How the War for Talent Will Shape AI
22:53 Share

Today on the AI Daily Brief, some incredible updates in the talent war between OpenAI and Meta. Before then in the headlines, Anthropic has started tracking the economic fallout from AI. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. All right, friends, quick announcements as we get into today's show. First of all, thank you to today's sponsors, Blitzy, Vanta, Agency.org, and Superintelligent. And to get an ad-free version of the show, go to patreon.com slash AI Daily Brief.

Welcome back to the AI Daily Brief Headlines Edition, all the daily AI news you need in around five minutes. We kick off today with a new program from Anthropic. The initiative is called the Economic Futures Program, and in short, it seeks to support research on how AI is impacting employment and the broader economy. The goal is also to help develop policy proposals to get ahead of any consequent shifts.

Sarah Heck, Anthropic's head of policy programs and partnerships said, Everybody's asking questions about what are the economic impacts of AI, both positive and negative. It's really important to root these conversations in evidence and not have predetermined outcomes or views on what's going to happen.

Now, if you are a regular listener to this show, you will know that Anthropic CEO Dario Amadei has long been of the view that AI labs have a responsibility to participate in the policy discussion and also has been increasing the volume of his warnings about potential job displacement and workforce disruption over the past few months. His most stark warning came in May when he predicted that AI could wipe out almost half of entry-level white-collar work and could drive unemployment as high as 20% over the next five years.

In January, while discussing the possibility of a post-work society, he said, I think AI companies are going to be at the center of that or in the crosshairs of that, and we have a responsibility to have an answer there.

So, this new initiative has three pillars. First, Anthropic will hand out research grants to further academic work on the topic. Second, they also intend to create forums for researchers, policymakers, and AI practitioners to come together for policy discussions. Finally, the company wants to scale their AI Economic Index to create longitudinal datasets that track AI's economic usage and impact.

Another way to think about this: is Anthropic trying to drive the policy discussion around jobs and AI to something that is a bit more evidence-based? Sarah Heck again commented: "I think the key goal is to figure out what is actually happening. If there is job loss, then we should convene a collective group of thinkers to talk about mitigation. If there will be huge GDP expansion, great. We should also convene policymakers to figure out what to do with that. I don't think any of this will be a monolith."

Interestingly, given how historically job losses from new technologies have tended to be blue-collar first, one of the interesting dimensions of AI is the first places that some of this is showing up seem to be in the industry that's creating the technology in the first place. Dee Das of Menlo Ventures wrote, "...there's a deep malaise in tech right now. New grads can't find jobs, big tech middle managers are trying to justify their existence."

Everyone not in AI wants to be in AI. Founders struggling with their startup for years see Roy rewrite the rules. In that one, he's referring to Roy from Clueley, who is basically going virality first, product second. Didi continues, combat security is at an all-time high with the meta offers. In other words, why am I working so hard? Tech, net of AI, is just not as sexy a job as it used to be 10 years ago.

He did caveat, of course, this is a massive generalization, but a majority of the conversations I've been having seem to reflect this. Still, while all of that might be true, Didi even points out that this is all just anecdotal right now. And when it comes to making policy, the discussion would likely be improved if we had real data about actual layoffs. For example, Microsoft slashed several thousand jobs earlier this year, largely from middle management and software engineering. They could be AI-related layoffs, or they could simply be a purge of excess headcount due to a deterioration in economic conditions. The

The distinction changes the policy implications dramatically. Now, if you are interested in these topics, Anthropic is fielding applications for grants to be issued in August and seeking submissions for symposiums to be held in Washington and Europe this fall. Next up, Meta is turning to private capital to fuel their AI build-out. The Financial Times reports that Meta is looking to raise $29 billion from private capital firms to fund their data center approach, with the reporting that they are in late-stage talks with

with some of the largest players in the space including Apollo Global Management, KKR, Brookfield, Carlyle, and PIMCO. Meta is apparently looking to raise $3 billion in equity alongside another $26 billion in debt. If it comes together, this will be one of the largest private capital raises in history, and sources say that Meta could even seek more capital once this round is closed. Meta is working with Morgan Stanley to structure the deal, and sources say there's been an exploration of ways to make the debt more easily tradable once it's issued.

