"To smooth his exit, Altman proposed he move from president to chairman. He pre-emptively published a blog post on the firm’s website announcing the change. But the firm’s partnership had never agreed, and the announcement was later scrubbed from the post."
I also thought this was interesting, not sure if this detail had been reported before this article, but seems like there were maybe more incidents than initially thought.
This list of 20 examples is new, albeit unsurprising given that there was no reason to think that there were only 2 incidents, and I highlight it in my longer commentary: https://www.lesswrong.com/posts/KXHMCH7wCxrvKsJyn/openai-fac...
"Graham said it was his wife’s doing. “If anyone ‘fired’ Sam, it was Jessica, not me,” he said. “But it would be wrong to use the word ‘fired’ because he agreed immediately.” "
The more ability you have to be kind of waterproof so to speak - like a duck - because of a mixture of positioning, ruthlessness, access, intelligence, perseverance, etc the better you are. Most importantly you have to regularly show allegiance to specifically investors over all other considerations, and boom you’re perfectly crafted.
Altman has brilliantly positioned himself as that perfect fulcrum between venture and technical skills such that he is (in his own mind and to a tiny subset of incredibly powerful people in SV) indispensable to the silicon valley crowd who doesn’t really know how to navigate the intersection between venture and technical skills.
They believe in his focus to get things done, not in his ethics. And with ChatGPT, he got things done. Which means, he connected the right people and raised money so they can work on the important things.
"Oh you think that going ahead before a decision is made and just publish a post will get us to decide in your favor because we're afraid of looking like we're backtracking? How about we don't decide in your favor and delete your post? What you gonna do now?"
The article talks about his "luck" in "surviving." Which is about all he has, apparently. No consideration for the future of the product, the company, or the industry. The triumphant conclusion of this article is "a bunch of CEOs pulled weight and Altman got his old job back, again."
The more of these I read, the more convinced I am that his company is going nowhere, as they seem more skilled at producing "internal drama" than "new product."
My take on it is that they have an exceptionally skilled marketing department that's good at creating the impression that they've actually created an AGI, with stories trickling out to the press about ChatGPT 'getting lazy' or 'taking a break for the holidays' and other stuff that serves to humanise what is, at least for me, a product which routinely fails to accurately answer basic questions, the answers to which are contained in hundreds of blogs and forum postings.
If others are experiencing something different that's interesting. Personally I note that the 'I got this weird answer' posts are never accompanied by screenshots or any other evidence, they're just claims which all subsequent repliers appear to take as the unvarnished truth.
I congratulate their marketing team; I suspect that Mistral will ultimately present a far superior product.
Don't you think most entrepreneurs would call Sam if they have a chance?
What business has he aced?
CEO of Reddit, how'd that go? Project Covalence, how'd that go? Andrew Yang? Our next president, right? 2018, The year Sam Altman became the governor of California, no? And of course: "I have guns, gold, potassium iodide, antibiotics, batteries, water, gas masks from the Israel Defense Forces, and a big patch of land in Big Sur I can fly to." - that aged well.
But he's now in a position where all his failures will be forgotten and his successes amplified forevermore (unless he buys a social media company, but he seems too smart for that so far). It is clear he has reached Jeff Dean levels of ridiculous hagiography though, but I guess worshipping CEOs is a step up from worshipping invisible friends or stinky geriatric ex-presidents so there's that.
Just by probabilities alone it's happened hundreds of times. Probably the source for original prophet claims.
If he followed the norms, he'd be a mid-level manager at FAANG company.
Imagine I win the lottery and hire really smart architects and tell them to make a revolutionary new building. They go on to make some amazing building that gets all sorts of awards and recognition. Is anyone really so naive as to suggest that I am deserving of any credit whatsoever? In the words of Urkel, "Did I do that?"
There are some leaders in the tech world who have clearly done something. John Carmack did something. Steve Wozniak did something. There are some questionable cases as well. As far as Altman goes, I put him firmly in the same camp as Musk. They haven't done jack shit. I give them no respect, recognition, or credit. They were just wealthy people spending money. I seriously question the people who look at them positively in any way whatsoever.
You need both to do something great. And there’s thousands of Woz’s out there whose names we don’t know because they never met their Jobs.
And that's a really common case.
It is absolutely true that you need both technical and non-technical people to make a business work. There are lots of jobs to be done, hats to be worn, and they are all valuable - many are necessary.
The issue is that our society and economy massively overvalues two roles in particular: those who bring the capital and those who carry the title executive. It's not that those roles don't contribute, they do. But they currently get the vast majority of the generated wealth, credit, and recognition (which then translates into more opportunities to access more capital and thus into an exponential feedback loop).
You should try hiring a group of really smart people and getting them to work together to create something great before you dismiss the skills of the person doing that.
There wouldn't be a counterfactual in your hypothetical scenario, but I'm confident that the difference between someone that got lucky and hired "smart" architects versus a visionary doing the same task would be drastic.
Money can only get you so far.
