AI Industry Drama: OpenAI's Controversial Rehiring and the Race for AI Talent

The artificial intelligence industry continues to be marked by high-profile personnel moves and corporate drama, as highlighted by recent events involving OpenAI and Thinking Machines Lab.

The OpenAI-Thinking Machines Controversy

OpenAI recently announced the rehiring of Barret Zoph and Luke Metz, cofounders who had previously left to join Mira Murati’s AI startup, Thinking Machines Lab. This move has sparked controversy with conflicting narratives emerging:

According to sources, Thinking Machines leadership believed Zoph was involved in serious misconduct while at the company.
Murati reportedly fired Zoph before learning about his return to OpenAI.
Thinking Machines raised concerns about whether confidential information was shared with competitors.
OpenAI’s CEO of applications, Fidji Simo, claimed the rehires had been planned for weeks and that OpenAI doesn’t share concerns about Zoph’s ethics.

Beyond Zoph and Metz, another researcher, Sam Schoenholz, is also rejoining OpenAI, with at least two more Thinking Machines employees expected to follow suit.

A Pattern of Industry Turmoil

This incident adds to a growing pattern of drama in the AI industry, reminiscent of OpenAI’s brief ouster of Sam Altman in 2023 (known internally as “the blip”). Other major AI labs have also experienced high-profile departures, including xAI’s Igor Babuschkin, Safe Superintelligence’s Daniel Gross, and Meta’s Yann LeCun.

Many AI researchers, who began their careers before ChatGPT’s breakout success, appear surprised by the constant scrutiny their industry now faces.

AI Labs’ Pursuit of Work Automation

Beyond the personnel drama, AI labs are intensifying efforts to create AI agents capable of performing economically valuable work:

Companies like OpenAI are collecting real-world work examples from professionals at prestigious firms.
Data suppliers such as Mercor, Handshake, Surge, and Turing are paying up to $100 per hour to contract top talent for AI labs.
Labs are creating “environments” – simulations that teach AI agents how to use enterprise software applications.
The goal is to train AI to execute office tasks across industries including legal, healthcare, consulting, and banking.

Recent improvements in AI agents, like Claude Code, suggest progress is being made, though questions remain about whether these agents can consistently perform complex tasks with the required accuracy.

The Stakes

With billion-dollar funding rounds becoming common in the AI industry and the technology contributing significantly to economic growth, the intense competition for top talent and technological breakthroughs shows no signs of abating.