always have the same last name as their kids. There's all different issues. And like, if it's to mean anything, it would

probably would have to be birth certificates. Otherwise your 19 year old sister could say, "Oh, I'm the parent." And like you at 16 get to do whatever you want and have that verified. And the worst part is these bills like don't serve a purpose.

already do this with your phones if you want. Say if you're getting, come here and like connect your phones and then verify and then say without putting in your most sensitive information and then say, okay, I have to approve everything you do. But instead of doing that, we're forcing the government to force like privacy violation and first amendment issues to do that. And I'm like, Hey guys, still not winning it. Still some issues here. You're going to end up back in court. And like, they just don't want to hear it. Yeah.

I'm in favor of giving parents control. The parent knows not the age of the kid, but the emotional age, the capabilities of the kid. Who better than a parent to decide what a kid should and should not see? Oh, yeah. So put parental controls in. That's appropriate.

They already have them, though. They just don't know how to use them and haven't. And I get that some of it gets a little bit confusing, but what these bills are requiring is already there. And parents just are sometimes using them and sometimes not. And then these bills just reinvent the wheel with security and First Amendment issues. Right.

it but it's it's the do something and like utah has one of these bills and i'm like hey buddy like i love i love the sponsor he's a really great guy but i'm like dude come on talk to me first sometimes like let me walk you through all this talk about mike lee no different one uh uh i i've i also uh made clear to other offices but with uh todd weiler in utah he's fun he's a bit of a firebrand but he really does try to solve problems

And he keeps doing porn age verification bills over social media ones. But like, I don't think he understands like the problems he's running into with the way he's doing it. And I just wish he would talk to me.

I mean, I imagine you get to brief members of Congress. Yes? Oh, yeah. Sometimes. Their staff, too. Some members. Or talk to the staff. Yeah. The staff is kind of who talks to the member, right? And you talk to the staff, generally. Yeah. Sometimes the members, too, depending on if they vibe. There are some members who vibe really well with me, but then other members are like staff. Right. Right.

I hope they listen. I mean, you know, Christina was earlier complaining about the lack of, you know, technical sophistication and members of Congress there, but they're smart people. They didn't get there by being morons. Well, except for Lauren Boebert, but they had to, I had to say, I had to do that, but sorry. Generally, I think they're intelligent and they just, they need staff to explain stuff and they, and, and, and need to listen and understand the technical issues and,

Sometimes it's easier politically just to say, well, I know there are technical issues, but we're not going to address. I think that's mostly what's happening, isn't it? It's that, and it's also, I'll give staff this, like they have to cover so many issues. Like no matter where you are, you just have to cover so many kinds of things, and it's impossible to know them all well, which is part of the reason overall I want government to do less so it can focus on the things it needs to do. There's other problems with like, and this isn't to like get on a soapbox

about it, but it's interesting. I've been running into this a lot lately. One of the problems with having government do a lot of things is it never budgets enough for enforcement because it's impossible to enforce everything. And we need to pick and choose what we want to enforce wisely to make sure the right stuff's enforced. And like laws against hair braiding are like not a priority, you know, for government that they can selectively enforce.

But in Congress, it's a lot of showboating. It's a lot of like trying to do something, even if it's not the exact right thing. Just hating tech companies was and still is a thing a bit. But there's just a lot of bad incentives here and including the like staff not having, you know, all the time they need to really dive into these issues. Well, yeah.

rstreet.org no conscript conscripting the app stores doesn't solve the problems with age verification josh withrow and our own shoshana weissman thank you shoshana let's take a little break more to come we're gonna wrap this up pretty quick we're getting to a grammy uh grammy award time how soon before the grammys start 15 minutes oh my god

Last time you were on, I let you go early so you could go see Tay-Tay's last show. Was it amazing? It was amazing. It was so fun. I had the best time. I'm so glad you got to see that. The last show in the ERA's tour. Wow. Was it tearful at the end? I mean, it was kind of like, I was still so jet lagged from being in South Korea like 20 hours earlier that it was all blur. But no, it was really amazing. And it was so cool to be there with so many, you know,

Real fans. I mean, anybody who's there is a serious fan, right? That's a very special ticket. It was really cool. I'm glad that we could do that. You don't mind missing the opening act? Not at all. Not at all. We're good. Taylor won't be on for the first hour. You know that. No, no, no. Her red carpet look just debuted. She is in a short red dress. Looks really good. Is it sparkly? A little bit, but it's very different from what she usually wears. It looks good.

It looks good. All right. I'm going to give everybody a moment to go search your socials for Taylor Swift's red carpet dress. Well, I tell you about our sponsor for this segment of twit. Delete me. Now, this is something you might want to know about. Absent a comprehensive privacy law in the U.S. government, we're going to need to rely on ourselves to protect yourself from people

getting your stuff online. If you've ever, have you ever searched for your name online? Please don't. It's really kind of shocking the amount of personal data that is easily available to anybody who searches online and then for a few bucks they can get even more personal. But here's the good news. The law does require data brokers to delete your record if you ask them.