We know it's possible for someone with no technical ability to look like genius just by having lots of money. I knew people who believed Musk was going to bring about an AI revolution all on his own because of his very special brain, but his tenure at Twitter was laughable to anyone who has passed CS 101. What makes you convinced all the rich "geniuses" who haven't revealed themselves to be idiots are actually smart for real this time?
Each one will end up pulling in their own direction, fighting with others over minor things, and nothing will ever get done.
Managing smart people, resolving conflicts, and making sure stuff gets done is a rare and underappreciated skill.
>claimed to have funded the invention of the piano and opera, financed the construction of Saint Peter's Basilica and Santa Maria del Fiore, and were patrons of Brunelleschi, Botticelli, Leonardo da Vinci, Michelangelo, Raphael, Machiavelli, Galileo, and Francesco Redi, among many others in the arts and sciences.
Sure they didn't do the paintings but a lot of that stuff probably wouldn't have happened without them.
This example doesn't work, at least, since the industrial revolution. If an average person wins the lottery they would need to solve a lot (really a lot) of issues to move a business forward and most probably go bankrupt except if he/she puts the money in low risk financial instruments.
There is no "I do that" in the complex world of business. It is a nice Disney story with a few examples. Organization does the work and that includes anonymous engineers, not only CEOs and founders. The great ignition could start at a garage (e.g. Apple) but the road is long.
What Musk did with SpaceX is in no way comparable to just paying an architect to build a building. Unless the entire biography from Isaacson was fabricated, he was extremely hands on in design decisions, company direction, and hiring.
If you think none of those were relevant to the success of spacex, you’re completely delusional.
Elon Musk's involvement in design decisions was limited to the typical executive level involvement in hardware design, in the sense that he was presented multiple options by the engineers and designers and got to pick which one they went with.
Indeed, with SpaceX it's usually immediately obvious when Musk was involved: does something fail spectacularly in a way it wasn't supposed to? That was Musk, getting involved in something he didn't understand and overriding the engineers and other professionals. (Contrast the first Starship launch this year, which Musk micromanaged, vs the second launch in November in which he had minimal involvement due to being distracted by X and Tesla.)
Why are you surprised by folks reaction? What did he successfully build before OpenAI? Nothing as far as I can tell? His main skill seems to be networking, not building product. Also, doing it fast, does not require lying, as seems his habit.
A lot can be accomplished by skirting pesky rules and regulations. It isn't something commendable.
I'm not saying Sam has never done anything illegal, but most of the outrage seems to be that he doesn't follow social/business norms (not that he doesn't follow rules/regulations/laws).
There is nothing fallacious about my comment. Saying I'm not allowed to compare him to other CEOs is ridiculous.
As another commenter said: he isn't unique.
Well that depends.
Is the thing that is being accomplished very good and are the regulations bad?
For example, what if there is a law that was passed for the sole purpose of protecting a well connected monopoly, so that consumers have to pay more money, only so that this one monopoly could profit.
I'd say that such a law would be bad and should be under minded.
Hey, if you don't like Sam, that's fine. But that doesn't make him a sociopath.
This sort of misattribution seems to show up when you have individuals with sociopathic tendencies running firms that do cool things: Elon Musk is the king of this. The man has not, on his own, had any sort of good results on the world - Tesla and SpaceX seem to do a lot better when he steps away from them - but he has found ways to get himself into a position to take credit from the engineers who do. Hyperloops and Twitter are the results of Elon Musk's original work. That's not to say that the management is necessarily negative: someone needs to fund the work of the good engineers.
If we followed the norms, he's be in prison. If you think fraud and self-dealing are justified if it moves a lot of value to shareholders, write to your senator. Until then the laws should be enforced against the rich as well as the poor.
Of course, I can't much bring myself to care in this case, because all that's at stake is a chatbot.
There's an art to building a rapid growth trajectory, cutting all kinds of corners to ramp up your customers/assets/users/whatever -- and then pirouetting into a bigger new role, while leaving someone else to stabilize your unstable creation.
Take it up a level, and the best practitioners are also very good at blaming their successors for messing up a good thing. Also very good at cultivating a certain subset of journalists who will keep burnishing the legend, never stepping outside the bubble.
Loosely related: does anyone in the brave new AI community have the patience and humility to try to solve the hallucinations problem? Because it seems as if Sam & Co. are quite determined to ignore it, deny it or minimize it, while continuing to push for breakneck growth of products with a very troubling design flaw.
I think this is akin to asking if anyone in the software engineering community has the patience and humility to try to solve the software bugs problem.
As an aside, I think these are both roughly equivalent to detecting code for which P(halts) is high!
No it never eliminated all bugs, but if we're conflating the two problems you don't need to eliminate all hallucinations either, you just need to make them very unlikely, or a consequence of your bad specifications.
Both have been characterized as being manipulative liars though.
It was also downweighted as a follow-up [1] because there had been a similar profile article / megathread the day before:
"King of the Cannibals": How Sam Altman Took over Silicon Valley - https://news.ycombinator.com/item?id=38744021 - Dec 2023 (162 comments)
In the absence of significant new information [2] people mostly just repeat what has already been said, and the more reptition there is in a thread, the dumber and nastier it gets.
