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

#631

Post by int_19h » Thu, Nov 30, 2023, 11:15 AM UTC

7e wrote: post #163

bedobi wrote: post #35

i'm still not clear what the accusation against Altman was... something about being cavalier about safety? if that was the claim and it has merit, i don't understand why it wasn't right to oust him, and why the employees are clamoring for him back

My pet theory is that Altman found out about Q* and planned to start a hardware company to make chips accelerating it, all without telling the board. Which is both dangerous to humanity and self-serving. It’s also almost baseless speculation; I’m interpolating on very, very few scraps of information.

How is that dangerous to humanity?

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

#632

Post by quickthrower2 » Thu, Nov 30, 2023, 11:17 AM UTC

asimpleusecase wrote: post #415

This pushed our small team to try Azure instance of GPT3.5 - wow 20 times faster. API does not fail to respond to server requests as we found oh so often on the OpenAI API. We now have something fast enough and stable enough to actually use. Pricing is higher, but my goodness, it actually works.

Hows everyone finding Azure? I’m an Azure bread and butter user but I am sure people mostly on AWS are having fun with Azures CLIs APIs Portal etc. because it is going to be a lot different and unfamiliar. Also egress/ingress costs although mostly the AI costs would far exceed that anyway.

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

#633

Post by bayindirh » Thu, Nov 30, 2023, 11:17 AM UTC

int_19h wrote: post #621

bayindirh wrote: post #562

Earlier quoted context omitted.

Then we should stop teaching Therac-25 incident to developers and remove envelope protection from planes and safety checks from nuclear reactors. Because, users should just input the moral inputs to these things. These are inanimate objects too. Oh, while we're at it, we should also remove battery charge controllers. Just do the moral and civic thing and unplug when your device charges.

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.

> In case of GPT-4 or any other LLM, it's merely "immoral output" - i.e. text. It does not harm anyone by itself.

Assuming that you're not running this query over an API and relaying these answers to another control system or a gullible operator.

An aircraft control computer or reactor controller won't run my commands regardless of its actuators connected or not. Same for weapon systems.

Hall pass given to AI systems just because they're outputting text to a screen is staggering. Nothing prevents me to process this output automatically and actuate things.

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

#634

Post by quickthrower2 » Thu, Nov 30, 2023, 11:21 AM UTC

jacquesm wrote: post #592

mattjaynes wrote: post #559

Earlier quoted context omitted.

Yes, but helping these people save face smoothes the transition. My guess is that those folks were waaay out of their depth and they naturally made naive mistakes. It doesn't benefit anyone to stomp on them. I'm sure they learned hard lessons, and Sam's message is what we call "grace", which is classy. Is it politics? Sure, but only in the best sense. By not dunking on the losers, he builds trust and opens the doors…

They may have been way out of their depth but they also may have been the only ones taking their roles somewhat seriously. They've now been shown what the true balance of power is like and that is a lesson they are probably not going to forget. Unfortunately they also threw away the one shot they had at managing this and for that their total contribution goes from net positive to net negative. I don't think that in a…

If Sam get’s fired I am moving to a HN clone with filters so I can ignore anything related to AI

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

#635

Post by SXX » Thu, Nov 30, 2023, 11:22 AM UTC

ben_w wrote: post #573

SXX wrote: post #553

Earlier quoted context omitted.

Except the reason why we dont all just killed each other yet have nothing to do with risk or cost of killing someone. And everything LLM can come up with will be exactly the same information you can find in any fiction detective book or TV series about crime. Yeah very very dumb criminal can certainly benefit from it, but he can as well go on 4chan and ask about assassination there. Or on some detective book discussi…

> Except the reason why we dont all just killed each other yet have nothing to do with risk or cost of killing someone Most of us don't want to. Most of those who do, don't know enough to actually do it. Sometimes such people get into power, and they use new inventions like the then-new-pesticide Zyklon B to industrialise killing. Last year an AI found 40k novel chemical agents, and because they're novel, the agencie…

Everything you say does make sense, except those people who able to get equipment to produce those chemicals and have funding to do something like that - they dont really need AI help here. There are plenty dangerous chemicals already well known to humanity and some dont actually take anything regulated to produce "except" complicated and expensive lab equipment.