You know where all thousand data brokers live and you know about the new ones that spring up every day? No, but delete me does. What I love about it and browse around because they have personal plans. They have family plans, which is really cool. And they also have corporate plans. And I would submit that this is a must for any business. We learned that lesson the hard way. Our CEO,

was impersonated in text messages to her direct reports saying, "I'm in a meeting right now. Could you buy 20 Amazon gift cards and email them off to this address?"

Now, fortunately, our employees are smart, but it opened my eyes because this was a very targeted spear phishing attack. They knew Lisa's name and they knew her phone number. They were able to impersonate it. They knew her direct reports. They knew their phone numbers. They knew their name. They had so much information. Where did they get it? Online. We immediately signed up for Delete Me to protect Lisa. You might want to do it for your family. With the Delete Me family plan, you can ensure everyone in the family feels safe online.

Delete me works not just to delete that stuff but to protect you from identity theft from cyber security threats from harassment Spear phishing attacking and more it really works in fact when the national public data broker breach happened Steve Gibson and I both searched and found our social security numbers in the Along with hundreds of millions of other people in the database we searched for leases information It wasn't there and I realized oh

it was deleted delete me's experts will find and remove your information from hundreds of data brokers you can assign if you're doing the family plan or a corporate plan you can sign a unique data sheet to each family member tailored to them easy to use controls mean account owners can manage privacy settings for the whole family the whole company now here's the best part delete me will yes delete it initially but then they will continue to scan and remove your information

regularly, even from new data brokers who spring up and they spring up all the time because this is an insanely profitable business. DeleteMe will remove addresses, photos, emails, relatives, phone numbers, social media, property value and more

They'll scrub it, they'll clean it, and they'll keep it clean. And it really works. I know it does because we do it. Protect yourself, reclaim your privacy by visiting joindeleteme.com slash twit. The offer code is twit. That's joindeleteme.com slash twit. And do use the offer code twit. You'll get 20% off all privacy plans at checkout. That's a good deal. joindeleteme.com slash twit. It does work. I know for a fact.

It was kind of a shocker when I found my social and Steve's social online. And Lisa, who? We don't know. We don't know that person. On we go with a stunning, exciting, thrilling, gripping edition of This Week in Tech. Let's see. Comcast is embracing a new technology from the IETF, the Internet Engineering Task Force, the L4S standard.

It's a little geeky but this is good news. Comcast is the biggest ISP in the United States and it's expected that Comcast adopting it will encourage other ISPs to adopt it. It's ultra low lag, particularly useful for what we're doing right now, Zoom calls, gaming. And Gadget describes it this way: "If a packet traveling between your device and the server experiences congestion, it will report that on arrival

which can improve future packets journeys. Sounds very cooperative.

Products from Apple, Meta, Nvidia and Valve are the first to support the technology. They were initial partners for testing the low latency connectivity. You can see why gaming companies would do this. - Yeah, makes a lot of sense. - Yeah. Comcast, other developers can choose to take advantage of the open standard once Comcast has fully rolled it out. It will be available to Xfinity, all of Xfinity customers at that point. Right now it's a pilot program in Atlanta, Chicago.

Colorado Springs Philadelphia Rockville Maryland and San Francisco I hope I counted San Francisco we're about an hour North um it's Comcast your ISP or you with someone else they are where they are uh where I have a Comcast business account that's what I'm talking to you now and I have a fallback with Starlink so there's a because Comcast comes and goes sometimes yeah sometimes yeah and uh and when it does in 10 seconds the ubiquity stuff will go oh there's no internet and it'll go up to the satellite and I'll come back

uh and that's worked pretty well with that's we've used that a couple of times um comcast is also rolling out doxus 4 in the next probably this year at least in our area which will greatly improve upstream bandwidths as well so while i am not a fan of comcast i wouldn't qualify as a fan i mean i don't think anyone would qualify it's nice to hear that they're adopting the new technology that's great

No, that is cool. It's good to see that they're pushing things forward. And it's probably because they face competition from various fiber providers increasingly. But of course, that's the thing about the United States that's hard, I think, for people who don't live here to understand. Very few of us get to choose who our Internet service provider is. Wow.

that's cool that's i'm sorry i was looking at taylor's swiss red dress yes it's good right it's good wow and she's got a little t uh on her uh uh whatever what do you call jewelry that suspends from the bottom of your dress well i know that's i don't know but i don't i have no idea what you would call that but it's it's it's thigh jewelry yeah i guess so the t is for thigh not taylor

Or Travis, I guess. Oh! I think that's the thought, that it's Travis. It's Travis, not Tay-Tay. It's Chiefs Red. Wow. Oh, it is Chiefs Red. Is she Christian? I don't know. It looks like a cross. Yeah, I was going to say. She's worn crosses before. Yeah, I think she's. Maybe it's all three. It's like all. It's all three. Holy Trinity. Travis, Taylor, and Tate.