Now, if you're wondering if it is usual for a big tech company to take on private capital in this way, the short answer is that no, it isn't. But the AI build-out is obviously no regular capital cycle. Meta recently upped this year's CapEx forecast to around $70 billion, a 10% increase on what was already the most capital-intensive year in the company's history.

The takeaway of this is that even the largest companies in the world are starting to run up against the limits of public equity to fund these mega projects. Now, of course, we've already seen AI startups take private capital to fund infrastructure. OpenAI and SoftBank's Stargate project is already courting private investors using a debt funding mechanism more typically used in the oil and gas industry. And more recently, XAI tapped Morgan Stanley to help them market $5 billion in debt.

On the other side of the deals, private capital firms like Apollo are increasingly pitching large companies on these deals as an alternative to corporate bonds. The private equity giants now own major insurers and annuity providers, so need a steady supply of high-quality loans to put on their books. As the deals are often structured as special purpose vehicles or joint ventures, they also serve to keep the debt off the balance sheets of the tech firms.

You remember on Friday when we talked about how Goldman Sachs has revised its opinion that we were at the end of the capex cycle to now they believe that we are actually in the middle of that cycle and this would seem to be evidence of that. Lastly today, a super cool one that I hope we will expand into a full episode at some point in the future. AI Video Platform Runway plans to launch a world-first generative video game platform.

This is a tiny first step in a vertical that seems entirely inevitable. That is, of course, AI games that can be generated on the fly. Alex Heath of The Verge was given access to the platform called Game Worlds, which Runway plans to make available as soon as this week.

He writes, "The consumer-facing product is currently quite barebones, with a chat interface that supports only text and image generation, but Runway says that generated video games are coming later this year. Runway is also in talks with gaming companies about both using its technology and accessing their datasets for training." So then, at the moment, the platform seems to be basically text-based adventure games with generated still images attached.

It's unclear how much of the gameplay is generated on the fly or if the experience is just about generating visuals and text to flesh out a pre-scripted storyline. Users will be able to create their own text and image game in the same style, the consistency of which is a whole technical challenge on its own. Regardless of where they are starting, Runway's ambition is very clear.

The company is already talking to TV and movie studios. The recently aired Amazon show House of David was made in part using Runway's technology. And they say that they're working with, quote, pretty much every major studio and most of the Fortune 100 companies. Now looking to take on the gaming world, CEO Cristobal Valenzuela said, If we can help a studio make a movie 40% faster, then we're probably going to be able to help developers of games make games faster. They're waking up and they're moving faster than I would say the studios were moving two years ago.

Look, I think, like I said, that this is a completely inevitable thing. Generative games are such a natural extension of many of the types of gameplay experiences that people already love now, where you can customize your experience and really choose your own adventure, that it's really just going to be constrained by what the technology can do.

One last interesting note from the Verge piece about this new launch. Heath writes, Naturally, I couldn't let Valenzuela get off our Zoom call without asking him about his recent acquisition talks with Zuckerberg. This is something that we reported on last week as well. Said Valenzuela, I think we have more interesting intellectual challenges being independent and remaining independent for now.

Well, for an update on all of that, let's jump over to the main episode. This episode is brought to you by Blitzy. If you're a technology leader, here's something that probably sounds familiar. Your organization's competitive edge is buried in legacy code that desperately needs modernization, but the resources required feel out of reach. That was the case for a global investment analysis firm. They needed to migrate 70,000 lines of complex MATLAB financial algorithms to Python. Algorithms that drive investment decisions for trillions in assets.

Their estimate? Months of high-cost specialized engineering work. Instead, they partnered with Blitze. Blitze's autonomous AI preserved mathematical precision and generated over 80% of the new codebase, completing the migration with just five days of engineering time. They cut the timeline by 95% and saved 880 engineering hours.

If your organization is facing similar modernization challenges, visit blitzy.com to schedule a consultation and discover how AI-powered development can transform your technical capabilities. Today's episode is brought to you by Vanta. In today's business landscape, businesses can't just claim security, they have to prove it. Achieving compliance with a framework like SOC 2, ISO 27001, HIPAA, GDPR, and more is how businesses can demonstrate strong security practices.

