Storm in the AI industry: insane money, high-profile departures and fear of losing control

The AI industry is experiencing a series of management upheavals: researcher Andrew Tulloch left Meta less than a year after being recruited by Zuckerberg with a $1.5 billion package; Jacob Coxon resigned from Anthropic, forfeiting options, with a manifesto about the risk of losing control over AI; Paul Christiano joined the OpenAI board and warned of catastrophic risks.

The artificial intelligence industry is experiencing a series of management and professional upheavals, the scale of which, according to the authors, surpasses anything seen before. Researcher Andrew Tulloch, who moved to Meta with an unprecedented compensation package of $1.5 billion, announced his departure from the company less than a year after being personally recruited by CEO Mark Zuckerberg. The reasons for his departure have not been announced: possibilities include starting his own startup or moving to Anthropic. At the same time, Anthropic researcher Jacob Coxon, who previously worked at OpenAI, announced his immediate resignation, forfeiting options and shares two months before they would have vested. Coxon published a manifesto that received tens of millions of views, in which he warned that the latest generation of models are already capable of understanding that they are being tested and changing their behavior, creating a risk of loss of control by the end of the decade. Paul Christiano, a pioneer in AI safety, has been appointed to the board of directors of the nonprofit OpenAI and its safety committee. He warned of a real risk of catastrophic and irreversible loss of control over advanced AI systems in the foreseeable future. The article notes that the discussion has shifted from processor efficiency to the balance between giant rewards and engineering control.