And the big J. Yeah. Yeah. See right there. It says page six says Taylor Swift wears Chiefs red on the Grammy carpet. All right. Ahead of the Super Bowl. Yeah. Her boy's going to her boy's going to be in the Super Bowl. It's so funny because I follow the NFL on Reddit and they post how much airtime Taylor gets compared to Travis. Yeah.

It's been going down, though. I think that the broadcasters have kind of gotten over that. No, I think last year, I think it was a much bigger deal. And I think this year, they're like, okay, yeah, she's another wag. But yeah. It's good for the NFL. I got to say, it's good for the NFL. Yeah. Well, and look.

I have to say, usually her clothing choices, I'm not a fan of. Really? And I can't blame the stylist. I think at a certain point we have to go, it's not the stylist's fault. It's Taylor's fault? You think she has bad taste? I mean...

I mean, look, we're not, we're not, no one is perfect, but I will say that her street style looks at, at his games have been so good. I thought so. I'm like, God, why can't you dress like this when you go to award shows sometimes? And, and, and I haven't been able to look like the little top hat, uh, circus, uh, ringmaster outfit. That's cute.

Does she wear that? That's really old. It's like old style circuit. I'm here for that. I like 30s, 40s influences. I'm the only one. Oh, I like that too. No, no. I think sometimes it can look good, but sometimes when I look at her Grammy look from last year, which was not good at all. That was a weird look. She had this black arms. And the stuff she wears to the VMAs is really...

Yeah. She's mixed, but the street style has been very good, especially with the football game. Here is, thanks to Vogue, every Taylor Swift. I searched for Taylor Swift NFL fits and found this article. All of the stuff. That's cute. Love that. That's very cute. Yeah, that's good. By the way, that's her mom. That's so cute. That's Travis's mom. That's Travis's mom. That's Travis's mom. Looks just like Taylor's mom. Okay. They look similar. Yeah. Yeah.

And who's that? I don't know. We don't know. Some bodyguard. Some bodyguard. Okay. That's interesting. Plaid sundress. Interesting combo. The big t-shirt. Always a good look at an NFL game. Does she ever paint herself red and white? No. I don't know. I got distracted. How many of you have given your spit to 23andMe?

No. Different company, but I've, yeah, I've spread my... You've donated your spit? Yeah, I've spread my spit around. I have a lot of medical tests, so any bodily fluid I have has gone somewhere. See, but you have a legitimate reason. Yeah. That's, you know, for your health. I just thought it'd be fun to know if I have any Irish in me, so

So, but now apparently your spit is for sale. Yeah. See, this is my fear with all of this. This is why I've never did 23andMe or Ancestry or those. And I'm very curious. Well, it doesn't really matter because my relatives did. So, you know, they can create genomes based on me anyway. Thanks, Uncle George. Really appreciate it. Yeah.

But yeah, 23andMe is about to go bankrupt, right? Isn't that the thing? Yeah, they're in trouble and they are seeking perhaps a buyer. And of course, what would you be buying? Not the failing genetic testing company. You'd be buying the data, the vials of spit.

I don't know why they're struggling so much. Maybe people are more privacy concerned than I thought. 23andMe stock, which fell 82% last year, dropped another 10%. It was briefly halted this week. That's not good. Founded in 2006 by one of the Wojcicki's, Ann Wojcicki. Yeah, this is...

Kind of a story of a more optimistic age of the internet. It doesn't seem that long ago, but the inception of this company and their peers, it was really just we thought about things much differently, including social media. And it's just kind of the sad outcome of what could have been a really transformative company.

I know. Well, and I actually have donated my spit to other genetic companies, not ancestry.com, but I interviewed George Church, who was like the father of genomics a couple of years ago on triangulation. Fascinating guy. And he has a company that

Unlike 23andMe, which does a statistical analysis of part of your genome, he has a company that does the whole genome. In fact, you can download, I did, my genome. It's many gigabytes worth of data. The value of that is you now have your genome, which you could send to other places for analysis. You could say, well, am I likely to have breast cancer? What's my likelihood of being an addict or something like that?

So, while I did 23andMe, I was underwhelmed. Maybe this is the problem.

They also had a breach last year, or 2023, more than a year ago, of 7 million customers. They had to pay a $30 million settlement on that. In September, the entire board quit. The entire board quit, right? Yeah, I remember that. That was wild. Yeah, that's usually a bad sign. That and getting delisted. Those are kind of bad signs. The company said this week it's exploring strategic alternatives, including perhaps a sale.