The problem is that navigating security and compliance is time-consuming and complicated. It can take months of work and use up valuable time and resources. Vanta makes it easy and faster by automating compliance across 35+ frameworks. It gets you audit-ready in weeks instead of months and saves you up to 85% of associated costs. In fact, a recent IDC whitepaper found that Vanta customers achieved $535,000 per year in benefits, and the platform pays for itself in just three months.

The proof is in the numbers. More than 10,000 global companies trust Vanta. For a limited time, listeners get $1,000 off at vanta.com slash nlw. That's v-a-n-t-a dot com slash nlw for $1,000 off. Today's episode is brought to you by Agency, an open-source collective for interagent collaboration.

Agents are, of course, the most important theme of the moment right now, not only on this show, but I think for businesses everywhere. And part of that is the expanded scope of what agents are starting to be able to do. While single agents can handle specific tasks, the real power comes when specialized agents collaborate to solve complex problems. However...

Right now, there is no standardized infrastructure for these agents to discover, communicate with, and work alongside one another. That's where Agency, spelled A-G-N-T-C-Y, comes in. Agency is an open-source collective building the Internet of Agents, a global collaboration layer where AI agents can work together. It will connect systems across vendors and frameworks, solving the biggest problems of discovery, interoperability, and scalability for enterprises.

With contributors like Cisco, Crew.ai, Langchain, and MongoDB, Agency is breaking down silos and building the future of interoperable AI. Shape the future of enterprise innovation. Visit agency.org to explore use cases now. That's A-G-N-T-C-Y dot org.

Today's episode is brought to you by Superintelligent, specifically agent readiness audits. Everyone is trying to figure out what agent use cases are going to be most impactful for their business, and the agent readiness audit is the fastest and best way to do that.

We use voice agents to interview your leadership and team and process all of that information to provide an agent readiness score, a set of insights around that score, and a set of highly actionable recommendations on both organizational gaps and high-value agent use cases that you should pursue. Once you've figured out the right use cases, you can use our marketplace to find the right vendors and partners. And what it all adds up to is a faster, better agent strategy.

Check it out at bsuper.ai or email agents at bsuper.ai to learn more. Welcome back to the AI Daily Brief. Today we are talking about some big updates in this incredible talent war story that has been emerging over the past few weeks that of course has meta and open AI at the center of it.

The specific context for today's story is that OpenAI has felt the need to actually recalibrate some compensation. They're very clearly acknowledging the pressure that Meta is putting on them. But the reason that I think that this matters for all of us here, and I just want to lay this out right up front, is that while it would be tempting to write this off as something that doesn't really matter to you until it manifests in the technology that actually gets built, there are a couple reasons why I don't think that's exactly true. The first is, with

With a technology this powerful and that has this much force to shape society, who builds it and who profits most from it will have an impact on how the whole thing plays out. Now, perhaps one could argue that everything is so likely to be commoditized that it doesn't really matter because all of these companies will win. And that's at least a reasonable take.

But we don't really know what AGI and superintelligence are going to look like, how fast they're going to accelerate once they're here. And so I think engaging with the question of who is bringing these things into the world is kind of an important conversation for all of us to be a part of.

The second reason why I think it matters for us to be paying attention to this is that if there is one watchword for AI in general, it is unprecedented. Everything about this technology, the speed with which it is developing, the rapidity with which it is being adopted, it is just unprecedented. And while yes, for decades we have highly rewarded, very valuable engineers, the changes that this talent war is bringing in terms of expectations for compensation and the value of top talent...

are unlike anything we've ever seen with talent before. Period. Full stop.

Now, by way of background, a few weeks ago, we started getting reporting that Mark Zuckerberg was personally out recruiting a new super intelligence team. It started around the time of the semi-acquisition type thing of scale, which brought their CEO, Alexander Wang, into the meta-fold to lead this new team. And when that story came to light, reporters also found that it was just the first move of this larger talent recruitment drive. Next, we got reports that there were these crazy offers being thrown around. Up

Up to $100 million signing bonuses, for example. That was the number that captured everyone's attention. As part of a strategy, I think, to combat that, Sam Altman seemed to confirm it in a way that I think was meant to paint both Zuckerberg and anyone who took that offer as completely mercenary. And going back to the point that I was making before, perhaps not the type of people that you want in charge of this incredible technology.

