OpenAI, Microsoft, and the OpenOffspring
This is, as ever, a fantastic post and great commentary on the weekend’s events.
Satya and Microsoft have pulled of a god-mode corporate coup in getting Sam and Greg in, one large factor in this has no doubt been access to compute.
This is the major limiting factor in the industry right now and in the short term, I think it will be incredibly difficult for OpenOffspring to compete because compute is so hard to come by right now.
For me this was probably the main motivator behind Sam and Greg joining Microsoft - they knew it was the best option to get access to the compute they need. They just couldn’t do what they want to do on their own right now.
Ethan - you are a voice of balance and sanity in this rapidly evolving space Your academic focus, and your ability to parse through hype is very useful. Thanks for doing what you are doing!
Fantastic summary. I’ve been glued to my phone and laptop ever since this started. I am genuinely bummed and have a hard giving the board the benefit of the doubt. Meaning I don’t. There’a a dozen better ways to have done the same thing they did. There’s almost worse way. If they were worried about dangers of AI, can’t see how splintering and losing their staff has helped that. But maybe Microsoft was always the more responsible one.
A great analysis. Your comparisons to the Traitorous Eight, and the Paypal Mafia are especially apt. I can't help but wonder how an inveterate startup guy like Sam Altman will fare inside a behemoth like MSFT. If MSFT is smart, it will structure its relationship with Altman such that he has sufficiently free rein that he will not feel constrained by corporate bureaucracy. But I don't know if MSFT is capable of pulling that off.
One thing though, to note: Elon Musk did not found Tesla.
The collective intelligence of the broader AI ecosystem is applauding in standing ovation! Why? Because collective intelligence is our capacity to evolve towards higher order complexity and harmony through such evolutionary innovation mechanisms as variation-feedback-selection, differentiation-integration-transformation, and competition-cooperation-coopetition>
We'll see the blossoming of all three triplets through the OpenOffsping startups and the migration of OpenAI talents to other AI labs. The only dark cloud on the horizon is if Sam 's accelerationist hunger for compute blindsided him to the dangers of putting AGI in the hand of a corporation.
Well, both Sam and Greg commented on Nadella's tweet about their leading a new company in the MS ecosysytem, saying [for the benefit of all humanity] "the mission continues."
Following that, 90 mins ago, Nadella wrote:
"We’ve learned a lot over the years about how to give founders and innovators space to build independent identities and cultures within Microsoft, including GitHub, Mojang Studios, and LinkedIn, and I’m looking forward to having you do the same."
Let's hope so...
Thank you for the cogent analysis Ethan. Ostensibly everyone at OpenAI are in their 20s & 30s...one Silicon Valley trope they /didn't/ follow was to hire a seasoned senior exec à la Google, Netscape, etc. If there were someone with more life experience and a steady hand, they may have been able to steady the tiller before the ship sank.
Can you please say a little more about how you see start-ups succeeding, when the compute power is getting locked up by the largest companies, who are also pushing for regulation to make it even less accessible to others?
This is a nice article, well, written as usual. I appreciate the perspective on what Sam Altman’s Exodus implies for the future of AI. Thank you for this.
"Microsoft will be one beneficiary, but it will not be the only place OpenAI veterans go" - as of last night, that's the destination.
Good summary, though I think the assertion that it doesn't change the present is debatable. One of the clear issues with Gen AI uptake among enterprises is trust and this does not help one bit. IT heads who are already sceptical about the value etc will be further put off by this added confusion. This effect will be short term, say 6 months or so, until its clear what the new platforms are, but there is definitely going to be a short term impact. Who would commit to building anything other than a POC on Open AI right now?
If the primary responsibility of a board of directors is to protect the long-term interests of shareholders and stakeholders, then the OpenAI board's decision is a momentous miscalculation, especially in the long run, the kind of case study typically taught at HBS.
Very insightful. Thank you
Interesting! Curious to see how it develops!
I so appreciate your usual calm, rational, discussion of the drama and noise. It's no wonder those not in "Tech" think it's a Funhouse. No matter what the issues the handling of this is nonsensical.
This article urges me to make it a topic for my students in the Emerging Technologies unit that I teach at the university, as it is truly an article that needs analysis and study.
I am truly grateful for this wonderful article, Ethan, and how you analyze the events because of your intelligence, which I find to be a guide to many of the questions that run through my head. The points you mentioned actually raise many future questions, but I certainly agree with you that the impact of generative artificial intelligence will continue with us for a long time and affect and disrupt our lives in both a positive and negative way. We are in a transitional phase to accept the new product produced by OpenAI, and I was very saddened by Sam Altman’s departure from the company. This raises many concerns about upcoming companies, but it may be a good thing for humanity, but I hope that Sam Altman does not join Microsoft, and joins companies with higher accuracy and quality, such as Google or Apple.
Interesting preview of the future: The board representing the humans trying to pull the plug or hit the off switch, but the super-intelligent AI won't allow itself to be shutdown...