Top
Best
New

Posted by adrianhon 1 day ago

Sam Altman may control our future – can he be trusted?(www.newyorker.com)
1901 points | 783 commentspage 8
BrenBarn 17 hours ago|
Of course not. No one can be trusted to control our future.
sph 13 hours ago||
Excellent article, truly well-researched. As someone close to a pathological liar [1], the idea that one could be at the forefront of the creation of an artificial superintelligence confirms all the existential risks of such a piece of technology and how naïve, if not ignorant, the average starry-eyed tech worker and investor is about this whole endeavour. It's easy to believe there is a lot of idealism and wish for a better world, but underneath the greedy drive for money and power is excellently summarized in Greg Brockman's own thoughts: “So what do I really want? [...] Financially what will take me to $1B.”

Literally, the only hope for humanity is that large language models prove to be a dead-end in ASI research.

---

1: “He’s unconstrained by truth,” the board member told us. “He has two traits that are almost never seen in the same person. The first is a strong desire to please people, to be liked in any given interaction. The second is almost a sociopathic lack of concern for the consequences that may come from deceiving someone.” — I guess now I know of two people with these traits.

flippyhead 6 hours ago||
No
cedws 14 hours ago||
Sounds like a snake pit. None of them can be trusted. If we have to rely on companies to self appoint a benevolent ‘AI dictator’ we’re fucked.

The only high profile person in AI I’d consider perhaps worthy of trust is Demis Hassabis.

CyborgUndefined 13 hours ago||
ugh, i don't understand why only altman scares you? what about google, china, and other players?

for me, the answer >>> we need to create our own systems. decentralized agent networks and etc.

if you don't want to depend on one person or one company controlling your AI, build your own infrastructure.

the concentration of power in one/two persons is the problem.

lenerdenator 23 hours ago||
If you are asking if a single human can be trusted with such a responsibility, the answer is, by default, no.
pdonis 21 hours ago||
Does the article ever actually answer the title question?
mohamedkoubaa 21 hours ago||
The answer is no, he can't be trusted
pdonis 21 hours ago||
Oh, I agree that's the correct answer. I just don't see the article actually ending up with that answer. I see it waffling. Basically, the article ends up saying that, well, we told you about all this dodgy stuff, but what he's doing is working.
Wyverald 18 hours ago|||
God forbid an article presents all the evidence from all parties and asks you to reach a conclusion by yourself...

Sorry for the snark. But I genuinely think the way they did this was perfect.

pdonis 17 hours ago||
> I genuinely think the way they did this was perfect.

Evidently we disagree. I responded about that to another commenter downthread.

mohamedkoubaa 19 hours ago|||
Trusted to increase shareholder value is also questionable
kubik369 20 hours ago||
I think you are misunderstanding the point of journalism. It can be debated whether the title should be such a question. Nevertheless, the article should just present information, ideally in a balanced way, without author's bias, so that you can decide for yourself. You can see the attempts at the balanced part in the article where an allegation/statement is made about Altman followed by parentheses saying that Altman recalls the exchange differently/does not remember.
pdonis 19 hours ago||
> the article should just present information, ideally in a balanced way, without author's bias, so that you can decide for yourself.

I get that this is the claimed ideal of journalism, at least for straight reporting. The problem is that it's impossible.

There isn't time or space to present all the information; the journalist has to filter. And filtering is never unbiased. Even the attempt to be "balanced" is a bias--see next item.

"Balanced" always seems to mean "give equal time and space to each side". But what if the two sides really are unbalanced? What if there's a huge pile of information pointing one way, and a few items that might point the other way if you believe them--and then the journalist insists on only showing you a few items from the first pile, so that the presentation is "balanced"? You never actually get a real picture of the facts.

There's a story that I first encountered in one of Douglas Hofstadter's books, about two kids fighting over a piece of cake: Kid A wants all of it for himself, Kid B wants to split it equally. An adult comes along and says, "Why don't you compromise? Kid A gets three-quarters and Kid B gets one-quarter." To me, the author of this article comes off like that adult.

In any case, all that assumes that this article is supposed to be just straight reporting, no opinion. For which, see the next item.

> It can be debated whether the title should be such a question.

Yes, it certainly can. If this article is just supposed to be straight reporting--no editorializing--then that title is definitely out of place. That title is an editorial--and the article either needs to own that and state the conclusion it's trying to argue for, or it shouldn't have had that title in the first place.

kubik369 12 hours ago||
> "Balanced" always seems to mean "give equal time and space to each side". I agree with you that this seems to be the idea people have when "balanced" is mentioned. I don't think this is correct. You can easily have a balanced article which has lots of evidence pointing one way or the other. I think that this article is like that. Boatload of pointers towards Altman being a sly person with reporters asking him about those exchanges and him basically shrugging each time.

The journalists credibility is doing quite a bit of lifting here as we have to trust that they put in the effort. One such example is the molesting accusations which the reporters say they heavily looked into and were not able to find any corroborating evidence.

> You never actually get a real picture of the facts. Yes, it is a fundamental impossibility in lots of cases. That's why we trust the reporters that they did as good a job as they could to present all pertinent information.

> That title is an editorial ... I do not perceive it to be editorialised. It states an arguably real possibility that Altman may/does have lots of real power. I am guessing that you believe that the "can he be trusted" is an editorialisation that points towards him being untrustworthy. If that is the case, I think those would be your biases knowing that he is probably not trustworthy. I see it just as an objective question.

Imagine a different situation: you have local elections into your small town. There is a new mayor candidate and during the next term, there will be some money to be given to residents for renovations and such, but not enough for everyone. You don't know this candidate. A local reporter, whom you trust, writes an article "New mayor candidate favoured in polls - will he be fair with the renovation money?". It is a piece trying to shed light on who this candidate is as a person, what was his life before moving into your village, etc. so that voters like you can decide whether to give him your vote. It is not editorialised, as it does not point either way.

shevy-java 14 hours ago||
I don't trust him. He already made statements that convinced me I don't want to touch anything he controls. In a way it is similar to Meta and co. For some reason the US corporations behave very suspiciously once past a certain threshold size. With Win11 from Microsoft I always wonder whether there is a not so hidden subagenda in place.
eximius 14 hours ago||
Fuck no! Of course he can't be trusted. We know that. Nobody questions that. We know that about most of the "elites" running the show.

We're just in this shitty pit of despair where people are desperate. It's difficult to campaign for good when you're struggling and capital can jerk people around.

People pursue good for the sake of good at cost to themselves when times are very good or times are very, very bad.

Right now times are only merely very bad.

calf 8 hours ago|
The last quote, to a layperson, may sound completely sinister, but therein lies a deep and open computer science question: AIs really do seem to get their special capabilities from having a degree of freedom to output wrong and false answers. This observation goes all the way back to some of Alan Turing's musings on how an AI might one day be possible. And then there were early theorems related to this e.g. PAC learning. I'd love to know about what's happened since on this aspect, such as the role of noise and randomness, and maybe even hallucinations are a feature-not-bug in a fundamental sense, etc.
More comments...