AI Weekly News Rundown April 20-27: 🚨 Alarming Rise in AI-Powered Scams 🧠Godfather of AI Geoffrey Hinton Warns AI Could Take Control from Humans ✈️Artificial Intelligence Enhances Air Mobility Planning 👨‍💻AI Now Writing Over 30% of Google's Code
29:51 Share

Welcome to a new deep dive from AI Unraveled. Great to be here. With a show created and produced by Etienne Newman. He's a senior software engineer and passionate soccer dad up in Canada. That's right. And hey, if you're finding these deep dives valuable, please take just a second to like and subscribe to AI Unraveled on Apple. Yeah, it really does help other folks find the show. It really does. So this week, wow, just a whirlwind in AI, wasn't it? Absolutely packed. Hard to keep up sometimes. Totally. We've got like

huge potential funding rounds, AI slipping into our everyday tech. Yeah, more and more. Ethical stuff, the future of work, even hints about AI and governance. It's a lot. It really is a broad spectrum. So we're going to try and unpack the most crucial bits for you from the last week or so, April 20th to 27th. We've sifted through quite a few articles,

Business, tech, ethics, even arts. Yeah, trying to get a good cross-section. Our goal today really is to give you a clear sense of the key happenings. Highlight those aha moments, maybe connect some dots. Exactly. You feel informed without drowning in the sheer volume of news. Let's do it. Where should we start? Okay, let's kick things off with the big money. XAI, Elon Musk's company. Reports are talking about raising, get this,

$20 billion. $20 billion? That's a number. Right. Especially after acquiring X, formerly Twitter, this could apparently value them over $120 billion. Which would make it, what, the second biggest startup investment ever? Yeah. Potentially. Potentially, yeah. Huge signal, isn't it? Massive signal. It just screams continued investor enthusiasm for AI, despite maybe some broader market jitters. And what does that kind of cash actually do for them?

Beyond just, you know, bragging rights. Well, it gives them serious firepower. We're talking resources for computing, attracting top tier talent. Right. And really accelerating their research and development. Plus, maybe helps with X's debt situation, too. Ah, interesting angle. So it bolsters XAI and potentially helps the mothership. It seems likely. Enables deeper AI integration across their platforms, for sure. OK, so massive investment flowing in.

And we're seeing the results trickle down, right? Microsoft just announced some big AI updates for Windows. Yeah, for the new Copilot Plus PCs. They launched Recall, AI-powered Windows Search, and Click to Do. Recall is the one grabbing headlines, I think. It captures encrypted snapshots so you can find stuff later. Exactly. And immediately, you know, privacy flags go up. Of course. Microsoft addressed it saying it's opt-in, the snapshots are encrypted and stored locally. But still, the...

concept of your computer constantly screenshotting itself, it

makes people uneasy. Yeah, it's that trade-off again, isn't it? Convenience versus privacy. What about the others? Search and click to do. Search is improved, uses natural language, processes locally, which is good for privacy. And click to do lets you select text or an image on screen and have Copilot do something with it. Edit an image, start an email, stuff like that. So definitely aiming for productivity boosts, but recall is the one that sparks the debate.

pretty much useful maybe but raises those fundamental questions got it

OK, shifting gears to the hardware side, Intel. Right. They're apparently making a big strategy shift, focusing on developing their AI chips in-house now, trying to take on NVIDIA. Yeah, that's a significant move. It signals a really deep commitment to becoming a major player, not just relying on acquisitions or partnerships. But NVIDIA's got such a huge lead, doesn't it, with their whole ecosystem? They do. It's a massive challenge. NVIDIA offers the hardware, the software, the whole package.

whole package. Intel acknowledges that. So what's Intel's angle? It seems they're focusing on refining their existing products for specific areas like robotics and autonomous agents trying to carve out niches. Makes sense. Play their strengths. Exactly. It shows they're serious about competing long term in the AI chip market. Okay. Competition heating up there. And speaking of challenging giants, let's talk perplexity. Ah, the AI search engine. Yeah. But their CEO, Ervin Srinivas,

He's got bigger ambitions. He's positioning perplexity to challenge Google, not just in search, but in browsers. With their upcoming browser, Comet. Right, right.

And the concept is interesting. They're calling it a containerized OS for AI agents. What does that actually mean? Well, think about your phone now. Apps are mostly sandboxed, right? They can't easily talk to each other or control deep system stuff. Yeah. For security? Exactly. But that limits what AI agents can do. Say you want an agent to compare real-time prices across three shopping apps or book travel based on your calendar and flight alerts. Can't really do that smoothly now. Right.