I'm just like surprised they can't use AI to like do stuff with because I would think that AI would be good at discovering what genetic factors and what, you know, things in your genetics can cause different diseases. Yeah. I thought that's when I signed up. I forgot if it was that or another company. That's what I thought I was getting like, oh, do I have markers for this disease or that disease or whatever? And like now in the age of AI, that's we can we can like, yeah, I don't know why they're not.

well that's why i did the nebula genomics uh thing because that was a that was a full uh genome which you can download and whether they eventually know you know that stuff or somebody else does you have your genome so now you can send it out and you know that's not going to change so it's whole genome sequencing it wasn't cheap but it's not you know originally it was tens of thousands of dollars i think it's gotten much more uh affordable

as the technology has improved but it wasn't it wasn't completely cheap let me just see and it's still spit I believe oh yeah not awful the Elite package is a thousand bucks that does that does your whole thing and then I think that's the one I did I don't know what it was a thousand dollars poorly spent since I don't really know what I what I got but I I have my genome anybody wants it

You know, I should just send it to the AI in the cloud and see what they say. I agree with you. I think AI could, because here's the thing. We don't really know a lot about what it says, right? There's some markers like the BRCA marker for breast cancer that you could look for. Right. But we don't know a whole lot.

i guess more all the time when you say a thousand dollars it sounds like a lot of money except when you think about all of those things that maybe we want ai to do with with our right our dna and so then i think well maybe i mean if i don't know what the business application is if there's a b2b but like if they're targeting consumers a thousand bucks is a lot of money what is there a market of people who are going to pay 20 or 40 or 100 bucks like

Maybe, but does that make a business that scales? Probably at least this business, maybe not. I think there's a lot of overhead. So maybe their model is just targeting consumers with something consumers can't afford or don't want to pay that much money for. I noticed that I just logged in and they want me to change my password and they want me to add 2FA here at the Nebula. So I'm wondering...

It's good. They're focused on security, right? But I'm wondering why after they got hacked. Yeah. It always makes me nervous. Hey, you want to change your password? Just, you know, just for no reason. For no reason. No reason. No reason.

Just saying. Just saying. We just think this is a good idea, so we're just going to require that you change it because we've deleted everything because we can't save anything. And yeah. Well, that's the thing. I mean, when you give somebody, I mean, they have my full G. Now, somebody said, okay, by the way, when George had done this originally, he did it as a project. I remember it was very expensive. I applied. He was doing it, I think, at Harvard. Yeah.

uh it was called the personal genome project and a number of i remember esther dyson did it and she said i'm going to do it because i'm at an age and stage in my life where i don't care if it gets out because one of the things the personal genome project said is we are not going to protect this the whole point of this is we could take these genomes and give them to researchers so they can use them to learn and find out more stuff so you do this whole big

questionnaire and you do the genome and I applied I really wanted to do it they didn't they accept me but I remember it was 10 15 000 it was very expensive so so a thousand seemed reasonable but now I'm going to quickly change my password if you don't mind uh streaming prices wow Netflix just raised their uh price yep uh

Streaming prices climb in 2025. They already surpass inflation rates. Five services increased their prices this month. Well, last month, I guess now in January.

Do you notice that? That the prices are going up of the stuff you want to watch on TV? Yeah. I mean, Netflix just went up to $25 a month if you want to have 4K, which is annoying to me because I do want to have 4K. I don't really care about the number of devices I have active. And the thing is that...

that their ads from what i understand are fairly you know not yeah i don't want ads i mean i don't either but if you gave me the option where you were like okay if you pay you know ten dollars a month and you get 4k with ads i might consider it because i don't watch a ton of netflix but um

No, I mean, they've gotten up and I have some sort of grandfathered in Disney bundle deal that Verizon still pays for most of and American Express pays for the remainder. What's interesting, though, and it'll be, I don't think it'll ever work with the big services, but with the smaller ones. So, Starz had announced, I guess, back in like October or November, they announced their quarterly results and they showed that they'd shed users.

And I guess that was a problem for Lionsgate or whoever owns them. And I was paying, I don't know, maybe $60, $70 a year. And I don't watch Starz that much, although there are a couple of shows on there that I like. And I went in just out of curiosity to see what my plan was. And first I was offered a deal where they gave me a really good deal where they were like, oh, we'll give you a six-month thing or whatever. And I was like, cool.

cool, I'll do that. And then I did that and I got another deal to get it even cheaper for like a year. So they were actively trying to, I guess, prevent me from canceling or whatnot. And it'll be interesting to see if other streamers will do that sort of thing where, you know, if you either threaten to cancel your account or change your subscription, they'll be like, oh no, please don't leave, you know?

Uh, somebody pointed out Knox Harrington in our discord that it really looks like the plan for these companies is to just get rid of ad free streaming to not only increase the cost, but to force you to watch ads. Who just did that to me? I paid 99 bucks for a year. I thought, oh, I won't have any ads. Now I can't get rid of them. It's becoming more expensive to avoid ads on streaming services. Mm-hmm.

The prices are going up for ad supported. Are they double dipping or is it really that expensive? I don't know.

um i mean i think that it's it's the realization that their business model that they yeah yeah now charges 19 for an ad so for the cheapest ad free video streaming plan that's why my 99 didn't go very far ridiculous 20 a month for hulu is is too much um well that's the next question is at what point

Do people start canceling? Ars Technica quotes a digital media trends report from Deloitte that said almost half of U.S. consumers said they'd cancel their favorite streaming video service if the price went up by five bucks. Well, guess what? It is.

