OpenAI started as a "countervailing force" to Google — did Elon Musk and Sam Altman torpedo DeepMind's plans to dictate AGI?

AGI (or artificial general intelligence) is becoming a buzzword that tech leaders blurt out when describing what the endgame in the AI race could look like. At this point, there seems to be a different meaning for the term emerging every day.

Your guess is as good as mine when it comes to identifying who will eventually emerge as the winner of the ever-competitive race, though AI safety researcher Roman Yampolskiy narrowed down the answer to the company with enough money to buy enough computing power to build AGI, eliminating the illusion of time as a limitation.

Now, former OpenAI researcher Daniel Kokotajlo has shared some interesting insights about an alternate and seemingly dangerous AI path. The executive claimed that DeepMind CEO Demis Hassabis had initially planned to build and rapidly develop towards the coveted AGI benchmark while leveraging Google's vast resources (via vitrupo on X).

While Microsoft's CEO Satya Nadella believes that Google already missed its opportunity in AI, Hassabis planned to ship AGI first, then leverage Google’s lead in the sector to fine-tune security guardrails.

However, the plan was "torpedoed" when Elon Musk, Sam Altman, and Ilya Sutskever joined forces and launched OpenAI.

Perhaps more concerning, Kokotajlo reveals that the executives launched OpenAI because they didn't trust DeepMind's Demis Hassabis to handle all the power that AGI entails responsibly, "when he was in charge of the project and all the fate of the world rested in his hands."

According to Kokotajlo:

"They wanted to create OpenAI to be this countervailing force that could do it right and distribute it to everybody and not concentrate power so much on Demis' hands."

The executive talked about leaked emails that came up in the lawsuit, which corroborate his claims and highlight the concerns among OpenAI's founding team about Demis becoming a "dictator" using AGI.

Interestingly, OpenAI has since lost the majority of its founding members, with a vast majority citing safety concerns and claims that the ChatGPT maker is prioritizing "shiny products" like AGI while safety processes and culture take a backseat.

Kokotajlo claimed that OpenAI is confident that it is well-equipped with technical talent to sort out alignment and safety issues compared to its rivals. Last year, OpenAI CEO Sam Altman boasted about his confidence in his team to build and develop AGI, further revealing that the company was shifting its goal post to superintelligence.