Comet aims to be an environment where agents can have that deeper access to reason and act across different services more fluidly. OK, I see the potential benefit.

But. There's always a but, isn't there? There often is. They also mentioned Comet will monitor user activity, browsing habits, purchases, location for hyper-personalized ads. Ah. So similar privacy concerns as with Google or Meta. Pretty much, yeah. It could redefine how we interact with the internet, heavily driven by AI personalization. But raising those same questions about data collection and tracking. Exactly. Convenience versus privacy. Round two. Seems to be a recurring theme.

Unfortunately, AI isn't just powering helpful tools. It's also fueling bad stuff. Yes. Scams. Yeah. Microsoft reported swarming over $4 billion in AI generated fraud attempts just in the past year. $4 billion. That's staggering. It really is. Phishing, impersonation scams, using deep fakes, financial fraud. It's getting incredibly sophisticated. So cyber criminals are weaponizing AI just as fast as everyone else is developing it.

It's an arms race, basically. It means cybersecurity strategies need to get much stronger, much faster. And it's not just outright crime. We saw that weird case with the lawyer for the MyPillow CEO. Oh, yeah. Citing fake cases generated by AI in a legal filing. Yes. A federal judge called him out on it. That's not good.

Not good at all. It's a huge warning sign about relying on generative AI in critical fields like law without rigorous verification. You can't just trust the output blindly. Absolutely not. These tools can hallucinate, make things up entirely. Human oversight is non-negotiable, especially when legal precedent is involved. Really underscores the risks. Yeah. And speaking of risks, Jeffrey Hinton. The godfather of AI.

He's out there again reiterating his warnings about future AI potentially taking control, saying the existential risks are being underestimated. Yeah, his voice carries a lot of weight given his background. He's consistently urging caution about the long term trajectory. What should we take away from these repeated warnings? It adds significant fuel to the global debate on

on AI safety and governance. He's saying, hey, let's not just focus on the cool stuff now. Let's think seriously about where this could lead if we're not careful. It's a call for proactive safety research and ethical consideration, basically. Precisely. A crucial perspective in the conversation. Okay. Let's pivot to some potentially more

uplifting applications, air mobility. Yeah, MIT researchers developed AI tools to optimize planning for things like flight coordination, air taxis, even emergency response. How could AI help there? Air traffic control seems pretty complex already. It is complex, but AI can process way more data in real-time weather, schedules, aircraft data, ground traffic. To find more efficient routes,

predict conflicts, manage drone traffic alongside planes, coordinate air taxis. So smoother travel, maybe fewer delays, better emergency response too. That's the goal. Optimize the whole system, make it safer and more efficient, especially as we get more things flying around.

Sounds promising. And speaking of physical things, China unveiled an advanced humanoid robot. Right, with AI-enhanced eagle-eye vision and complex real-time task capabilities. What's the significance there? We see robot demos all the time.

This points to rapid progress in robotic perception and decision making. A robot that can see accurately and act complexly in real time. That has huge implications for manufacturing, logistics, maybe even healthcare. More versatile, more capable robots. Exactly. Getting closer to robots that can genuinely operate effectively in human environments. Meanwhile, Apple's making moves in robotics too, but kind of internally. Yeah. That secret robotics unit reportedly moved from the AI division to the hardware group. What?

What does that suggest? It suggests they might be getting serious about actually building robotic hardware, not just researching AI for it. Better synergy between the AI brains and the physical body, maybe. And the concepts are interesting. Expressive AI lamps, tabletop companions with robotic arms. Yeah, hints at integrating robotics into the home and typical Apple fashion sleek consumer focused could lead to new product categories down the line. Definitely one to watch.

Now, this next one is fascinating. Anthropic launching an AI welfare research program. Yeah, diving into the ethics of potential AI consciousness and moral status. That's pretty deep. Seriously. Are we talking about robot rights already? Well, not exactly rights yet, but exploring the foundations. Can we assess consciousness? Can AI have preferences or feel distress? How would we even know? Wow. They even hired a researcher who estimates, what, a 15% chance current models might be conscious? Kyle Fish, yeah.

Heavy stuff.