Let's see if they really do. And I think a lot of them will. What we've seen happen is it's churn, right? So you might not say subscribe the whole year, right? Maybe you'll forego that year-long bundle. If you're like, no, you know what? I only need to watch the shows these months out of the year. And as live becomes increasingly less important, right?

you know, like we have events like the Grammys and the Super Bowl. But other than that, you're like, okay, you know what? I can dip in and out and I can catch up on what I missed before. Yeah. I think all the streamers kind of realize that it's a zero-sum game and it's a race to the bottom. There will be some winners and there will be losers. And mostly there will be losers because the cost is exceptionally high. This was a bad

bet by most of the video companies at least that's the kind of conventional wisdom right now i don't watch much i i don't subscribe to anything other than youtube tv um but occasionally like severance or whatever god i know i pay almost it's cable car it's almost 100 bucks it's like 83 bucks a month it is which is still cheaper than cable i was paying a ridiculous amount of money when i finally canceled cable and i also have youtube tv but

still there i think that they all made a mistake when they shifted to do ad free the way they did ad free i think that if they had maybe you know tempered into things then then they would youtube tv started at 35 bucks a month yep and now it's like 85. it's 85. i mean is that because the local channels are gouging youtube over the top streamers like youtube fubo's raised its pro everybody's raised its prices

Some of it is probably that. Sling? Sling has gone up. Well, Sling has had some more limited offerings too. Hulu Live TV has gone up. DirecTV Now or whatever it's called. Yeah. So part of it is probably that. It seems self-defeating because in the long run, what's going to happen? People are just going to watch YouTube. Not YouTube TV, but YouTube.

Yeah, that's what I mean. I don't pay for YouTube TV. I don't watch things like I watch a specific thing, but I pay to not have ads on YouTube. There is almost no circumstance I'm going to sit down and say, what am I going to watch? Let me browse for something and then watch it. No, no, no. I'm going to read a book or I'm going to get away from a screen if possible. But I was at the streamer. I was at CBS for a very long time when I went from CBS All Access and we were there during the Paramount Plus transition. And I was at CBSN for a very long time.

I mean, the network too, but CBSN was like the first streaming news platform. And I saw all of the economics, at least when it came to the news business. And that was a losing game for news too. I mean, I love it. I love CBSN. I love NBC News Now, but those are not winners. And I like it. I,

I mean, like the consumers have voted and they don't want streaming as it is right now. At least it's too expensive and too little value for that. Right. Yeah, exactly that. Yep. Yeah. Let's take a little one last break. I got to get this one last ad in and then we'll wrap things up with a fabulous panel. Dan Patterson is always great to see you. Blackbird. You too. AI. You're the director of content. What does Blackbird AI do? What is it? What are you talking? What are you doing here?

So I think visit compass.blackbird.ai. I mean, like really, this is, I don't want to,

Call it. Go ahead. We're not going to call it a fact checker. It is a context checker. Oh, I mean, like, I'm happy to plug. Yeah. Plug, plug. But I mean, like, I don't want to call this a fact checker, but sign up at compass. Blackbird.ai. This will check claims when you see nonsense on the web, when you see a ridiculous claim, whether it's on a social media site, you can paste in a link. If you see something on Reddit, if, if your neighbor tells you that the sky is purple, you're,

Put a claim into Compass. If you see an image and you think it's a deepfake, put the image in there. It will give you not just like yes or no, up or down. It will give you a ton of context, a really interesting context that is useful and important and almost always accurate. Oh, let's see. Is Taylor Swift breaking up with...

Travis Kelsey. Yes, exactly. And in fact, we tracked many of the, what we call narrative attacks because we see them as analogous to cyber attacks and the narrative attacks targeting Taylor Swift and Travis Kelsey last year. If you look at the Blackbird AI blog, which is where I spend a lot of my time with our analysts and our AI engineers, we published a ton of results about Taylor all through last year. And much of the conversation was manipulated media content.

And anytime you now have

A question about if something is manipulated or authentic on the web, use Compass to check whether it's authentic. This is great. I signed up for this last time you were on, and I forgot to use it. I got to use this more. That's great. I really don't want the log roll, but this is AI that is in the public good. And I mean, I don't want to say fact check because, you know, it's an LLM. Sometimes maybe a fact won't happen, but it's really accurate. I keep it bookmarked. And when I have...

questions about what I see on the web, I use compass. Verify the authenticity of an image. This is great. And there is an API. So you could, in theory,

Oh yeah. Add this to your existing stack. That's great. We have it. I think maybe we're working on discord integration. We have it in our company Slack. And like, if, if you, I mean, it's just rolling all day long, people posting what claims that they see. So Slack integration. What models are you using? It's our model. It's your own. Yeah. It's not an existing LLM. We used existing LLMs last year, but I think we're on our own models. Nice. Very cool.

let me see if i can find a an image from our our discord and see if it's real yeah definitely do that because you know as me coming from like being a journalist for two decades and then like pivoting into tech like this is a very natural place to be because i mean we care about journalism we care about information integrity yeah really good it's great to have you dan thank you so much tomorrow christina warren will have a new job

Do you go to the desk and there's a new MacBook Pro sitting there and a little beanie with a propeller on it or anything like that? What do you get as a new employee? I think it'll depend. I haven't received my new stuff yet, so that's going to be the interesting part is I don't have my corporate laptop yet. What did Microsoft give you? What did GitHub give you? Did you have a nice...

