811–820 of 854 posts
To be honest, my bias is against boards, especially non-profit boards who have no skin in the game and often an idealistic view that clashes with operational realities. The case of Altman seems to be a “where there’s smoke” thing to an extent, but I also am not inclined to trust the board over the employees who do seem to like him. Insofar as he steered OAI to be a consumer product company rather than a research inst…
https://web.archive.org/web/20240529195305if_/https://www.bu...
Earlier quoted context omitted.
> Like before there was doubts he was technical. Now it's clear he is not and was acting (at least regarding software) ... He lies, he gets caught up, that's it Says you, random HN poster. Here's the assessment of a NASA astronaut who spent time on the ISS and who has a PhD in mechanical engineering and a degree in applied mechanics: "[Musk is] able to have conversations with our top engineers about software, and the…
Curious for your take on this: https://venturebeat.com/ai/elon-musk-and-yann-lecuns-social-... Or this: > I personally wrote the first national maps, directions, yellow pages & white pages on the Internet in the summer of 1995 in C with a little C++. Didn't use a "web server" to save CPU cycles (just read port 8080 directly). Couldn't afford a Cisco T1 router, so wrote an emulator based on a white paper. Or maybe thi…
The quotes you chose contain direct references to him coding, to code that he wrote himself. Where is the name dropping? Are you being confused by the reference to web servers in the first quote maybe? In 1995 the term web server meant something like Apache with CGI. Writing a custom implementation of HTTP in C was a fairly standard technique to improve performance back then. The first version of Amazon was written in C, Google's web servers were still frequently written in C++ when I joined.
I mean, you say he has only "surface knowledge" and is "acting" to "maintain the illusion", but how many programmers could knock out their own implementation of HTTP in C on their own, let alone implement BGP and all the other low level stuff you need to bring a T1 online? How many would even understand what Musk was talking about?
Earlier quoted context omitted.
And a guy who insulted Musk first.
I think responding to "shove your submarine" with "you're a pedo" is somewhat disproportionate.
Earlier quoted context omitted.
My understanding is that it was not so much peer pressure so much as an explicitly coordinated worker action, based on a rational assessment of what their comp packages would look like if the board's plan played out completely.
In my opinion this is most likely. Scenario A, Sam Altman continues as CEO and the for profit arm of OpenAI continues to call the shots, growth and market share continue to be priority number one. Scenario B, Sam Altman is fired, OpenAI non-profit board (re?)-asserts more control. Safety, alignment and other things like that take priority over growth and market share. x% of Employees are Sam believers, and when the r…
Earlier quoted context omitted.
totally agreed. I've been on the inside of two public fiascos. In both cases, rumor and intrigue ruled because the powers that be would not issue statements. The most prolific one I was on the inside(ish) for was Donglegate from the SendGrid perspective. I firmly believe more open communication would have made that go better.
I’m very curious how Donglegate was destined to be anything but a dumpster fire from moment number one. SendGrid wasn’t perfect by any stretch, but how could it have been better handled? Openly admitting that they fired someone as ransom for a DDOS?
Earlier quoted context omitted.
If I were going to work for someone I would prefer to work for a person who is so beloved by his employees that they’d rather quit than not work for him, absolutely.
> If I were going to work for someone I would prefer to work for a person who is so beloved by his employees that they’d rather quit than not work for him, absolutely. I think it's a stretch to assume he was "beloved by his employees," because they signed that document. Some of those very same people have resigned ( https://www.fastcompany.com/91126785/openai-resignations-are... ), which is not someone you typically…
Only two notable figures resigned. The others worked for the superalignement group (nothing of value, just a waste of time).
Earlier quoted context omitted.
Would you do the same in your company?
You mean pass something like that around? No. I’d seriously consider quitting on the spot if my boss at any place I’ve worked in the past asked me to sign something like that. Not signing isn’t quite the same thing as quitting, but there’s a clear implication not doing so would seriously damage any long term prospects at the company.
Can't be toxic culture if everyone wants to work there, and if 95% of people threatened to fire themselves to follow him. It's toxic be a board member and to write a research paper spilling trade secrets of the company that you work for. Maybe that's why the CEO doesn't talk to you about product launches anymore.
> Can't be toxic culture if everyone wants to work there Some counterarguments: 1) this only works if information flows out of the organization, but OpenAI has used NDAs to suppress criticism 2) OpenAI seems to me to polarize the people with the technical skills to work there, I think I'd bet that more of them are creeped out than eager 3) people, especially young people, are very capable of believing that being a je…