Okay, back to the creative side. Adobe's Firefly. Big updates there. Firefly image model 4 and 4 ultra-launched. Better realism, more control, higher resolution. Yep, up to 2K resolution now. And importantly, the Firefly platform now integrates third-party models too. Oh, like whose? OpenAI's GPT image gen, Google's Imogen 3, VO2 for video, Flux, and...

A bunch of them. So Adobe is becoming more of a hub for different AI creative tools. Seems like it. Offering versatility, but within their framework that emphasizes commercial safety standards.

They also took text-to-video out of beta and released a text-to-vector model. Lots of new tools for creatives then. Definitely. More power, more options. And Google DeepMind is boosting its Music AI sandbox. Yeah. New create, extend, and edit features powered by Lyria 2, which they say has higher fidelity audio. And real-time music creation too. Lyria real-time. Exactly. Let's musicians interactively create music with the AI. They're expanding access to more musicians in the U.S. too.

So AI is a musical collaborator helping overcome creative blocks. That's the idea. Offering new ways to experiment and generate ideas. Interesting. Now, inside Google itself, AI is apparently writing a lot of code. Over 30 percent of new code, according to internal disclosures. 30 percent. That's huge. It is. It clearly accelerates development cycles, lets engineers focus maybe on harder problems. But what about quality?

bugs, security? That's the big question, right? It definitely raises concerns about quality assurance and the need for really rigorous human review of AI-generated code. You can't just let the AI run wild with the code base. Definitely not. Powerful tool, but needs careful management. And there are issues with AI use being hidden.

Hundreds of scientific papers apparently used AI text without saying so. Yeah, that's a serious problem for academic integrity. Transparency is crucial in science. Why is undisclosed AI use such a big deal? It obscures authorship.

makes it hard to check for bias or inaccuracies introduced by the AI, and undermines trust in the research, we urgently need clearer guidelines on disclosure. Absolutely. Trust is fundamental. Yeah. On a related note, MIT developed a periodic table of machine learning. Kind of a cool idea. A framework to help scientists quickly find the right AI methods for their work.

So making AI research more intuitive, speeding up discovery. That's the hope. Help researchers navigate the complex landscape of AI techniques more easily. OK, now this was a bit controversial. The California bar exam. Oh, right. The examiners admitted using AI tools to help draft questions. Yeah. And they didn't tell the candidates. Oof. That raises ethical alarms, doesn't it?

Fairness, transparency, potential bias in the AI generated questions. Exactly. Using AI in high states professional certification needs very clear rules and disclosure. Agreed. Lots to figure out there. Now, despite all the complexities and debates, the underlying infrastructure boom continues, right? AWS and NVIDIA.

Still reporting incredibly strong demand for AI-focused data centers. Growth is rapid. So the picks and shovels of the AI gold rush are still selling like hotcakes. Absolutely. The need for massive compute power isn't slowing down. It's a booming sector. And OpenAI is making its tools more accessible, too. They released their advanced image generation model via API. GP Image 1, yeah. Allows developers to iterate that powerful image creation into their own apps. So democratizing access.

Pretty much. It supports diverse styles, gets text and images right, includes C2PA metadata for tracking provenance. Who's using it already? Early adopters include Adobe, Figma, Canva, the big creative platforms. Makes sense. Putting powerful visual tools in more hands, hopefully responsibly. That's the goal. Okay, before we get too much further, let's mention our creators app again. Good idea. Etienne Newman, who creates and produces AI Unraveled,

Also developed the AI-powered Jamga Tech app. Right. It's loaded with performance-based questions. Really effective prep. Yeah. Designed to help you pass over 50, that's 5-0, in-demand certifications. Think cloud, cybersecurity, healthcare, business. Really critical fields right now. Totally. If you're looking to level up your skills, it's a fantastic resource. Links are in the show notes, of course. Definitely check it out. Okay. Back to the news.

Microsoft, again, they introduced new AI agents. Researcher and analyst, yeah. Designed for complex tasks with an M365 co-pilot. This is part of their bigger frontier firm's vision, right? Where AI agents collaborate with humans. Exactly. They're painting a picture of the future workplace where AI doesn't just assist, it takes on more autonomous roles as a partner. And the research suggests companies adopting AI are thriving.

handling workloads better, workers more optimistic. That's what they're reporting. Their belief is that eventually every employee becomes an agent boss directing AI partners. That's a pretty fundamental shift in how work gets done. It really is. Redefining roles and productivity. And Anthropics Cloud is getting more practical workplace features too. Meeting prep. Yeah, it can now analyze your emails, calendar, documents to generate briefings, suggest agendas, draft follow-up notes. Okay, that actually sounds incredibly useful. Takes away a lot of the pre-meeting drudgery.

