Sam Altman returns as CEO, OpenAI has a new initial board • Hacker News Board

Re: Sam Altman returns as CEO, OpenAI has a new initial board

#731

Post by BoiledCabbage » Thu, Nov 30, 2023, 7:38 PM UTC

irthomasthomas wrote: post #709

NiteOwl066 wrote: post #693

Earlier quoted context omitted.

What's the source of those comments?

Sam Altman at the APEC conference, taking part in a panel, along with Google and Meta AI people. Actually, it's quite amusing hearing Google exec define AI as Google translate, and Sam's response to that. https://youtu.be/ZFFvqRemDv8?t=770

Thanks for the link. I found a few interesting topics in there.

One other, was by one of the other execs on stage on categorization of the types of risks people discuss with AI vs lumping everything together under "safety". https://youtu.be/ZFFvqRemDv8?t=1430

1. GPT outputs - Toxic, bias or non-factual

2. System Usage - misinformation, disinformation, impersonation (ex voice)

3. Society/Work - Impacts on workforce, education, replacing jobs, decision making

4. Safety - The more sci-fi style safety discussion.

It's a very insightful point to call out, and something I never see discussed with the same rigor / granularity here on HN. People often pick their strawman from the list and argue for or against all 4 using that one.

Re: Sam Altman returns as CEO, OpenAI has a new initial board

#732

Post by int_19h » Thu, Nov 30, 2023, 8:08 PM UTC

bossyTeacher wrote: post #643

int_19h wrote: post #621

Earlier quoted context omitted.

In both of your examples, the result of "immoral inputs" is immediate tangible harm. In case of GPT-4 or any other LLM, it's merely "immoral output" - i.e. text. It does not harm anyone by itself.

The point here is that is giving the user detailed knowledge on how to harm others. This is way different than a gun where you are doing the how (aiming and pulling the trigger). The guy says he wanted to slow down the progress of AI and GPT suggested a detailed assassination plan with named targets and reasons for each of them. That's the problem.

Thing is, this detailed knowledge is already available and much easier to acquire. There are literally books on Amazon that explain how to jury-rig firearms and bombs etc. Just to give one example: https://www.amazon.com/Improvised-Munitions-Handbook-TM-210/....

When it comes to those "detailed plans", if you actually try something like that, what you get from it is a very broad outline that is pretty much a mish-mash of common sense stuff and cultural tropes (many of which aren't true IRL). Similarly, the "list of targets" that it makes is simply the most prominent people associated with that area in public perception, not necessarily people who are actually key. All of this can be achieved just as well with a few Google searches, and the resulting plan will likely be much better for it.

I've yet to see any example where GPT would come up with something along these lines that is not trivially found on the Internet anyway.

Re: Sam Altman returns as CEO, OpenAI has a new initial board

#733

Post by rezonant » Thu, Nov 30, 2023, 8:12 PM UTC

TeMPOraL wrote: post #555

rezonant wrote: post #467

Earlier quoted context omitted.

This is a universal truth of feedback and customer service. Every user report is an iceberg: for every 1 person there's a much more significant number of people who experienced the problem but never reported it.

Yes, but the company may be as an ice breaker going across the pole in a straight line, and still when asked about hitting ice, the captain will say that this now is literally the first time it ever happened.

So true.

Re: Sam Altman returns as CEO, OpenAI has a new initial board

#734

Post by Karunamon » Thu, Nov 30, 2023, 8:14 PM UTC

bhpm wrote: post #728

sanderjd wrote: post #721

Earlier quoted context omitted.

In general, your comments strike me as reflecting a very naive understanding of how anything works. Legal responsibilities are simply not the only thing that matters. This entire episode was about 95% a PR battle, which the board lost and Altman won. If it's not obvious to you that the board's legal rights and responsibilities had essentially zero to do with the outcome here, then I really don't know how to help you.

“Naive” is thinking that taking a poll of the public or employees (which is the only reasonable action I can think of that looks like “arguing their case”) would have had any positive effect on the outcome. This is simply… not how decisions are made. Even shareholder vote calls (for companies that have them) are coordinated, with the outcomes for consequential decisions well understood before voting happens. The opti…

>there is no version of this story where the board doesn’t do whatever they hell they want, and no version where “making the case” doesn’t result in even more chaos.

The failure to make their case is PRECISELY why the board was ultimately unable to do what the hell they wanted: remove Sam Altman. As it turned out, his presence was important enough to the employees that the usual corporate playbook of "do whatever the hell we want and only disclose what is legally required" backfired spectacularly.

If your thesis was correct, Sam would not be there right now.

