AI leadership challenges, AI agents, and Microsoft-OpenAI tension

Welcome back to a new edition of AI Odyssey!

Mira Murati, former CTO of OpenAI, is reportedly raising over $100 million for her new AI startup, tapping into her industry influence and insider connections to create AI products with proprietary models.

^{Sponsored by Airtable}

It’s time to change the way we build digital products

Keeping busy is easy. Driving business impact is what matters. Consistently align your people to the most strategic priorities, discover product opportunities from deep customer insights, and gain total visibility on execution with Airtable ProductCentral.

Get a first look at Airtable’s powerful new solution, built to answer the needs of modern product teams.

Learn More

1: What OpenAI’s leadership team is missing

Greg Brockman, OpenAI’s co-founder and president, is returning after a sabbatical, but tensions could flare as he reunites with CEO Sam Altman. During Brockman’s absence, Altman took control of technical teams, and their differing leadership styles may create friction.

The bigger issue? OpenAI’s leadership turnover. The past year has seen an exodus of key executives, leaving staff anxious and outsiders puzzled. Many believe OpenAI needs a "Trillion Dollar Coach" figure like Bill Campbell, who famously helped tech giants like Google and Apple manage internal conflict and egos.

Airbnb CEO Brian Chesky played a similar role when he helped broker peace during last year’s board drama, but he’s got his hands full with challenges at Airbnb. CFO Sarah Friar, who’s been critical to OpenAI’s fundraising efforts, is gaining popularity with investors but isn’t exactly the mediator they need.

While OpenAI focuses on advancing artificial general intelligence, its real challenge may be dealing with human dynamics at the top.

2: The era of AI agents is just beginning

The news: Microsoft is gradually rolling out semi-autonomous AI agents for businesses. These tools, available through Copilot Studio, can handle tasks like sales leads and customer queries, offering customization options for users.

Why it matters: AI agents promise a boost in productivity by automating routine tasks. However, generative AI's unpredictability raises the risk of unintended actions, making human oversight essential to avoid costly mistakes.

The big picture: Companies like Microsoft are betting big on AI agents, but the real challenge is managing their unpredictability. By adding guardrails—such as limiting tasks and involving humans in complex decisions—they aim to minimize risks.

Looking ahead: As AI agents expand their role, success will hinge on balancing efficiency with oversight, ensuring automation doesn’t go off track.

3: Chinese AI firms get creative to cut costs amid U.S. chip curbs

The news: Chinese AI companies, including 01.ai, Alibaba, and ByteDance, are slashing costs to build competitive AI models, despite U.S. chip restrictions limiting access to cutting-edge technology.

Why it matters: Facing tighter budgets and limited access to high-end Nvidia chips, Chinese tech giants have managed to reduce "inference" costs—the price of running AI models—by over 90%, offering more affordable options than their U.S. counterparts.

The big picture: To cut costs, Chinese AI firms are using smaller data sets and hiring cheaper engineers, while adopting a "model-of-expert" approach—training multiple smaller neural networks rather than one large model. This method reduces the need for expensive computing power but increases complexity, as engineers must manage multiple networks.

Looking ahead: Despite challenges, China's AI companies are finding innovative ways to stay competitive, focusing on affordability rather than building the best models. As the AI race heats up, their low-cost strategies could reshape global AI competition.

4: Google moves Gemini app team to DeepMind

Why it matters: Google is shifting the team behind its generative AI Gemini app to DeepMind, marking a strategic integration into its core AI research unit. This move, announced in a memo from CEO Sundar Pichai, is part of broader organizational changes to streamline Google's AI development efforts.

Key changes:

• Gemini team, led by Sissie Hsiao, will now report to DeepMind CEO Demis Hassabis.

• Prabhakar Raghavan, head of knowledge and information products, steps into a new role as chief technologist, moving away from direct product oversight.

• Nick Fox, a 21-year Google veteran, will take on most of Raghavan's previous responsibilities, including overseeing products like Search, Ads, and Maps.

This reorg signals Google’s continued focus on AI innovation, consolidating efforts under its flagship DeepMind unit.

5: Microsoft and OpenAI’s partnership frays

The news: Microsoft and OpenAI’s once strong partnership is under strain as OpenAI pushes for more computing power and funding, while Microsoft grows wary of its dependence on the startup.

Why it matters: OpenAI, losing $5 billion this year, needs more resources, but Microsoft has started hedging its bets, hiring talent from competitors like Inflection and exploring alternatives.

The big picture: While OpenAI secured $6.6 billion in new funding, tensions remain. Microsoft agreed to reduce costs, but OpenAI has also turned to Oracle for additional computing power.

Looking ahead: If OpenAI achieves artificial general intelligence (AGI), Microsoft could lose access to its tech, signaling a potential split.

AI NOTES 🗒️

Watch: How AI is helping to decipher animal communication (ABC)

Extra read: How a Trump win could unleash dangerous AI (Wired)