Earlier quoted context omitted.
He should be making investors money. That's his #1 job. 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's his #1 job. That was not is #1 job as per the charter, that why all this controversy started.
Sam Altman's knack for dodging bullets with a little help from bigshot friends
151–160 of 204 posts
Re: Sam Altman's knack for dodging bullets with a little help from bigshot friends
#152As Paul Graham says, Sam is a Pied Piper who plays his flute and will take the Geeks to the town lake and dump them when he is done
Re: Sam Altman's knack for dodging bullets with a little help from bigshot friends
#153> “The senior leadership team was unanimous in asking for Sam’s return as CEO and for the board’s resignation, actions backed by an open letter signed by over 95% of our employees. The strong support from his team underscores that he is an effective CEO,” said an OpenAI spokeswoman. This is where I stopped reading. At the risk of sounding defensive of Sam Altman. If there's something so fundamentally wrong with him,…
Many years ago I once tried to share a History channel documentary on the JFK assassination with a friend. They watched up until the first part, the documentary dramaticized the assassination and talked about how the shot seemed impossible based off the shape of the car. 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 modifie…
Re: Sam Altman's knack for dodging bullets with a little help from bigshot friends
#154Earlier quoted context omitted.
Yes, the accomplishments of OpenAI represent that in action. It's an exceptional outcome to say the least. If everybody could do it, they would be doing it. Altman's abilities are every bit as rare as that of a 10x software developer. The same was true of Jobs, and the same is true of Musk (regardless of whether someone likes him or not; who cares if he's likable, it's an infantile emotional derangement to obsess so…
> Yes, the accomplishments of OpenAI represent that in action. It's an exceptional outcome to say the least. What specifically backs that up? Didn't others create the technology? Is it that hard to hire and organize when you've got a top researcher that everyone wants to work with? By many accounts, he was just using OpenAI to further other businesses and investments.
Hire a top researcher? Ok, you could get lucky with a few. But keep? Good luck if you are an incompetent manager. If Sam Altman was the "know-nothing" that so many on HN claim him to be, then the top researchers would leave and go to Meta, Google Brain, etc. It makes no sense that this core research team has been so stable. Is this not evidence enough for you?
Re: Sam Altman's knack for dodging bullets with a little help from bigshot friends
#155Earlier quoted context omitted.
Someone has to be skilled enough in fundraising to be able to raise the money to get the talent to do the thing. That alone is a difficult skill to master. It’s often harder to do than to be able to create great things. Carmack and Woz are legends for a reason.
None of these people were skilled at fundraising. Their opportunities were inherited.
I agree with your first sentence, but I am confused by your second sentence. Are you referring to this part? "Carmack and Woz are legends for a reason."
Re: Sam Altman's knack for dodging bullets with a little help from bigshot friends
#156Earlier quoted context omitted.
> As far as Altman goes, I put him firmly in the same camp as Musk. They haven't done jack shit. 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 del…
Per dozens of engineers at SpaceX...the biography from Isaacson contains numerous exaggerations and outright fabrications. 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…
I tried to a Google search for a source. I cannot find anything. Do you have a good source, or insider information?
Re: Sam Altman's knack for dodging bullets with a little help from bigshot friends
#157Earlier quoted context omitted.
This is a fallacy. You can't say "this other CEO did something" if you're trying to make a point about Sam. If you have a problem with Sam doing something illegal, say it. 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).
The outrage is mostly focused on the fact that he doesn’t seem to tell the truth.
Re: Sam Altman's knack for dodging bullets with a little help from bigshot friends
#158You don't have to like Sam. But people who do big things, by definition, have to get there by pushing things along at a cadence that differs from most other people. I'm not necessarily saying it's a good thing, but I don't get why every article about Sam seems to be surprised that the guy who helped build something nobody thought was possible (AI this decade) got there by doing things in a fast, unique way. If he fol…
He's not unique, instead very textbook sociopath. Lines up with the alleged abuse of his younger sister too. There's not much more to him than a wheeler and dealer, self serving abuser.
Re: Sam Altman's knack for dodging bullets with a little help from bigshot friends
#159Earlier quoted context omitted.
does anyone in the brave new AI community have the patience and humility to try to solve the hallucinations problem? 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.
Nope, those are very different. Fundamentally the reason that LLMs hallucinate is because they are predicting the next most likely word. It would be expensive to find a way to remove outputs where P truth is low, but it's not completely implausible like removing software bugs.
As an aside, I think these are both roughly equivalent to detecting code for which P(halts) is high!