Now at the time, and again this was just a couple of weeks ago, Albin proudly stated that so far no one had taken that offer. But that wasn't going to be the case for long. Last week news broke that Zuckerberg had successfully poached effectively the entire OpenAI Zurich office. Hyperbolic Labs CTO Yuchen Jin wrote, "They don't even want to wait for their OpenAI equity to vest. Wonder how big meta signing bonus is."

Now, one of the people who was recruited, Lucas Beyer, did also say that the reports of $100 million signing bonuses were fake news. But still, people wondered. By latest count, Meta has successfully poached at least eight OpenAI researchers. And the number seems to be rising. Chubby on Twitter writes, even OpenAI employees confirmed the extent of the loss of researchers to Meta. They reposted something from Chang Liu who wrote,

Not too many people outside the company know how talented and hardcore they are. Such a huge loss for OpenAI, and I feel really disappointed that the leadership didn't keep them. Now the reason that Chubby posted that as a screenshot and not as a tweet is that the tweet itself was deleted very quickly.

And now we have this new report from Wired. It's called OpenAI Leadership Response to Meta Offers. Someone has broken into our home. Sean, better known as Swix, writes, Wow, this Meta OpenAI talent war has resulted in 1. Going from, oh, the good people don't leave, to someone has broken into our home and stolen something. 2. A one-week company shutdown. 3. Pivoting the product launch calendar and AGI race, especially when Stargate comes online. And 4. For some reason, Meta is ignoring Anthropic.

Now, Sean caveats, of course, if this source is true, but it definitely feels like something big is going on. The specific genesis of the Wired piece was that Mark Chen, chief research officer at OpenAI, sent what they called a, quote, forceful memo to staff on Saturday that was basically a battle declaration, a pledge to not go quietly into the good night and fight this war for top talent against Zuckerberg. Wrote Chen, I feel a visceral feeling right now as if someone has broken into our home and stolen something. Please trust that we haven't been sitting idly by.

So what does that mean that they aren't sitting idly by? Well, Chen said that he was working with Sam Altman, as well as other leaders, around the clock to talk to those with offers. He said, we've been more proactive than ever before. We're recalibrating comp and we're scoping our creative ways to recognize and reward top talent.

The note from Chen also included messages from seven other research leaders at the company, where they apparently wrote notes to staffers to try to get them to stay. "Row one leader to their staff. If they pressure you or make ridiculous, exploding offers, just tell them to back off. It's not nice to pressure people in potentially their most important decision. I'd like to be able to talk to you through it, and I know all about their offers." And yet complicating all of this is that apparently there is a broader sense of drain in OpenAI even outside this battle.

Wired, open AI is largely shutting down next week if the company tries to give employees time to recharge. Wrote another leader in Chen's memo, Meta knows we're taking this week to recharge and will take advantage of it. Try and pressure you to make decisions fast and in isolation. If you're feeling that pressure, don't be afraid to reach out. Me and Mark Chen are around and want to support you.

They also tried to refocus people on the big battle. We need to remain focused on the real prize of finding ways to compute into intelligence. This is the main quest, and it's important to remember that skirmishes with meta are the side quest. And so it's pretty clear that as much as Altman might be correct that OpenAI's culture of experimentation and what he and team members perceive as their increased chance of achieving AGI would be enough to make many of the best researchers stick around, at least for some, money is clearly trumping culture.

Now, speaking of culture, the AI cash splash is also starting to cause friction among Meta's existing staff. At an all-hands-on Thursday, Meta CTO Andrew Bosworth was asked about these supposed $100 million signing bonuses. Talking about Sam Altman's statements, he said, Sam is just being dishonest here. He's suggesting that we're doing this for every single person. Look, you guys, the market's hot, but it's not that hot.

