Ex-OpenAI board member provides her first detailed account of CEO ouster | Hacker News Reader

Ex-OpenAI board member provides her first detailed account of CEO ouster

reuters.com

111 pointstestrun47 comments

Discussed today:

https://news.ycombinator.com/item?id=40506582 - 427 comments

The link you shared isn't on the front page now, though :(.

If I was at OpenAI making stupid money and I had to deal with a psycho boss once in a while, I’d probably be fine with it.

I think most people would, as was evident by the amount of people who backed him to come back as the CEO.

Not that every one of those people knew this, but there certainly were people who did.

From the article:

> On the podcast, Toner attributed Altman's swift return to employees being told that the company would collapse without him.

> Additionally, once a potential return seemed likely, employees feared retaliation from Altman if they did not support him, she said.

> employees being told that the company would collapse without him.

These employees are smart enough to decide for themselves how critical they believe Altman is.

I also get the vibe from her that she believes those employees are dumb sheep who can't decide for themselves who their lead should be.

I only listened to parts of it but I did not get that vibe. For me it sounds like there was a lot of pressure applied. Particularly given the secondaries it does not at all surprise me that it was a very likely outcome.

If a leader does not want to step down and the leader has power, that leader can wreck chaos where the rational move is to leave the leader in place. I don’t know enough about OpenAI or Sam Altman to judge the situation but given the absolutely chaotic situation when it unfolded it sounded very rational to restore the prior status quo.

Well she thinks the same of the general population, that is that they would believe everything generated by AI

> those employees are dumb sheep who can't decide for themselves

The point is that they are smart because they know not to put themselves at risk with actions that would not lead to get rid of a threat.

> (...) decide for themselves who their lead should be.

They didn't had a vote or a choice. People living under a dictatorship also don't decide who their leader is when the dictator prompts them to show their support.

You somehow missed the whole point, which is that once retaliation was extremely likely these employees decided to not put themselves at risk.

This is how dictators get their 99% approval rating, and why democracies enshrine the right to vote in secrecy.

Crowd dynamics and snap decisions are very different from individuals and carefully considering all aspects before jumping in.

> Altman's swift return to employees being told that the company would collapse without him.

Regardless of who said/who thought that... Why would it?

Engineers need to take a good look at the mirror and stop idolizing "visionary jerk" figures.

Um...because Microsoft offered to take on all staff at existing pay rates if OpenAI imploded? That was the pivotal point IIRC.

Altman impersonating Roko's Basilisk is very much on-point for an AI company.

Huh, its Roko's Basilisk all along

I think it’s complicated and more tied in with your personal ambitions and passions than money. I imagine that most people making stupid money at the forefront of LLM development are probably very passionate about their work. I imagine that will make the career/person side of things much more intertwined than say… making stupid money working in finance.

It’s fairly easy to deal with bad upper management, poor organisational decision making, political infighting, silly bureaucracy, pseudo-work and a range of other nonsense when you’re mostly in it for the tech and go home at 4pm (or whenever you go home in your country). It’s quite another thing if you actually care about the mission.

OpenAI dual structure also makes things super complicated and adds gas to the fire. Like Altman doesn't own anything but has some weird VC fund, and all fuckery around employee compensation.

Past a certain point, money doesn't matter to people of principle.

Surely you mean principles dont matter to people with money?

"If there's a steady paycheck in it, I'll believe anything you say." -Winston Zeddemore

I think you underestimate the power of cults of personality and the extreme tech bubble people in Silicon Valley live in. People who are backing these individuals genuinely want to work with them, be around them, and most are hoping to become them one day. The percentage of those who deal with their shenanigans just for the money is likely very low.

you don't threaten to walk out in retaliation while vague-tweeting heart emojis because you think your CEO is a mere sociopath, IMO you do it because A, you won't build generational wealth without him, B, you feel there's an existential risk to humanity under any other leadership or C, you're a pragmatist whose social rank is proportional to how enthusiastic you are about the emperors' fashion

(OpenAI feels cultish to me as an outsider, some north korea shit going on there, us against the world)

Hm.

I can see how it looks cultish. I know people who work there, and I know someone who managed to escape (with outside assistance) from scientology, and… if they're a cult, it's one about how dangerous AI could be, not how great Altman is.

But both of these examples are mere anecdotes; I have no formal qualifications to discuss cults, and contact with only two current staff members.

I'm sure there is a truth to it. But as soon as Sam Altman & Elon musk get mentioned this entire forum gets highjacked by hating on them instead of having an interesting discussion.

With all the negative press Altman is getting, I wonder if he’ll step down at some point to appease public opinion and run things from the background instead. Seems somewhat reasonable but unlikely with his ego.

What makes you think Altman cares what anyone thinks?

It seems pretty obvious that he's a dark triad type personality and will keep going as long as possible.

