One of my favorite aspects of Microsoft for years now has been its exposure to AI. They are exposed through their own AI developments, but my favorite part of the Microsoft AI ecosystem is their access to OpenAI. Microsoft has invested $13 billion into OpenAI over the past five years, the companies are extremely close on AI research, they have a revenue sharing deal, and Satya Nadella kept Sam Altman in power as CEO when he was fired by the board. Microsoft got a non-voting board seat out of that arrangement, but I think they gave it up to avoid Federal scrutiny. In short, Microsoft is extremely close to the preeminent AI company.

Microsoft virtually possesses exclusivity with OpenAI, despite a push by Apple to join via an investment round. Microsoft’s AI future, which I believe is a highly profitable future, is heavily dependent on OpenAI. Microsoft’s internal AI tools are mediocre right now, so continuing to source OpenAI gear and get a small percentage of sales, as well as pirate a bit of their tech for Microsoft’s own AI, is important. When I heard that OpenAI was going for-profit and that there were mass quits, I was worried for Microsoft.

                  It is absolutely paramount above all else that OpenAI keeps its talent. Staying ahead in the AI race means keeping and finding the smartest, most knowledgeable people. OpenAI has failed to do the former. OpenAI had 11 cofounders, Sam is the only one remaining at the company. This worries me. Sam Altman is great, he is a smart dude, and I am relatively confident in his ability to lead, despite that board rebellion last year, but he has not built this company alone, and I worry that OpenAI has already/will lose its advantage.

                  Anthropic is an AI company that both Google and Amazon are bankrolling. DeepMind is Google’s internal AI wing. IBM Watson is sort of there. Cohere is growing. And Stability AI seems like a highly practical competitor. I have hated and will continue to hate on Elon, but as I have said before, he is great at growing projects to scale, and his xAI is a concern.

                  ChatGPT improvements are also becoming more difficult, as OpenAI looks for increasingly marginal gains. I know some people believe that something like Moore’s Law is applicable to AI and that the technology will exponentially improve, but I disagree. I think there is a hard ceiling on AI development and we have already reached 60% of it. The next 40% is the best, the world-changing AI, but it is also the hardest to get. ChatGPT-5, for example, has already been delayed so much, and all the while, OpenAI is shedding critical employees, who are going on to found their own projects or rake in lucrative contracts from the big boys like Google.

                  Ilya Sutskever was one of the OpenAI board members who fired Sam, and infamously/famously publicly apologized, before leading the charge to bring Sam back. OpenAI had the option to bring him back, and Ilya had expressed interest in returning to OpenAI in some capacity, but OpenAI executives decided not to bring Ilya back, a mistake. He has gone on to begin his own AI company because he is a super-talented AI-scientist, called Safe Superintelligence, which has already pulled in over $1 billion in funding.

John Schulman is another highly important co-founder who left the company. He soon found a contract at Anthropic. Jan Leike was OpenAI’s best research scientist, and he found a contract at Anthropic. On top of these moves, OpenAI’s chief technical officer, chief research officer, and vice president of research all resigned on the same day about a week back.

                  These extremely smart, extremely important people are not getting chewed up in the OpenAI ecosystem and leaving it all behind once they resign. These people are leaving OpenAI and taking their intelligence and experience elsewhere. In the short-term, maybe this lets Altman do his thing easier, but in the long-term, all of these departures will harm the company and will allow for competitors to surpass them in increasingly damaging ways.