Justin Sullivan/Getty
One of Google's top product leaders believes a new approach to AI could offer a "straight shot" to artificial superintelligence, which promises abilities far more advanced than humans in all areas.
Logan Kilpatrick, product manager for Google's AI Studio, wrote on X that he believes that aiming straight for artificial superintelligence, or ASI, without focusing on intermediate milestones is "looking more and more probable by the month."
He said he believes the success so far of scaling up test-time compute— the moment when an AI model must perform a task or answer a question — is a "good indication" that a direct path to ASI might be possible.
To date, a lot of conversation has focused on when or how artificial general intelligence, in which a model matches or surpasses humans on a broad range of tasks, will be reached. Kilpatrick wrote that he still believes we will get to AGI, but "it's likely going to just look a lot like a product release" rather than one big explosive moment.
With increased evidence that pre-training AI has plateaued, companies and researchers have explored new ways to improve AI models, such as having them "think" through problems in a similar way to humans. Both Google and OpenAI have recently unveiled new models with improved reasoning abilities.
Kilpatrick wrote on Monday that he believes OpenAI cofounder and former chief scientist Ilya Sutskever may have seen "early signs" of the potential of scaling test-time compute.
Sutskever, who has publicly claimed the industry is running out of data to train AI models with, left OpenAI earlier this year to launch his own startup named Safe Superintelligence. "We will pursue safe superintelligence in a straight shot, with one focus, one goal, and one product," Sutskever wrote on X in May.
Kilpatrick said he used to believe Sutskever's method would be a mistake, but now thinks it may just work.
Kilpatrick's opinion will carry some weight. He previously led developer relations at OpenAI and joined Google earlier this year in a move seen by many as a boon for the search giant. One person previously described him as a "secret weapon" for Google in the AI race.
When asked on X about the advantages of taking a more iterative approach to superintelligence, Kilpatrick hedged his response.
"I'm more bullish on iterating than I am straight shot," he wrote, "but the latter just might work."
Got more insight to share? You can reach the reporter Hugh Langley via the encrypted messaging app Signal (+1-628-228-1836) or email (hlangley@businessinsider.com).
