Sometimes a comeback story starts with a fumble.
A former top OpenAI researcher said Google’s AI renaissance is as much about OpenAI’s missteps as it is about what the search giant got right.
“Personally, what I think you should consider Google’s comeback, I think it’s OpenAI’s fumble,” Jerry Tworek, a former VP of research at OpenAI, said on a Wednesday episode of Ashlee Vance’s “Core Memory” podcast.
Tworek, who spent almost seven years at OpenAI, said earlier this month that he left the startup “to try to explore types of research that are hard to do at OpenAI.”
OpenAI CEO Sam Altman declared a “Code Red” in December amid increasing competition from Google. The tech giant received wide praise across the industry for the capabilities of its Gemini 3 AI model, which some observers said had surpassed ChatGPT.
While declining to detail what he described as OpenAI’s missteps, Tworek said that the pioneering AI company should never have lost the lead it established with the release of ChatGPT in 2022.
“If you are a company that is ahead and has all the advantages that OpenAI has you should always stay ahead,” he said.
Overall, Tworek said, “Google did a lot of things right.”
“Very clearly, Google started treating seriously at that moment, training large language models and, like, through OpenAI fumbling its lead, they are very, very close now in capability and in terms of models trained,” he said, adding that the whole industry began to up its investment in AI when OpenAI showed ChatGPT could generate revenue.
As for OpenAI, Tworek said that the sheer toll of the AI race has led the non-profit-research lab-turned-public-benefit-corporation to place less of an emphasis on risky research that may not yield results. A spokesperson for OpenAI did not respond to Business Insider’s request for comment.
“There are multiple aspects of certain things that are just hard to do in a company that has to compete in an extremely, extremely brutal and demanding race for having the best AI model in the world right now,” he said. “One dynamic is there is naturally how much willingness of risks companies are willing to take from the perspective of trying to not fall behind.”
Tworek said “all major AI companies” are facing pressure to show user growth and pay for GPUs while simultaneously competing to be the best available model.
“That does affect somehow your appetite for risk that you are willing to take,” he said.
Do you work at OpenAI or Google? Contact the reporter from a non-work email and device at bgriffiths@businessinsider.com
