OpenAI in throes of executive exodus as three walk at once

> Marcus wrote: "GPT-5 hasn't dropped, Sora hasn't shipped, the company had an operating loss of $5 billion last year, there is no obvious moat, Meta is giving away similar software for free, many lawsuits pending.

> "Yet people are valuing this company at $150 billion dollars.

> "Absolutely insane. Investors shouldn't be pouring more money at higher valuations, they should be asking what is going on."

I've been saying for a while now that the lack of GPT-5 is a huge red flag for their future. They burned all the hype for GPT-5 on 4o, letting the media call their upcoming model GPT-5 for months before coming out with "4 but cheaper". o1 is impressive but again not a new generation—it's pretty clearly just the result of using 4o's cost savings and throwing tons of extra tokens at the problem behind the scenes in a technique that's easily replicable by the competition.

OpenAI has no moat and there has been no serious movement by governments to give them the moat they're so desperately lobbying for. Everything about their actions over the past year—the departures, the weird marketing tactics, the AGI hype, and the restructuring—says to me that Altman is trying to set up an exit before investors realize that the competition has already caught up and he has no plan to regain the lead.

> She said she is leaving "because I want to create the time and space to do my own exploration." Murati was joined by McGrew, who said: "It is time for me to take a break." Regarding his departure, Zoph stated: "This is a personal decision based on how I want to evolve the next phase of my career."

What a coincidence they all want to go on a break at the same time.

The Atlantic [0] says that it's clearly the Sam Show, if that was not crystal clear in the past. And what's the point of being an executive if one person makes all the decisions?

quote:

The departure of executives who were present at the time of the crisis suggests that Altman’s consolidation of power is nearing completion. Will this dramatically change what OpenAI is or how it operates? I don’t think so. For the first time, OpenAI’s public structure and leadership are simply honest reflections of what the company has been—in effect, the will of a single person. “Just: Sam.”

end quote

[0] https://archive.md/BzzYS

I think the data will show that it is going to take exponential costs in either train-time or inference-time for linear or even sub-linear improvements. We're scraping the bottom of the ice-cream container. I expect a lot of benchmark shenanigans to conceal this like what we saw with o1. Most use-cases won't tolerate using orders of magnitude more time and tokens for marginal improvements. These companies are still valuable but the multiples will need to be reassessed.

I am not convinced that going for profit is in the long term financial interest of the organization. Productizing will require diverting resources from foundational research to monetization. Moreover, the brain drain from this move is predictable because the top people in the field are largely wealthy enough that they don't need to choose where they work solely on compensation.

Even if they are motivated by compensation, they may be looking around and thinking both that Altman's approach is reckless and that it is a distraction from the core research. This would give them confidence that they can beat Altman in the long term by focusing on research and creating an environment that is most attractive to the best researchers in the field while he is busy chasing dollars today by scaling out what they already have rather than truly innovating. There is plenty of money sloshing around (see Ilya, e.g.) so it's not like they will struggle finding funding.

Now, it could be the case that there is some incredible insider tech that we don't know about publicly, but if they are way ahead of everyone it is hard to see from the public information. It certainly does not look good when the early employees who actually built the technology are leaving and the businessman who has a track record of lies, manipulation and a general lack of human empathy is consolidating power, abandoning the company's core mission and asking for huge amounts of funding along with a plan to extract an enormous amount of the world's natural resources to serve his profit margins.

In other words, I am quite suspicious that Sam Altman is a ruthless con man. That is what all the reporting I've seen around him most strongly suggests. I don't believe it is in the long term interest of an organization to be helmed by that kind of leader. I could be wrong, but I am still waiting for the people who have left to clarify that his leadership had nothing to do with their departures.

If it walks like a duck..,

Any other company seeing this kind of exodus would generate massive ridicule and concern.

But “AI” is supposedly the “next big thing” and the media and a lot of people with “get rich quick schemes” are trying to keep the hype alive.

The hype is clearly overblown and reality is setting in.

The bubble is about to burst.

And Tesla survived because the federal government bailed them out. Not because Merlin was some badass businessman.

> The nonprofit is core to our mission and will continue to exist.

Mere existence is the lowest of bars, kind of a funny way to clarify how core something is.

It's probably tied to fundraising. Execs probably have "key person" clauses and need to negotiate stock or secondaries as part of the deal.

This again proves that realistically we're hardware bound (waiting for nvidia to release bigger, more efficient and better hardware.

We won't see any improvements until the next generation of hardware comes out and allows us to do things that were computationally impossible before.

Yes, of course we can develop better transformers and improve reasoning capabilities, but that's not what I am talking about. I am mostly referencing the ability for newest "AI" hardware to multiply 256 matrices in a single clock cycle per core.

All this non-stop drama and Anthropic just quietly putters along in the background.

If OpenAI were anywhere near having achieved AGI, they would not have stepped back.

isn't this just Sam consolidating power

A rather banal explanation would be that OpenAI is a hot company, and it's execs likely receive very enticing offers from other companies or even potential investor backing.

I wonder if the humanity is catching up to the executives here realizing life is short, bullshit not worth it and they made enough to get out while they are young enough

More discussion re: Mira: https://news.ycombinator.com/item?id=41651038

> Jason Wong, Gartner analyst, told The Register: "It's clear with the departures of the co-founders, and high-profile engineering leaders, that OpenAI is being remade with Sam's vision. His manifesto and the shift to a for-profit entity also reinforces his vision for the business.

It seems like Altman was a poor choice to run an organization meant to benefit and protect humanity, as reports increasingly make him sound like a lying, manipulative sociopath (albeit a powerful, competent, and lucky one).

> "This could have significant impact on OpenAI's partnership with Microsoft, which clearly stated they view OpenAI as a competitor. Microsoft has already started to downplay the importance of OpenAI models in their overall AI strategy. For enterprises, uncertainty is not good for business and key tech investments like generative AI. Other frontier models – especially more open ones – have caught up to OpenAI, which will further influence decisions to derisk by moving away from OpenAI or spread their risk using other models."

Now that's some happy news. According to https://www.wheresyoured.at/subprimeai/, OpenAI is already losing absurd, unsustainable amounts of money even with massive discounts from Microsoft. I wonder what fun things could happen to OpenAI if Microsoft started charging their competitor full price?

If it were to close, who is well-positioned to acquire OpenAI's patents?

Can someone more knowledgeable tell me how to square this with the news that OpenAI is trying to spin up a for-profit arm? Are they afraid of layoffs or are they smelling things turning rotten?

I mean may be investors are just stupid which really makes me ponder how life would just be better if resources were allocated differently in society.

In the end, there will just be Sam Altman sitting in a grand chair with the pre-eminent sentient superhuman AI at his side and at his command. Think of the power he will have!

If the future unfolds where AI does really take off, his power may be unmatched by anyone in the past if he stays in control of it.

I think this is purposeful consolidation of power by Sam. He sees what is coming and the value of being in singularly control of it.

Feels like a movie script.

(I think he needs to acquire an android company next, AI needs at least some type of embodiment. Boston Dynamics?)

> OpenAI in throes of executive exodus as three walk at once

They feel the danger of the law. /s

Just like crypto, the AI hype is coming to an end.

What baffles me is how OpenAI keeps its doors open. They're paying Microsoft to be able to exist, plus their product requires huge amounts of electricity and infrastructure. OpenAI is unsustainable.