(Edit: I double checked that by skimming through the comments in this thread and I think it was the right call—they're generic, i.e. could just as easily have appeared in any similar thread, and range from ok-but-repetitive to outright-bad.)
[1] https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
[2] https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so...
Who among us ...
Like all human beings, he wasn't a one-dimensional character from a crappy sitcom. He had negative and positive traits. Whatever one feels about him personally, it is absolutely a mistake to dismiss him. What he achieved -- multiple times -- puts him in extremely rarified air in the business world.
Billionaires need on the ground types like him to actually do stuff.
It’s hard to find people who are young and ambitious and effective, and yet fundamentally and ruthlessly aligned with corporate hegemony.
And I suspect he must be spectacularly good at flattery.
We share this thought.
My pet theory regarding the recruitment efforts of this class: two broad fronts, one to drive the herd and the other to pick candidates.
The first front permeates our head space in form of content in the full spectrum of the bandwidth of this class: entertainment, news, organizations. Here the message is insistently that “Only suckers play straight”, “crime pays”, “winners and losers”. This is a consistent message.
The second front is more selective and here I am speculating (unlike the former which one gets to experience first hand, whether one likes it or not.) Here I suspect a more ‘direct’ sales pitch is used, and likely references the former.
Imagine you are a bright young thing from ‘not money’ raised in that “cultural” environment just mentioned. Now you open your favorite “journal of record”. You read barely disguised lies, note how these lies lead designated demographics by the NOSE with them not even aware of this fact, and finally your mentor has a conversation with you:
“Do you really think that the general population, these same people we were discussing being led by the nose with lies, are actually capable of self-governance and should have a say in “important” matters?”
This is where I stopped reading.
At the risk of sounding defensive of Sam Altman. If there's something so fundamentally wrong with him, why does he have such broad support? He's just one man, say that what this article alleges is true, how many tricks does he have up his sleeve to fool the entirety of Silicon Valley?
I mean, we can beat around the bush all day but he managed to get his old/new?? job back at OpenAI. This means to me that his ouster was either unwarranted or the board was acting blindly and should have stepped aside just on account of being clearly incapable of making sound decisions.
This article does not ask the right questions. Doesn't answer any questions either like the many articles written before it riddled with speculation about Sam's much belabored firing.
He makes people money. Plenty of terrible people commanded great loyalty. All the loyalty tells you is he’s a good leader. Not where he’s leading us.
(Not saying anything about Sam. Just pointing out that loyalty is orthogonal.)
Aside from that, who cares about loyalty to the CEO of a tech company. He's a replaceable employee and OpenAI should be able to continue without him.
What were his actual crimes. It's not clear from the previous board nor anyone who's spoken about the matter what those were.
That was not is #1 job as per the charter, that why all this controversy started.
Boards don’t judge crimes. And unlike courts, they have no disclosure obligation. Particularly for a private company.
See I had already watched this documentary in full, and knew that this was just a cliffhanger for the second part that explained the car was modified, making the seemingly impossible shot very possible. Yet, my friend said "See? That's enough evidence." and walked away.
I asked, "don't you want to hear the rest?" the response: "nope."
Your comment reminds me of this experience
Stock options and a carefully built cult of personality, my friend
75 million people voted for Trump.
As for what the benefit is, the ongoing report is one benefit. If the law firm (named here for the first time, I believe) didn't know about, say, Sutskever's list of 20 incidents, well, they do now.
Up until recently, the press about OpenAI has been mostly "look at the shiny lights" and now it's going to get the Elon Musk treatment.
You could shine a light into most people's lives and find in most cases broken egg shells here & there. Fixating on them at the expense of the bigger picture is a mistake.
Regarding the openai board, the problem imo is that too much weight was given to a nebulous & subjective mission, allowing those sitting to make up ridiculous positions and find them tenable (ie the company going kaput being in line with the mission, like srsly??)
However, there won't be a monopoly or anything of the sort that will form, our society's institutions wouldn't be able to properly enforce it and it's too easy for weights to leak, too much alpha for people to not fight back against it. Not worried.
The ai safety crowd's problem is that they're focused on a hypothetical entity that will take over humanity, imagining all sorts of scenarios in lalaland which are ridiculous (some of their "thought leaders" even advocating for strikes on datacenters to save humanity???), where the real risk is criminals/terros being empowered by the ai's cognitive ability. For that, we'll need the gov involved one way or another.
I cant picture a case where charges need to be brought under a novel framework that couldnt be captured by existing laws. If you deploy a computer virus, thats a crime, if you hack someones bank account, thats a crime, if you slander someone, thats a crime.
Im personally not convinced that any of the public efforts made around AI safety have been honest. If these people were serious then like you say, they wouldnt be chasing evil scifi ai overlords. I can think of some wonderful use cases to counter things like AI generated fakes; I am actually uncomfortably suprised that every social media company hasnt added a banner below every picture that says whether scans indicate a picture was deepfaked, twitter could scan a picture when you post it and show what a variety of algorithms think. That would have a real shot at improving information asymmetry at the social level.
OpenAI going kaput might indeed be the best for humanity as a whole, only time will tell.
