Normally I am the cynic but this time I’m seeing a potential win-win here. Altman uses his talent to recruit and drive forward a brilliant product focused AI. OpenAI gets to refocus on deep research and safety.
Put aside cynicism and consider Nadella is looking to create the best of all worlds for all parties. This might just be it.
All of the product focused engineering peeps have a great place to flock to. Those who believe in the original charter of OpenAI can get back to work on the things that brought them to the company in the first place.
Big props to Nadella. He also heads off a bloodbath in the market tomorrow. So big props to Altman too for his loyalty. By backing MS instead of starting something brand new he is showing massive support for Nadella.
I get that funny money startup equity evaporates all the time, but usually the board doesn’t deliberately send the equity to zero. Paying someone in an asset you’re intentionally going to intentionally devalue seems like fraud in spirit if not in law.
And if you separate out the products from OpenAI, that leaves the question of how an organization with extremely high compute and human capital costs can sustain itself.
Can OpenAI find more billionaire benefactors to support it so that it can return to its old operating model?
For all we know, OpenAI may actually achieve AGI, and Microsoft will still want a front row seat in case that happens.
I personally expect the chat.openai.com site to just become a redirect to copilot.microsoft.com.
For one, I'm not sure Sam Altman will tolerate MS bureaucracy for very long.
But secondly, the new MS-AI entity can't presumably just take from OpenAI what they did there, they need to make it again.
This takes a lot of resources (that MS has) but also a lot of time to provide feedback to the models; also, copyright issues regarding source materials are more sensitive today, and people are more attuned to them: Microsoft will have a harder time playing fast and lose with that today, than OpenAI 8 years ago.
Or, Sam at MS becomes OpenAI biggest customer? But in that case, what are all those researchers and top scientists that followed him there, going to do?
Interesting times in any case.
The dataset is more challenging, but here msft can help - since they have bing and github as well. So they might be able to make few shortcuts here.
The most time consuming part is compute, but here again msft has the compute.
Will they beat chat-gpt 4 in a year? Guess no. But they will come very close to it and maybe it would not matter that much if you focus on the product.
MSR leadership is probably a little shaken at the moment.
Not doing that would be participating in illegal wage suppression. I'm not sure how following the law means OpenAI and MSFT can't continue a business relationship.
Exhibit A: this weekend, lol.
Part of me thinks that Nadella, having already demonstrated his mastery over all his competitor CEOs with one deft move after another over the past few years, took this on because he needed a new challenge.
I'd wager Altman will either get sidelined and pushed out, or become Nadella's successor, over the course of the next decade or so.
It's an interesting time!
But I think it is probably sufficient to point to the language in the contracts granting illiquid equity instruments that explicitly say that the grantee should not have any expectation of a return.
But I think this is an actual problem with the legal structure of how our industry is financed! But it's not clear to me what a good solution would even be. Without the ability to compensate people with lottery tickets, it would just be even more irrational for anyone to work anywhere besides the big public companies with liquid stock. And that would be a real shame.
What I meant is, most likely assuming that you are using pytorch / jax you could code down the model pretty fast. Just compare it to llama, sure it is far behind, but the llama model is under 1000 lines of code and pretty good.
There is tons of work, for the training, infra, preparing the data and so on. That would result guess in millions lines of code. But the core ideas and the model are likely thin I would argue. So that is my point.