Now, obviously this matters because alongside the newly recruited superintelligence team, Meta also has multiple divisions packed with AI researchers that are currently on payroll that they still need to keep happy. Bosworth continued, "...what Altman neglects to mention is that he's countering all these offers, creating a market for a very, very small number of people who are for senior, senior leadership roles. That is not the general thing that's happening in the AI space. And of course, he's not mentioning what the actual terms of these offers are. It's not a sign-on bonus. It's all these different things."

Bosworth then suggested that their efforts are bearing even more fruit than what's been reported so far, saying, "There are quite a few more in the pipeline that I can't announce or share right now." Taking one more dig at Sam Altman, he said, "Sam is known to exaggerate and in this case I know exactly why he's doing it, which is because we're succeeding at getting talent from OpenAI. He's not very happy about that." Now, I don't know. To be honest, that is not a particularly strong denial of the $100 million signing bonuses thing.

It sort of makes it seem like Bosworth is trying to say, well, guys, that's not actually just for signing bonuses rather than denying the totals. And also he's suggesting that it's for a smaller number of people that Sam Altman seems to be implying, but that still could be offers for a meaningful number of people at these incredibly inflated levels. Whatever the actual numbers are, it is at this point very clear that there is a very expensive and well-planned poaching campaign going on. The Wall Street Journal reported on the existence of something called The List,

a compilation of the most talented researchers and engineers that Zuckerberg put together personally over the past few months. It features PhDs from elite schools like Berkeley and Carnegie Mellon that typically have experience at either OpenAI or Google DeepMind. One recruit that spoke with the journal described Meta's goal as obtaining a transfusion from the country's top AI labs.

Interestingly, the article discussed how the tight-knit group of elite AI talent are in many cases swapping notes with each other, planning team-ups and also generally figuring out if they should work together at Meta. The recruitment of three researchers from OpenAI Zurich, for example, was reportedly a package deal that forced Meta to sign the entire team. Now, a less discussed angle of this story is that none of these researchers entered the field in hopes of cashing in on a nine-figure offer, at least not primarily.

When most of these folks were beginning their PhDs, AI was a niche and fairly unglamorous computer science topic that goes double for specialized areas like speech synthesis, where reports are that academic advisors would often try to steer students away from.

Alexei Efros, a Berkeley professor who has supervised numerous top AI researchers said, "For my students and postdocs, the objective was never to do hot things and become a millionaire. The objective is to try to solve cool, interesting, important unsolved problems." Now, of course, what has happened is that those cool and interesting problems are now some of the most valuable in the world, and Meta has started to pay accordingly.

And lastly, of course, how it might shake out the dynamics. When someone asked Juchin Jin, "How bad do you think this is going to hit OpenAI in future models?" They said, "I view OpenAI more as a product company rather than a model company."

Now, I don't think it's quite that dramatic yet. I think that there is a successful talent campaign being waged. It's repricing the cost of the most elite set of talent for everyone in the market. This is what markets do. Someone comes in, says, we think this thing that used to be worth that much is actually worth this much, gets a bunch of people to come over and join them because they're paying more than everyone else, and either they're wrong and the price resets to where it was, or everyone is forced to pay the same as they are, lest they get out-competed because the talent gravitational pull.

My guess is that no one's going to wait around to see if they're wrong or not, and everyone is just going to pay up. When it comes to OpenAI, however, and their destiny, I think it's way too early to count them out just because there's been a successful recruitment of so far less than a dozen senior researchers. And when it comes to this idea of OpenAI as a product company, not a model company, the reality is that they are doing that in addition, no matter what.

It's very clear if you are watching closely that OpenAI is at least a little concerned about the ultimate commoditization of models and is very determined to keep a close relationship with the end user. A recent story from the information, which we will most certainly get into later in the week, was titled "OpenAI takes a page from Palantir and doubles down on consulting services." That, I do not believe, is about them losing talent to Meta. It's about their assessment of what it looks like to compete in this very new era.

Still, it is interesting times. We are in a very up-in-the-air kind of moment, and I'm excited to be here covering it with you guys. For now, though, that is going to do it for today's AI Daily Brief. Appreciate you listening or watching, as always. And until next time, peace.