Capuchin monkeys explore decorations filled with their favorite treats at Edinburgh Zoo.
PA Images/Reuters
Mark Zuckerberg has been offering $100 million+ packages to lure AI talent from rivals.
Some AI experts have rebuffed Zuck, citing loyalty to their company's mission.
I know what my father-in-law would say about this. It includes the letters B and S.
Since my dad died, I've grown close to my father-in-law Bill. He has a mix of joy and realism I admire, and his advice tends to stick.
One of his favorites: "We're all prostitutes when it comes to work." He's from an older generation, so I'll swap in "sex workers" as the appropriate phrase here. What he means is that we work mainly for money.
That's been on my mind as the AI talent war heats up. Mark Zuckerberg has been offering $100 million-plus packages to lure AI researchers and engineers from frontier labs and Big Tech rivals. Some have refused, citing loyalty to their company's mission.
"They are trying to buy something that cannot be bought. And that is alignment with the mission," Anthropic CEO Dario Amodei said recently, noting his top talent has stayed despite Meta's offers. Staff at Thinking Machines Lab have also turned Zuck down, either over leadership concerns or mission loyalty, according to Wired.
My father-in-law would say a phrase that includes the letters B and S. According to the gospel of Bill, this is all about getting paid, as usual.
A Business Insider scoop backs this up: This week, Charles Rollet reported Zuck's recruiting drive has created tension among existing Meta AI experts, who resent newcomers getting higher pay in its new Superintelligence team. That's made some easier to poach. xAI has nabbed several, and Microsoft has a Meta talent wish list. On August 6, Laurens van der Maaten, a top Meta scientist, announced he was joining Anthropic.
Reacting on X, former Meta engineering director Erik Meijer wrote: "Every action has a reaction; the unintended side effects of creating a SI team," referring to the Superintelligence group. When asked for comment, he shared a YouTube clip of an experiment in which two monkeys performing the same task were given different rewards. The one that got a less tasty treat hurled it back and angrily shook its cage.
Human treatments are often tested first on laboratory monkeys.
Mladen Antonov/Getty Images
If we're all metaphorical monkeys or sex workers, what about the "mission-driven" folks staying put? One possible explanation: equity.
Many engineers and researchers get stock in their startups, and these awards typically vest over several years. If you're at a hot AI lab, your unvested equity has probably soared in value lately, or there's a chance it could.
For instance, Anthropic could be worth $170 billion soon, up from about $4 billion two years ago. If you got equity back then and you're waiting for it to vest, there's no way you're leaving right now.
No surprise: most folks are staying at Anthropic. Until that equity vests, anyway.
I want to hear back from AI experts who are getting big offers. Are you mission-driven and staying put, or will you take the $$$ like most of us monkeys? Let me know: [email protected].
Is it just me, or does this summer feel super busy? New AI models are launching all the time. M&A dealmaking is unusually active. Funding rounds are getting done. How's your summer going? Let me know (if you have time!).
Apple CEO Tim Cook at the White House with President Donald Trump
Jonathan Ernst/REUTERS
Who's on Apple's AI shopping list?
If you're a Big Tech CEO, what's the worst way to spend a summer day? Attending a White House event as President Donald Trump rants about Jeffrey Epstein coverage while you stand there, waiting. That's gotta be up there.
I'm guessing Tim Cook had better things to do than go through this ordeal on Wednesday. He was there to stop Trump from crushing the iPhone with bigly tariffs. Much higher on his to-do list: getting Apple back in the AI game. An AI update to Siri is delayed, and the company has lost AI talent to rivals.
What happens when a huge, cash-rich company has its back against the wall? One common outcome is a flashy, strategic acquisition. These deals make a statement, and (if they work) they can catapult a company into new sectors or technologies by quickly amassing talent and intellectual property, along with new users and customers.
On a recent call with analysts, Cook dropped a major hint that this is what Apple might do in AI. "We're very open to M&A that accelerates our road map. We are not stuck on a certain size company," he said.
Apple's biggest acquisition was $3 billion, so I took Cook's "size" comment as a particularly strong hint that the company could go large here.
