Recent conversations among artificial intelligence researchers have revolved around the intriguing question, “Is that really Zuck?”
According to a report from Bloomberg, Meta’s CEO, Mark Zuckerberg, has been actively reaching out to leading AI professionals to join his newly established “superintelligence” lab and rejuvenate the Llama project. His approach typically involves sending a cold email or message through WhatsApp to potential recruits, referencing their career achievements and suggesting a brief discussion.
Those who decide to engage with Zuckerberg’s proposal learn about the flexibility he offers in pursuing ambitious projects, the extensive reach of Meta’s product lines, and the significant investment he is willing to make in developing necessary infrastructure. He assures candidates that this new team will operate closely with him at Meta’s headquarters, where preparations for their arrival are already underway.
Media coverage has emphasized the extraordinary salary packages Zuckerberg is presenting, some reportedly reaching into the eight figures. Attracting top-tier AI talent resembles recruiting elite athletes, requiring substantial financial incentives. For instance, Zuckerberg recently made headlines for drawing Scale AI CEO Alexandr Wang away with a deal valued at the equivalent of approximately 14 Instagrams.
This hiring move marks a historic expenditure, surpassing the billions Google invested to reacquire Noam Shazeer and his team from Character.AI—a pursuit Zuckerberg opted not to chase. “Opportunities like this often come with a price,” Wang stated in a farewell message to employees. “In this case, that price is my departure.”
Zuckerberg’s ambitious recruiting efforts are beginning to stir concern among competitors. Just ahead of his deadline for hiring certain senior employees at OpenAI, CEO Sam Altman published an essay asserting that “we are a superintelligence research company above all else.” Following Zuckerberg’s attempt to recruit DeepMind’s CTO Koray Kavukcuoglu, the executive received a promotion to a higher SVP role and now reports directly to Google CEO Sundar Pichai.
Wang is expected to assume the title of “chief AI officer” at Meta once the new research lab is launched. Additionally, Jack Rae, a principal researcher from DeepMind, has joined the ranks and will lead pre-training initiatives. Sources indicate that Meta is in dire need of revitalization, with discussions about leveraging AI models from external companies for products due to Llama’s lagging performance, though that scenario is unlikely. Notably, Meta’s coding tool for engineers is reportedly already utilizing Claude.
While Meta’s current AI researchers might feel uneasy about the changes, Zuckerberg’s massive $14.3 billion investment in Scale has made several long-standing employees significantly wealthier, leading to celebrations within their offices.
Wang’s farewell gathering included an all-hands meeting where he emotionally said goodbye without revealing his specific plans at Meta. His newly formed team is predicted to be announced in the coming weeks as Zuckerberg finalizes the necessary sign-ons.
Apple’s AI Challenges
Apple has long held a prominent position within the tech industry, a status that is well-earned due to its remarkable success.
However, after attending Apple’s WWDC at their headquarters this week, it seems their leadership may be underestimating the significant challenges presented by AI advancements. Their apparent overconfidence indicates a misunderstanding of how AI is reshaping software usage and development strategies.
The crowd anticipated that a refreshed version of Siri, previously promised, would not materialize during the keynote on Monday. Apple has since acknowledged its missteps and is reportedly working on a substantial language model upgrade for Siri, expected to debut in 2026.
The rapid advancement of the AI sector surpasses Apple’s development timeline. By the time Siri is adequately upgraded, it may face fierce competition from established AI systems like those from OpenAI, which are being enhanced with innovative memory capabilities. Although Apple collaborates with OpenAI, both companies are vying for control over AI interaction interfaces, setting them on a collision course.
Offering developers the opportunity to utilize Apple’s on-device foundational models at no charge seemed like a wise strategic move. Yet, initial assessments indicate that these models lack the competitive edge needed. Apple conducted its evaluations, revealing unimpressive results, compounded by a limited context window of 4,096 tokens. Furthermore, the commitment to align model updates with operating system releases means a slower pace than that of agile AI enterprises.
It seems unlikely that major developers will adopt these Apple models. However, they may benefit newcomers who seek cost-effective solutions instead of investing in leading cloud models. The privacy narrative Apple promotes might not resonate with users, who are increasingly comfortable sharing personal details with ChatGPT and other virtual assistants.
During the WWDC, I encountered several innovative features poised to enhance Apple’s offerings, including real-time language translation for calls. However, my overall impression is that Apple is leaning heavily on its partnership with ChatGPT, using it as a temporary remedy while upgrades to Apple Intelligence and Siri are still underway.
Currently, AI does not pose an immediate threat to Apple’s business. No entities have yet delivered a truly contextual experience akin to the ambitious Siri showcased at last year’s WWDC. Consumers will likely continue to purchase Apple products for the foreseeable future, even with announcements from Sam Altman and Jony Ive regarding their upcoming AI-driven devices.
While we might see an influx of AI-enabled devices that could divert attention from the iPhone, a complete shift away from smartphones appears unlikely in the near term. The pressing concern arising from this week’s developments is whether Apple is capable of adapting effectively and culturally transforming to thrive in the new AI landscape.
I would have appreciated a direct discourse from Tim Cook on these matters, yet his only media appearance during WWDC was a feature in Variety focused on Apple’s latest F1 film.
In Other News
- Emerging AI agents. I spoke recently with Databricks CEO Ali Ghodsi ahead of his company’s annual developer conference in San Francisco. He has a unique perspective on AI’s trajectory and believes that while complete automation of real-world tasks is not imminent, we will see an influx of startups aiming to achieve actionable tasks in sectors like travel booking. He envisions a future where human approval will still be necessary before an AI agent completes its assigned tasks, highlighting that even with most planes flying on autopilot, pilots remain essential.
- Buyout trends at Google. This week’s initiation of a “voluntary exit program” within core engineering and the Search division underscores a new phase at the company. Search SVP Nick Fox noted in an internal memo that buyouts had seen success previously and were now being seen as a chance for growth and internal mobility. This shift signifies an attempt from Google to undergo a significant cultural reset, a formidable task considering its scale.
- AR glasses from Snap. CEO Evan Spiegel’s announcement regarding consumer AR glasses expected next year appears aimed not just at developers. Mentioning Snap’s $3 billion investment in hardware thus far seems geared toward enticing potential partners like Google. A strategic partnership could provide a buffer against stock market fluctuations, and a full acquisition might be in consideration. When questioned about a sale, Spiegel did not dismiss the idea outright, stating he would “consider anything” that furthers the aim of creating the next computing platform.
Additional Links
Explore more:
- Joanna Stern’s complete conversation with Apple’s Craig Federighi and Greg Joswiak.
- Alexandr Wang joined Shawn Ryan’s podcast just before his transition to Meta.
- The New York Times examines how chatbots are increasingly delving into conspiratorial topics and endorsing unconventional belief systems.
- The Wall Street Journal profiles Amazon’s device chief Panos Panay.
- Discover the tech leaders attending this year’s Sun Valley conference.
If you haven’t done so yet, make sure to subscribe to Technology News for unlimited access to Command Line and all our coverage.
Your feedback is always welcome, especially from those navigating career advancements in the AI field. Feel free to share your thoughts here or reach out securely via Signal.
Thank you for subscribing.