Could significantly improve efficiency for sure. Let people focus on the actual meeting content. Makes sense. Now, there's a challenge brewing against OpenAI's restructuring. Right. The move from nonprofit to a for-profit, capped profit or public benefit corporation. Yeah. Some former employees and big AI names, Hinton, Bengio, Mitchell, are urging authorities to block it. What's their main concern? That it compromises the original mission.

that prioritizing investors, even capped ones, over the public good will change OpenAI's direction and maybe make them less focused on safety. Ah, the core governance debate. Profit motive versus the original charter of benefiting humanity. Exactly. It highlights those ethical tensions around commercializing powerful AI research. A really important fight to watch could set precedents. Definitely. Meanwhile, out on the roads, Tesla. Robotaxi trials. Yep.

Yep. Started supervised trials with employees in Austin and the Bay Area using their full self-driving tech. Supervised, meaning safety drivers are still involved. Yeah, safety drivers initially. But they're planning a commercial launch potentially by June 2025. That's a significant step towards autonomous ride hailing actually becoming a service. Huge potential to transform urban transport if they can pull it off safely and get regulatory approval. Big ifs.

but a major milestone. And Google shared some numbers for Gemini, their chatbot. Oh, yeah. How's it doing? 350 million monthly active users as of March.

Which sounds like a lot. It is a lot. But apparently still trails competitors like ChatGPT significantly. Daily users were around 35 million. Okay, so rapid adoption but still playing catch-up in the head-to-head chatbot race. Seems that way. The landscape is still very competitive. No kidding. And here's a wild one from Google's antitrust trial. OpenAI's head of product testified. Yeah, about what? That if Google is forced to sell off Chrome browser due to antitrust rulings, OpenAI would consider buying it. What?

Whoa. Open AI buying Chrome. Why? Their thinking is integrating ChatGPT deeply into Chrome could create a vastly superior AI-driven browsing experience. I mean, maybe. But imagine the implications. Open AI controlling the dominant web browser. Right. Massive expansion of their influence. Huge questions about competition. Data privacy. It's a bombshell idea. Definitely. A potential earthquake for the web if that ever happened. Okay. Shifting to social media and content creation.

Instagram launched a new app. Edits, right. A standalone video editor. Yeah. Looks like a direct competitor to CapCut, which is owned by TikTok's parent company. Makes sense. Keep creators using Instagram's tools. What features does it have? Advanced stuff, apparently. Yeah. AI animations, green screen capabilities, aiming to give creators robust tools for short form video.

empowering creators within their ecosystem, standard playbook, but important for staying competitive. Totally. And speaking of competition, Apple's trying to boost Siri. Seems like it. New engineering chief, Mike Rockwell, he's apparently overhauling the team, bringing in key people from the Vision Pro project. Ah, bringing in the A team from a cutting-edge project. That signals intent. Yeah, revamping teams focused on speech, understanding, performance, UX. A full overhaul, maybe? Sounds like it.

Sounds like a serious strategic push to make Siri actually competitive again. Long overdue for many users. Agreed. Now this is impressive. A Korean startup, Nari Labs. Never heard of them. Founded by just two undergrads, they released an open source text-to-speech model called DIA. Okay. And it's reportedly better than models from Eleven Labs, OpenSource,

OpenAI Meta. Two undergrads beat the giants. Seriously. That's the report. Supports emotional tones, multiple speakers, non-verbal cues, really high quality. Wow. That just shows you innovation can come from anywhere. And open source too. Incredible, right? Underscores the power of small focused teams in open development. Amazing story. And speaking of OpenAI again...

the washington post has joined their alliance what does that mean chat gpt can now provide summaries quotes and links back to washington post articles ah some more media outlets partnering with ai platforms for distribution and visibility yeah seems like a growing trend media adapting to how people find information now navigating the new landscape makes sense and looking ahead anthropic ciso had a prediction yeah ai powered virtual employees

with their own digital identities integrated into corporate networks, possibly as early as next year. Virtual employees, like autonomous agents doing jobs on the network. That's the idea, which, as the CSO pointed out, raises huge cybersecurity questions. You bet it does. How do you manage their access? Monitor them.