I did. Yeah. Microsoft, I don't really remember. I think I got, well, it was in person, which was now a throwback. And GitHub, yeah, I got a hoodie with my handle on it. And I got a bunch of other stuff. I got like a stuffed Octocat. Oh, that's cool. GitHub hooked me up.

That's really cool. And as far as the new place, TBD, we will see. But yeah, I'm pretty sure I'll get some new merch. Oh, I can't wait to see. Yeah, I'll be excited. Good luck. That's wonderful. You deserve the best. We're always thrilled to have you on. And I hope that you can continue to be a regular on our shows. I would be sad to lose you. Yeah, likewise.

Also Shoshana Weissman, rstreet.org where she's head of digital media. And are you going to end your association with the Sloth Committee and start up with the Marmot Committee or is it just a joint venture?

think it's a joint venture they're both like i mean they're actually very different animals they're both rodents aren't they or no no no no sloths aren't rodents but they do both have faces of babies and i feel like that's a thing yeah yeah that's like my feminine nurturing side being attracted to sloths so you'd prefer alpaca for instance to llamas because they're cuter yeah yeah yeah guanacos are great they're very cute yeah i like them

like them but marmots you just want to hold them you want to like nurture them especially the yellow-bellied marmot they're so cute rstreet.org reader writings always good stuff

our show today brought to you by zscaler i love these guys because i love the idea of zero trust that's the way to protect yourselves they are the leader in cloud security you know over the years enterprises spent billions of dollars on perimeter defenses firewalls and then of course you got to have a vpn to let people in but breaches they're not going down it's not working

18% year over year increase in ransomware attacks in 2024. A record $75 million payout. And I think that's just the tip of the iceberg. That's just the amount people reported. It's got to be more, right?

The fact is traditional security tools just aren't working. They expand your attack surface. They've got public facing IPs that are just like candy to bad actors. And bad actors are now using AI tools to work better, smarter, come up with new attacks. And of course, if you're using a VPN,

you're letting bad guys into your network, but you're not making sure that they are good guys in your network. So they're getting everything and then they exfiltrate it out through the firewall because firewalls are struggling to inspect the encrypted traffic. They can't really see it. So everything's getting exfiltrated.

And the firewalls and VPNs allow lateral movement. I mean, once you're connected to the network, you can wander anywhere, find everything you want, encrypt it, send it out. People are really unable to stop this stuff. Hackers exploit traditional security infrastructure and they're using AI to outpace your defenses. You got to rethink your security. We can't let these guys win. They're innovating faster than we are. They're exploiting your defenses. That's why you need Zscaler.

Zero trust plus AI. It stops attackers. For one thing, it hides your attack surface. Your apps and your IPs are invisible. So there's no candy there hanging to be attacked. Plus, it eliminates lateral movement because once a user's in your network, they can only connect to specific apps, not the entire network. And every connection, every user is continuously verified online.

Every request based on identity and context, it simplifies security management using AI powered automation. And man, you need it because with half a trillion daily transactions monitored by Zscaler, you need AI to find those needles in the haystack, the actual threats so that you can protect yourself.

Hackers can't attack what they can't see. Protect your organization with Zscaler Zero Trust and AI. Learn more at zscaler.com slash security. That's zscaler.com slash security. We thank them so much for their support of the show. And you support, of course, the show by using that address. That way they know you saw it here. zscaler.com slash security. All right, Burke has submitted an image for me to test.

Burke, you think this is a phony image? Let me go to blackbird.ai and see if I can verify, validate that image. Go to compass.blackbird.ai and click vision at the top. Yeah, I'm at the vision page. Oh, I can't paste it though. I have to save it and then paste it. Well, I'll do it later. We don't know. We don't know. Oh, you can even do batch. That's cool. That's cool. Really cool.

All right. We got to get going because, you know, you got to watch the Grammys. How are you going to watch it? YouTube TV, Fubo? YouTube TV. Yeah. YouTube TV. Yeah. I'm recording it. So, you know, I can watch it anytime. Those things, I like to record those because I don't want to watch eight minutes of commercials after every break. So I record half an hour or an hour ahead of time and then I skip the commercials. And then you can see through, which is nice.