So, which AI startups could be on Apple's shopping list? Ben Bergman, Rebecca Torrence, and I spent this week asking bankers, venture capitalists, and analysts which M&A targets might make sense. Check out the full list below, but one that came up several times was Thinking Machines Lab, a startup run by former OpenAI CTO Mira Murati.
"It would be game-changing for Apple to partner with an independent, leading AI lab like Thinking Machines," said Sarah Guo, a leading venture capitalist who's backed many startups in the field, including Murati's company. "They have massive threats and opportunities across the Apple experience."
Radically improving Siri experiences with Apple's mobile data, using memory and personalization, is a huge opportunity, even if Siri has failed to keep up with the breakneck pace of AI capabilities to date, Guo noted, while stressing that Apple should partner with Thinking Machines, rather than buy the startup.
You know what they say in Silicon Valley, though. Partnerships can be a prelude to a deal.
My take on who's up and down, including updates on tech jobs and compensation.
UP: You know who really knows how to take a victory lap? Palantir CEO Alex Karp. Shares of the defense tech company surged this week after it smashed through Wall Street expectations. The stock is up almost 600% in the past year.
DOWN: Airbnb slumped this week after issuing guidance that disappointed Wall Street. This stock has lost at least 15% since Brian Chesky took the company public in late 2020. So much for "Founder Mode."
TECH JOB UPDATE:
In BCG's 2025 "AI at Work" survey, employees and managers showed a significant divide in how generative AI is integrated into their work. While 85% of leaders and 78% of managers use GenAI regularly, only 51% of frontline employees reported similar use.
Concerns about job displacement due to AI remain high. However, this also differs depending on seniority. Forty-three percent of managers and leaders in BCG's survey expect their jobs will certainly or probably disappear in the next decade. For frontline workers, that number was notably lower at 36%, BCG found.
From the group chat
Other Big Tech stories I found on the interwebs:
OpenAI just gave out juicy bonuses to fend off recruiters (The Information)
OpenAI can probably afford it. The startup could be worth $500 billion in the secondary market (Bloomberg)
AI labs use everyone else's data without permission, but they get grumpy when it happens to them (Wired)
Uber under-reported sexual assault and misconduct complaints (The New York Times)
Q: What were the main differences between Study Mode and the main ChatGPT tool?
Study Mode felt more proactive and conversational. When I asked if I should buy a car and briefly described my life, it introduced the concept of "being car poor" without me asking. Regular ChatGPT doesn't do that. Study Mode also gave concise bullet points instead of big paragraphs, which made things clearer and left less room for error. It also asked how I felt about the topic, not just practical things like budget or commute. For example, it asked if I'd feel anxious without a car or if I'd miss my family. When I mentioned my Uber and grocery spending, it offered a "budgeting exercise," like a thought experiment. Traditional ChatGPT would just crunch numbers if you gave it a figure.
Q: What tasks is Study Mode good for, versus the main ChatGPT?
Regular ChatGPT is fine for tasks like summarizing transcripts or proofreading a cover letter. But Study Mode shines when you want a complicated subject explained clearly, with relevant new concepts, pros and cons, and exploratory thinking. It doesn't give you a final product; it helps you create your own.
Q: Is Study Mode better for student learning?
Definitely. It explains concepts deeply and visually. Some charts reminded me of my old IB economics textbook. It balances ideal advice (like saving 20% of income) with realistic data (most Americans save 8% or less). It also has features like quizzes I haven't tried yet, which could be great for learning Spanish or exam prep. But it's no substitute for a human teacher's emotional connection.
Q: Any tips for Tech Memo readers using Study Mode?
Be open-ended. Don't just ask for facts — ask why something works, what alternatives exist, or how designs function. Like a good classroom, it's about curiosity.
User feedback
I love hearing from readers. What do you want to see more of, or less of?
Specifically, though: This week, I want to hear back about how work has interrupted your summer chillin' plans. The most gruesome stories = the best. Let me know: [email protected]
Sign up for BI's Tech Memo newsletter here. Reach out to me via email at [email protected].
Login