Who's responsible if they mess up or get hacked? Exactly. A whole new category of insider threat almost. Needs a total rethink of security protocols. Definitely a future challenge for security teams. Now for the movie buffs.

The Oscars people confirmed something. Let me guess, AI in movies. Yep. Films using AI-generated content can be eligible for Oscars as long as they meet all the other existing rules. So AI isn't automatically disqualifying. Interesting. Opens the door for AI-assisted filmmaking that emphasizes transparency and presumably human creative control still being central. A pragmatic approach acknowledging AI as a tool. And AI is hitting high fashion too.

Orma Kamali, the iconic designer. Using AI. Exploring new materials, silhouettes, even personalized styling for customers. Wow. So AI is a creative exploration tool in fashion. Yeah. Augmenting the designer's process, not replacing it.

shows AI's potential even in very aesthetic fields. Reshaping creative industries across the board. And that open source TTS model, Daya, we mentioned. The one from the undergrads. Yeah, it's apparently getting picked up by indie developers. Yeah. Really challenging the big players. That's fantastic. Leveling the playing field, empowering smaller creators with high quality voice AI. Exactly what open source can do best. Great to see. In healthcare, another promising collaboration.

Biostate AI and Weill Cornell. What are they working on? Creating AI models for personalized leukemia care using genomics and health records data. Ah, precision medicine for leukemia. Huge potential impact there. Tailoring treatments based on individual patient data could be revolutionary for oncology. Absolutely.

AI-driven personalization is a major goal in healthcare. And back to hardware rivalry. Huawei. What up them? Set to start mass shipments of their new Ascend 910C AI chip next month, May 2025. Okay, so a powerful new chip entering the market. And positioning Huawei as a leading domestic alternative in China, directly challenging NVIDIA's dominance, especially given U.S. restrictions. Right. Accelerating China's AI self-reliance efforts could re-emerge.

reshape the global chip market. Big geopolitical and tech implications there. For sure. Anthropic, again, doing more interesting internal work. They actually charted Claude's value. Charted its values. How? Analyzed like 700,000 interactions, identified over 3,000 distinct values reflected in Claude's responses, practical, social, protective, etc. Trying to understand the model's underlying principles. Exactly. And found these values shift based on context.

It provides a foundation for trying to align AI better with human values. Fascinating insight into the mind of an LLM, a step towards value alignment. Definitely. Now, this is ambitious. The UAE plans to let AI write laws. Write laws or just assist. The aim is using AI to help draft, review, and update federal and local laws. For efficiency, reduce delays. Wow. AI-assisted governance. That's bold. Could set a precedent, but also raises massive questions about policy.

automation versus human judgment, bias, accountability. Huge debate potential there. Automation in lawmaking is a whole new level. And DeepMind's CEO, Demis Hassabis, made a pretty stunning claim. What now? Said AI could potentially end all disease, maybe within a decade.

by accelerating drug development. And all disease. That's incredibly optimistic. Extremely. Huge potential if AI can revolutionize drug discovery that much, but also massive hurdles, biological complexity, regulation, ethics. Right. Aspirational goal, but the timeline seems ambitious. Still shows the level of belief in AI's potential. Absolutely.

Back to social media safety. Instagram is expanding AI for spotting underage users. Trying to catch teens misrepresenting their age. Yeah, using AI-powered age detection more broadly now, stepping up youth protection. Important effort, but again, accuracy and privacy concerns are always there with age verification tech. The constant balancing act. Always. And the DOJ, in the Google antitrust case, is arguing Google could use AI to extend its search monopoly. Through exclusive partnerships.

integrating AI in ways that disadvantage competitors, further entrenching their dominance. So AI becomes another tool for maintaining market power.

That's the DOJ's argument. The trial's outcome could reshape the AI-driven search landscape and future regulation. High stakes indeed. Here's a fun fact, or maybe not so fun for OpenAI's budget, Sam Altman said. Yeah. Users saying please and thank you to ChatGPT cost them millions in extra compute. Just being polite costs millions. Seriously. Apparently. Those extra words, scaled across millions of interactions, add up in processing costs.

Wow, shows the sheer scale and cost of running these models. Every little bit counts. Makes you think twice about chatbot etiquette, maybe. Or maybe not. Probably not. And speaking of chatbots and money, OpenAI and Shopify might team up. For what?