> It seems pretty obvious that he's a dark triad type personality

The obvious moustache twirling villains mostly end up in prison almost immediately, precisely because they're so obviously bad.

Subtle and careful ones, that normal people don't notice are being manipulative, are generally the ones who get furthest — the darkness isn't shown to those who must not see it.

This is why so many in the US blame all of Congress except their own representative.

1. You assume that what press reports and public opinion is the same thing. It's not.

2. You assume majority of the public actually has an opinion or time to think about whether Sam should run OpenAI. They don't.

3. You assume public opinion is an accurate metric of justice. It's not. We have judicial system for that.

4. You assume Sam makes decisions through his ego and not any other system. You don't know that. Your ego thinks it's true. But honestly you have zero information unless you are really close to Sam in that case you have no reason to make this comment.

I think you've actually made four assumptions there.

I am a member of the public, and I consider Sam Altman power hungry, egoistical and narcissistic.

My opinion has been formed through accounts in the press.

I think you missed the majory of public part.

Interesting. So you make decisions based on what the press tells you? Do you believe the news the press choose to report and not to is based on what they think is good for you? Do you believe the press actually can know what is good for you?

The press is the press and my opinions are my own. The press can be right or wrong about Altman and that doesn’t affect my opinions. I believe, regardless of the press, that Altman is operating for profit (as self interest dictates he should) and I politely disagree on his stance for OpenAI. Just because some of the press agrees with me doesn’t necessarily mean I’m wrong.

I think you misspelt majority.

How is that relevant? You understood the context and instead choose to focus on the misspell.

Because if you're trying to educate people, it's worth trying to get the basics correct first.

I am not trying to educate anyone. I am just asking questions. If you think I am trying to educate you that's your ego again trying to make you miss the point being conveyed here.

You initially made four assertions.

They were most certainly not questions.

Be on your way, and have a good day.

That is true. I guess my ego fooled me as well.

Listening to him on the lex Friedman podcast, gave me funky vibes about his motivations.

I believe everything she said about Sam. I don't doubt that he exaggerated how robust their safety processes were, or announced products without telling them first, or used company resources to benefit himself personally while denying he had a financial stake in the company.

I honestly believe that employees did talk about his toxic behavior, the role of a successful startup CEO can be achieved with surprisingly little people management skills.

But the board handled this horribly; it's unsurprising they lost the fight, and frankly, listening to her podcast, the main question I have is why she was on the board in the first place. The level of articulation, the weak framing of events, etc, are what I'd expect to hear out of a middle-level HR rep, not someone on the board of a billion+ dollar company.

She got played from day one and is mad because she was the last one to figure it out. Sam brought her on for academic credibility and she didn’t realize how powerless the role was until she made the mistake of attempting to flex that power. She was out maneuvered at every step and had zero business playing in that space, hard elbows are going to be thrown and this now is how she’s dealing with it.

This interview is exactly why she never should have been in that role to begin with.

It's surprisingly a bad look for her. Plus all the criticism to AI itself. She seems set on slowing down AI development, which is clearly at odds with what the company is doing. Why would she event be involved in this company let alone in the board.

> But the board handled this horribly; it's unsurprising they lost the fight, and frankly, listening to her podcast, the main question I have is why she was on the board in the first place. The level of articulation, the weak framing of events, etc, are what I'd expect to hear out of a middle-level HR rep, not someone on the board of a billion+ dollar company.

Your comment sounds more like ad-hominem and character assassination than an actual analysis of the accusations.

Only recently did OpenAI became a billion dollar company. A few years ago it was scrambling for financial support. Nevertheless this is immaterial to the fact that there's a solid accusation of a toxic work culture and shady practices that motivated firing the CEO.

I don't think that paragraph was meant to be an analysis of the accusations themselves, merely suggesting why the messengers fell on their own swords.

I don't think that complaining how articulate a board member was has any relevance to the outcome of a vote. Board members are there to decide, not to be professional entertainers or politicians.

You have a board member mentioning shady practices, lack of accountability, and toxic work culture, and you decided to ignore all that and instead attack the board by focusing on inane details such as how articulate you feel some members were.

> I don't think that complaining how articulate a board member was has any relevance to the outcome of a vote

The lack of articulation didn't have any relevance to the outcome of the vote.

The lack of articulation meant the vote didn't have any impact on who was the CEO a fortnight later.

If people want history to repeat, they are free to not learn lessons from this.

If they want success in the next vote to oust a leader deemed undesirable, they should learn why this one was undone despite the board having the theoretical power.

This is also why I take a dim view of those who can't see why an LLM could be a risk given "it's just words": the right words in the right ears can undo almost any human institution.

Uber-Successful Silicon Valley tech entrepreneur is a manipulative narcissist. News at 8.

She had months to come up with an explanation and this is all she can do?

I mean, I dislike Sam Altman as much as the next guy but if any of this was true she would have said it at the time.

Reply on news.ycombinator.com