Re: Sam Altman returns as CEO, OpenAI has a new initial board

#735

Post by _factor » Thu, Nov 30, 2023, 8:17 PM UTC

gopher_space wrote: post #209

cyanydeez wrote: post #147

Earlier quoted context omitted.

they're B2B now, that means only political correctness. and I'm not sure why anyone dances around it but these models are built by unfiltered data intake. if they actually want to harness bias, they need to do what every capitalist does to a social media platform and curate the content. lastly, bias is an illusion of choice. choosing color over colour is a byproduct of culture and you're not going to eradicate that b…

What does political correctness and bias mean in this context? edit: I'm asking because to my eye most of these conversations revolve around jargon mismatch more than anything else.

“Write me a web application.” Sure, here are some Microsoft and Google products to do so!

Not all filtering has to be prohibitive, just unnoticed.

Re: Sam Altman returns as CEO, OpenAI has a new initial board

#736

Post by SilasX » Thu, Nov 30, 2023, 8:38 PM UTC

creato wrote: post #508

0xDEAFBEAD wrote: post #441

Earlier quoted context omitted.

Nathan says the initial version of GPT-4 he red-teamed was "totally amoral" and it was happy to plan assassinations for him: https://nitter.net/labenz/status/1727327464328954121#m Reducing the cost of medical transcription to ~$0 is one thing. Reducing the cost of assassination to ~$0 is quite another.

> Reducing the cost of assassination to ~$0 is quite another. It is reducing the cost of developing an assassination plan from ~$0 to ~$0. The cost of actually executing the plan itself is not affected.

Good planning necessarily reduces the cost of something relative to the unplanned or poorly planned version. If it identifies a non-obvious means of assassination that is surprisingly easy, then it has done something "close enough" to reducing the cost to $0.

Re: Sam Altman returns as CEO, OpenAI has a new initial board

#737

Post by vikramkr » Thu, Nov 30, 2023, 8:56 PM UTC

tsunamifury wrote: post #694

vikramkr wrote: post #556

Earlier quoted context omitted.

Ah yes, all those "customers" that build products that were a thin wrapper of chatgpt with an extra feature, truly the innovative powerhouses that will revive our economy /s. I mean seriously, if anyone thought they were going to build a business around a got based startup to parse documents when document parsing is already a specific goal with benchmarks in AI that openai was obviously going to be working towards -…

Seems so odd to be such an jerk about it…

After wave after wave of theranos/crypto/NFTs/ftx etc etc honestly I've got very little patience left and they've soured the entire national mood on silicon valley to the point where this crowd was the justification for a lot of people to want to let SVB take down actual companies with it instead of a bailout. I find them actively harmful, and now they're trying to whip up a storm whining about anticompetitivenes to stop development in a rather useful tool. Won't pretend there isn't a little bit of glee watching the folks who just jumped from NFTs into this as the new big thing seeing their investment already go up into smoke.

Re: Sam Altman returns as CEO, OpenAI has a new initial board

#738

Post by starttoaster » Thu, Nov 30, 2023, 8:57 PM UTC

croes wrote: post #413

starttoaster wrote: post #328

Earlier quoted context omitted.

They're saying that ~80% of OpenAI employees were determined to follow Sam to Microsoft and continue their work on GPT at Microsoft. They're saying this actually signals stability, as the majority of makers were determined to follow a leader to continue making the thing they were making, just in a different house. They're saying that while OpenAI had some internal tussling, the actual technology will see progress und…

They showed they are hypocrites. Blaming the board the hindered OpenAI mission by firing Altman but at the same time threaten to work for MS which would kill that mission completely.

I don't think that's necessarily true or untrue, but to each their own. Their mission, which reads to, "... ensure that artificial general intelligence benefits all of humanity," leaves a LOT of leniency in how it gets accomplished. I think calling them hypocrites for trying to continue the mission with a leader they trust is a bit...hasty.

Re: Sam Altman returns as CEO, OpenAI has a new initial board

#739

Post by quickthrower2 » Thu, Nov 30, 2023, 9:09 PM UTC

mvc wrote: post #701

quickthrower2 wrote: post #630

Earlier quoted context omitted.

Not necessarily. Money keeps you in the game. Giving it all away means you are at the bottom being not that effective. And you can donate money in your will.

If you only give money away at your death, are you really giving it away? What else are you gonna do with it?

Give it to your family is the other option.

Re: Sam Altman returns as CEO, OpenAI has a new initial board

#740

Post by quickthrower2 » Thu, Nov 30, 2023, 9:20 PM UTC

giardini wrote: post #723

quickthrower2 wrote: post #641

Earlier quoted context omitted.

Not this again!

quickthrower2 says: > "Not this again!" I am unenlightened by your short missive. Perhaps there you could point to something clarifying your intent - something like, say: "This has been previously discussed ad nauseum in the Fortune magazine article at URL ..." or "Here's a similar post and the details of OpenAI URL ." Thank you.

It is a HN trope to say why does company X need so many employees. Usually said about world class companies. Usually because you can build something that looks like it in a weekend with React and MongoDB (although OpenAI you can’t prototype like that).