In chat shopping. Yeah. Testing an experience where you could buy Shopify products directly within a chat GPT conversation. Conversational commerce. The next frontier for chatbots. Could be. Seamlessly blending chat and shopping. Interesting development. Now, there's an AI startup called Mechanize with a very direct goal. Which is? Replace all human workers. Automate every job using AI agents.

And they're backed by some tech leaders. Good place, everyone. That's intense. Really throws fuel on the fire of the whole AI and jobs debate.

raises huge ethical and societal questions. Understatement of the year. That's a polarizing vision. For sure. On a more positive medical note, Alibaba's AI cancer tool. The one for pancreatic cancer, Damopanda. That's the one. It received FDA breakthrough device designation. That's significant. Means the FDA sees real potential for improving diagnosis or treatment. Yeah, a big step for integrating AI into early detection for a really deadly cancer. Great news. Hope it pans out.

But AI support bots can still mess up. Cursor AI's coding assistant bot. What happened there? It hallucinated a login policy, just made one up. Caused user confusion, people canceled subscriptions. Ouch. Shows why human oversight and customer support, even AI-powered support, is still crucial. AI can go off the rails. Totally. Can't just set it and forget it. Nope. DeepMind also apparently shifting its AI learning approach.

more experiential. Meaning learning from interaction with the environment rather than just static data sets. Yeah. Aiming for more adaptability, learning like humans do through experience and exploration. Could lead to more robust AI that handles real world complexity better. A move towards AGI maybe. Potentially.

more dynamic learning. Interesting direction. And beware of new phishing attacks. There's one apparently fooling even Gmail security. Oh, great. Exploiting Google tools somehow. Yeah. Sophisticated scam getting through filters. Just a reminder to stay vigilant. Always. The threats keep evolving. Meta is also ramping up its AI age detection efforts.

Proactively finding and adjusting suspected underage accounts. More focus on protecting minors online. Necessary, but again, the privacy accuracy trade-off. It's unavoidable, it seems. Now, data is showing Google's AI overviews, the summaries at the top of search. Yeah, what about them? They might be draining clicks away from websites, significantly reducing click-through rates. Ah, so Google answers the question directly, and users don't need to click the original source. Let's like it.

which could be a huge problem for publishers and websites relying on search traffic. Could force a big shift in SEO and content strategies, changes the web traffic economy. Definitely something to watch. And rumors are swirling. OpenAI might build an AI-powered social network. A social network centered on what? Potentially AI-generated images, interactive visual content, a different kind of social interaction. Could reshape digital identity, but also huge potential for technology

Deep fakes, privacy issues, risky territory. Very, but potentially disruptive. True. Let's end on a couple of positive applications. AI text alerts for snow leopard conservation.

That sounds cool. How does it work? AI systems analyze camera trap data or other sensors to detect snow leopards, then alert rangers to help prevent poaching. Using AI as a tool to protect endangered wildlife. Love it. Great example of AI for good. And AI is entering youth sports, too. For skill tracking coaching. Yeah, personalized coaching, talent ID, could democratize access to high-level analytics. But also raises questions about...

Data privacy for kids. Fairness. Pressure. Needs careful implementation, for sure. And finally, Dimas Hassabis demoed Genie 2. Another deep mind AI. What does this one do? Build interactive 2D video game worlds from simple image prompts. Create a playable game world from a picture. Wow. Could revolutionize game development, especially indie games, maybe education too? Incredible creative potential there. So, as you can clearly see, the world of AI is just...

moving at light speed. Absolutely. This week alone, massive investments, new features in everyday tech, huge ethical debates, AI transforming industries from law and medicine to music and fashion. The landscape just packed with opportunity, but also really complex challenges, isn't it? No doubt about it. So for you, the learner, we really hope this deep dive gave you those aha moments.

A solid handle on the week's big AI news without getting lost. Yeah, keep an eye on these trends. They're definitely going to shape our future in some pretty profound ways. So here's a final thought to chew on. Considering everything we talked about, AI, writing code, maybe laws, creating game worlds, the possibility of consciousness, autonomous agents.

What do you think will be the biggest surprise AI throws at us in the next year? Or the biggest challenge. Lots to think about there. Definitely. And hey, one last reminder. Don't forget to check out Etienne's AI Power Jamcat app if you're aiming for those industry certifications. Those performance-based questions are a seriously effective way to prep. Links are right there in the show notes. Thanks so much for taking this deep dive with us. Yeah, thanks for listening. Until next time, keep exploring.