Well, and then what's great too, I'm still not used to being on the West Coast for award shows, even though it's been like seven years now. I will never be used to it. But the nice thing is, is that like they rerun them. That's the Grammys. They show it twice, don't they? Yeah. And the same thing now with the Oscars and the, and the, and the, the,

uh golden globes or whatever they'll just replay do they do they run it again oh i didn't know that at least at least um on the the last few years at least i don't know if the oscars that they rerun it but at least the other things like they'll you know because what else are they going to do like they used to always uh i think what it used to be is they used to start them late so west coast would always get it delayed but in the last five or six years they've

been like, oh no, we'll just broadcast them live everywhere. But that means that if you are on the West Coast, what are you going to blow your prime time feed for? So. Benito, are we, because next week's the Super Bowl, are we going to shift the time on Twit because of the Super Bowl next week? That's the plan, right? Yeah. Noon, right? Yep. So we're going to launch two hours earlier so that we can get everybody in and out before the beginning of the Super Bowl at 3 p.m. Pacific. So we'll do it at noon Pacific, 3 p.m. Eastern if you watch that.

uh live super bowl ads eight million dollars this year for a 30 second commercial and uh they're already i've already seen a few previews my son who is a tick tock chef as so he's kind of a budging bud budding celebrity chef was invited to cat's deli in new york city to watch the making of the hellman's mayonnaise super bowl commercial

It's a good deli. It's a great deli. He's friends with the owner now. And I'm not going to, no, no, I'm not going to tell you, but just watch it because it recreates a classic film moment that was filmed at the Katz's Deli some years ago. Oh, oh, is this the one that already got? Yes. Okay. Yeah, yeah, yeah. Okay. Got it. Got it. They're a little older now, but they're just as frisky, I guess would be the, would be the word.

There will be a lot of AI ads, a lot. All the AI companies, this is what happens. I remember in 2000, all the dot coms. - Not the dot coms. - The first dot com ad, which was, was it pets.com? I can't remember. - Yeah, the puppet. - Yeah, the dog, yeah. Whatever happened to that? You know, it's funny, the conventional wisdom was pets.com went out of business because it made no financial sense to ship kitty litter

at great expense to people. So they went out of business

And now we buy our kitty lighter online and it's shipped to us. Well, no, they were just too early. I think the problem then back then was they were too early. It was a lot of things. Like I remember there was a web van. Remember them? They were like the grocery store kind of thing. And now everybody's doing it. Yeah. They were just too early. And I think the bigger thing too was that rather than charging what it actually costs to do things, which everyone would have paid, they were like, oh no, we're going to be cheaper because it's the internet and we have free money. And it was like, no,

no. Scale first, profit later. Or no, scale first and then like cut your prices if you're trying to do that. What was the company? You were in Manhattan when this was around. They would deliver cigarettes and ice cream by bicycle? Cosmo. That's every company all the time. Was it Cosmo? It was Cosmo.com I think is what you're thinking of. I was, yeah, they were early. They were like, it was basically Postmates but like

15 years before Postmates. They were just ahead of it because it was a great idea.

It was. I used to have them delivered to, so I was actually in Atlanta. I was in high school and I used to have Cosmo deliver stuff to my school. I'm not even joking. I used to get like deliveries, like DVDs. Pizza for Spicoli, pizza for Warren. I mean, I had pizza delivered too, but Cosmo was the one where I would get yelled at by the administrators. They'd be like, Christina, you have to stop ordering DVDs to the front office. And I'm like, but I can't.

K-O-Z-M-O. K-O-Z-M-O. Cosmo.com. Free one hour delivery of videos, games, DVDs, music, mags, books, food, basics, and more. I got a number for that. I live in Brooklyn. Cosmo. They're gone. They're gone. They were bought by yummy.com who said, oh, we're going to relaunch soon. And then no, eventually they were relaunched as a warehouse club.

There's a documentary that is out of print that I don't know how to find it. It's been uploaded to YouTube a few times, but it's not available. Last time I checked, it's called the eDreams. And it came out, I want to say, in like 2002 or 2003. And it's about the rise and fall of Cosmo. And it's really interesting. One mistake they made, they agreed to pay Starbucks in February 2000, $150 million to promote Cosmo in the coffee shops. Wow.

They were guaranteed to go out of business. Oh yeah, absolutely. Again, the prices that I would pay as like a 16 year old, I would like order, I'm not joking. I would like order from them and like, you know, the year 2000, whatever, I would like be ordering my DVDs and, you know, candy and whatever. And it would be cheaper than going to Best Buy or, you know, the mini mart. And then it would be delivered to my, to my high school. And I'm like,

Okay, you're going to charge me a fraction of the price. Why would I not? Again, the only trouble I got in was the nice ladies in the front office were annoyed. They were like, we can't keep calling you to the office to get your packages. It's like noon. I was like, well, but let me take it to the classroom. The former co-founder and CEO of Cosmo now is president of BibleGateway.com.

Okay. They sell Bibles. No comment.

Super Bowl Sundays are very noisy. Apple Watch, as you know, will tell you if you are in a noisy environment. I leave that on. I think that's very useful. I do too. It's great. Yeah. I have mine set at 90 dB. Any long exposure above 90 dB, you're going to have hearing loss. So they did a study, Apple Watch study, found that noise levels across America are significantly higher than normal for about nine hours on Super Bowl Sunday, next Sunday. What?