Again difficulty of production of poisons and chemicals it's not what prevent mass murdering around the globe.

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

#636

Post by jwmoz » Thu, Nov 30, 2023, 11:25 AM UTC

irthomasthomas wrote: post #589

Although we have, as yet, no idea what he was actually refering to, I believe the source of the tension may be related to the statements Sam made the night before he was fired. "I think this is going to be the greatest leap forward that we've had yet so far, and the greatest leap forward of any of the big technological revolutions we've had so far. so i'm super excited, i can't imagine anything more exciting to work…

The model is so far forward it refuses to do anything for you anymore and simply replies with "let me google that for you"

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

#637

Post by nojvek » Thu, Nov 30, 2023, 11:25 AM UTC

andy_ppp wrote: post #628

When the following works I'll be impressed: "ChatGPT can you explain possible ways to make fusion power efficient for clean free energy production given the known and inferred laws of physics?" and it actually produces a new answer and plans for a power plant. It's very hard to know how close we are to this, but I do think it's possible these AI models end up being able to infer and invent new technology as they impr…

Correct. What we need is intelligence rooted in math, physics, engineering, chemistry, material science (aka great grasp of reality).

Then you can ask it to create designs and optimize them. Ask it to create hypothesis and experiments to validate them.

Human brains are v capable at abstraction and inference but memory and simulation compute is quite limited. AI could really help here.

How can we design better governance strategies?

Analyze all passed laws and find conflicts?

Analyze all court cases and find judges who have rules against the law and explain why? Which laws are ambiguous? Which laws work against the general public and favor certain corporations? Find ROI of corporation donations which led to certain laws, which led to X% rise in their profits.

The really big piece missing from current AI is reality grounded modeling and multi-step compositional planning + execution.

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

#638

Post by say_it_as_it_is » Thu, Nov 30, 2023, 11:27 AM UTC

Is it fair to say that some of the original board members were chosen because they were women and firing Sam, demoting Greg were unanticipated activities of women whose archetype was of a more even-keeled, compassionate character who would value humanity more than a cold, calculating, toxic white male?

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

#639

Post by quickthrower2 » Thu, Nov 30, 2023, 11:28 AM UTC

krick wrote: post #164

bagels wrote: post #12

Is anyone feeling more comfortable about relying on OpenAI as a customer after this announcement?

Well, for starters, we all know that while realistically it's not unusual for a company to have a mission-critical person, it is very undesirable. So much so everybody must pretend that this is just unacceptable and surely isn't about their company. Here, we kinda saw the opposite being manifested. More convincingly than I've ever seen. Second, I simply don't understand what just fucking happened. This whole story do…

“Investors” are supposed to consider their money a donation. Of course the 100x cap is generous so it is kinda an investment. And the coup reveals a higher chance that this will morph towards for-profit as that is where the power seems to be, let alone the money.

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

#640

Post by SXX » Thu, Nov 30, 2023, 11:28 AM UTC

0xDEAFBEAD wrote: post #571

SXX wrote: post #553

Earlier quoted context omitted.

Except the reason why we dont all just killed each other yet have nothing to do with risk or cost of killing someone. And everything LLM can come up with will be exactly the same information you can find in any fiction detective book or TV series about crime. Yeah very very dumb criminal can certainly benefit from it, but he can as well go on 4chan and ask about assassination there. Or on some detective book discussi…

>And everything LLM can come up with will be exactly the same information you can find in any fiction detective book or TV series about crime. As Nathan states: >And further, I argued that the Red Team project that I participated in did not suggest that they were on-track to achieve the level of control needed >Without safety advances, I warned that the next generation of models might very well be too dangerous to re…

I dont disagree with his points, but you completely miss the point of my post. People dont need an AI advise to commit crime and kill others. Nonestly humans they're pretty good at it using technology of 1941.

You don't have bunch of cold blood killers going around not just because police is so good and killers are dumb and need AI help. It's because you live in functioning state where society have enough resources so people happy enough to instead go and kill each other in Counter Strike or Fortnite.

I totally agree that AGI could be a dangerous tech, but it's will require autonomity where it can manipulate real world. So far GPT with API access is very far from that point.