Return of Sam

Five days that have shook the AI World

The last time I posted something on this blog, Sam Altman seemed to be ousted from the role of OpenAI CEO. For most companies and for most CEOs this would have been the definitive turn of events, and even “second acts” like that of the legendary Steve Jobs take years, if not decades, to play out. But OpenAI is not just another startup, Sam Altman is not just another CEO, and AI is not, by a very long shot, just another technology. So for several days after the fateful Friday firing, the situation had seemed to be in flux, Sam had revisited the OpenAI offices as guest in order to renegotiate his return, and seemingly all of us in tech were glued to any piece of news that could be gleaned from whatever sources we could get access to.

The only place where all of the most relevant actors in this drama were posting up-to-date information was X, which cemented its status as the platform for the raw breaking news updates, especially in tech, and even more so in AI.

That whole weekend and the following days I was not getting much sleep, which did not help my ongoing struggle to get better sleep. :/ I was far from the only one who did not get much sleep during that period - the CEO of the smart mattress company 8sleep tweeted that their data indicated 27% drop in hours of sleep in the entire San Francisco for all of their customers.

So after all back and forth (including probably one of the shortest nominal stints as a high-level Microsoft executive), Sam is back as OpenAI CEO, and all the other high-level OpenAI personages. The public display of loyalty to Sam by the vast majority of OpenAI employees, a true rarity these days for any tech company, also reinforced the notion that people who work there are uniquely driven by the high sense of mission that they are on.

Greg Brockman has tweeted several times since then about this sense of mission, and seems to be singularly refocused and driven to work even harder than before. There is considerable sense that the rank-and-file OpenAI employees also share the same drive.

To this day we still don’t know what exactly prompted Sam’s firing, and who were the main decision-makers behind the scenes. Everyone seems to be in agreement that there had not been any outright malfeasance. There is also a shared sense that something about the recent work at OpenAI prompted the sense that the company has seriously deviated from its open and safetist mission to develop AGI, but we don’t know any details. There have been news about a new search algorithm that some teams developed - dubbed Q* - which could help with the agentic aspects of the AI being developed, but my sense is that the importance of this algorithm has been blown out of proportion, mostly by the journalists looking for a sensationalist angle.

One thing that I am fairly confident about, though, is that this near-death experience at OpenAI has imbued all the stakeholders with a new sense of urgency. Whatever project everyone has been working on will be even more accelerated. Tomorrow is not guaranteed, and long term plans will be turned into the short term variety. In a few months, if not much sooner, we’ll probably see some really groundbreaking developments. The World of AI is anything but boring.