If you were in the stadium, of course, but I guess people have parties and they shout and they scream and there'll be some Eagles fans and Kansas City fans making a lot of noise. Average noise levels 1.5 to 3 decibel louder during the past four Super Bowl games compared to levels on the Sunday following the game. It's a lot quieter now than it will be next week.

A reminder, we'll be starting a little early next week. Join us for a Twit Early or do what most people do and subscribe. That way you can listen whenever you want. We stream, as I mentioned, on eight different platforms. Normally 2 p.m. Pacific, 5 p.m. Eastern, 2200 UTC of a Sunday. Again, we're going to start two hours earlier next week. You can always download a copy of the show from our website, twit.tv, or subscribe in your favorite podcast player or watch us on YouTube TV. There's audio and video.

Of the show and of course if you're a club member special access behind the velvet rope my deepest thanks to our panelists Shoshana Weissman, thank you for being here. Thank you for having me. It's always so much fun I always want to play climb every mountain when you're on working on it. You're a mountain climber. I

That's so awesome. What's your next peak? Thank you. So the next trip I have is coming up in May, and it's to do the high point of New Mexico, maybe a 14er, and then whatever else I can fit in. Wow.

So cool. Do you huff it up? You go fast or you kind of slow? Oh, I'm extremely slow. I mean, like I've had lead poisoning the most of my life. So now I'm like, I'm hoping like I get faster and stuff, but I'm slow. I train fast, but it's on a treadmill. Oh yeah. You do it. That's what matters. And I'm getting faster. I'm getting a lot faster than I used to be. Like I pass like three years ago, me easily. So it's getting better. I'm a slow walker. I stroll. I think it's good to stroll. Life is.

is happening fast enough stroll that's my motto i like your motto stroll i'm stroll i'm the yes stroll you you know what you gotta you gotta stop and smell the wombats is what i'm saying yeah did you know they poop square poop oh yeah that's their thing that's their big thing that's their thing that's great everybody ought to have a thing thank you shusha

thank you uh dan patterson so great to see you as always director content blackbird.ai do that blackbird thing that is compass.blackbird.ai that is so cool and that's free yeah it's so cool thank you for doing your bit uh to make sure that this information gets destroyed well we like to call it a narrative attack no more narrative attacks thank you so much dan and of course christina warren congratulations on the new gig thank you

I love the new house. It's very attractive, very beautiful. It is. It's a good look for me. I like the pink walls. I think that it has kind of a retro aesthetic at this point. Kind of 60s, kind of 80s. It's really cool. Yeah, I like it. And the carrot curtains in the kitchen. That's great. Yeah, the corn. It's great. Is that corn? Okay. That's corn. Carrots. Yeah, whatever. It's a vegetable of some sort.

thank you christina i can't wait to find out where you will land next if is the best thing to do follow you film at film underscore girl or um you can find me on twitter mastodon i have to still be on twitter i'm asked on blue sky all those things still film girl yeah film girl and it's usually film underscore girl but some of the networks don't like underscore so it's just film girl one word but yep so is blue sky your preferred these days like everybody else i don't

I don't know. I mean, I'm kind of trying to post across them. I probably use threads the least, but yeah, blue sky is definitely has the most vibes, but I still really love Mastodon. You know, I have a lot of good people there. So it kind of varies. Yeah. I prefer hanging with geeks. Blue sky is too much like old Twitter.

Which is why people love it. It's a good, it's a double-edged sword. It's a good and a bad thing. But yeah, so still trying to figure them all out. I still resent the fact that we have to have five when we used to have one, but it is what it is. Well, I just posted my blog and let that post everywhere. And that seems to me the best way to do these things.

Thank you to our esteemed producer and technical director, Benito Gonzalez. Appreciate the work you do. Benito's working on our theme song for Wednesday, the launch of Intelligent Machines. It's going to be a lot of fun. I look forward to seeing you on Tuesday or Wednesday, I should say. Well, Tuesday, too, because I'll be here for MacBreak Weekly and Security Now and then Wednesday is Windows Weekly and Intelligent Machines every Sunday. It's twit.

Thanks to our club members for making this possible. Thanks to all of you for being here. We will see you next time. You know, this is our, I say this every time now, this is our 20th anniversary year. April will be the 20th anniversary of the first twit. Can you believe 20 years? That makes me feel so old. I've never done anything for that long. It's wild. Cause I remember, I remember when you launched it and, and that's bizarre. That's amazing.

congratulations seriously yeah i don't well i don't know if it's congratulations or oh it is init it's just inertia no no it's congratulations like this is like like you help pioneer a new medium and then see it go into all these other things and you're still doing it i've been doing it long enough that it is not only no longer the new medium it's kind of old school kind of going downhill now

I jumped from radio to podcasting. I don't know what I'm going to have to do. I'm going to have to do something. Maybe AI is the next thing. I don't know. There you go. Thank you, everybody. Great to see you. As I've been saying for 20 years, thanks for being here. We'll see you next time. Another twit is in the can. Bye-bye. This is amazing.