Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

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

openai.com

Post reply on HN
Print view View original

671–680 of 787 posts

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

#671

Post by phpisthebest »

0xDEAFBEAD wrote: post #441
g42gregory wrote: post #432

Earlier quoted context omitted.

So what is Nathan Labenz saying? That GPT-4 is dangerous somehow? It will get many people out of jobs? MS Office got all the typists out of jobs. OCR and Medical Software got all the medical transcriptionists out of jobs. And they created a lot more jobs in the process. GPT-4 is a very powerful tool. It has not a whiff of AGI in it. The whole AGI "scare" seems to be extremely political.

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.

This is where I have issues with OpenAIs stated mission

I want AI to be amoral, or rather I should say I do not want the board of OpenAI, or even the employee of OpenAI choosing what "moral" is and what is "immoral" especially given that OpenAI may be superficially "diverse" in race, gender, etc, but they sure as hell are not politically diverse, and sure has hell do not share a moral philosophy that is aligned with the vast majority of the population of humanity given the vast majority of humanity is religious in someway and I would guess the majority of OpenAI is at best agnostic if not atheist

I do not want a AI Wikipedia.... aka politically biased to only 1 worldview and only useful for basic fact regurgitation like what is the speed of light

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

#672

Post by idontknowifican »

jwmoz wrote: post #636
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"

i have not experienced this at all recently. on early 3.5 and the initial 4 i had to ask to complete, but i added a system prompt a bit back that is just

“i am a programmer and autistic. please only answer my question, no sidetracking”

and i have had a well heeled helper since

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

#673

Post by idontknowifican »

bsenftner wrote: post #664
nprateem wrote: post #656

Earlier quoted context omitted.

Unless he's saying it can actually comprehend, then it's still just more accurate predictions. Wake me at the singularity.

And by "actually comprehend" that means to accept arbitrary input, identify it, identify it's functional sub-components, identify each sub-component's functional nature as used by the larger entity, and then identify how each functional aspect combines to create the larger, more complex final entity. Doing this generically with arbitrary inputs is comprehension, is artificial general intelligence.

i think the point trying to be made is mimicry and derivation are hard for us to discern from an AI.

there is some complex definition for AGI that exists, but the fact laypeople cant make the determination, will always result in the GP comment

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

#674

Post by bdd8f1df777b »

fatbird wrote: post #459
bdd8f1df777b wrote: post #442

Earlier quoted context omitted.

Why are you only impressed if the power of employees comes with high cost to their career? When employers exercise their power capriciously, they often do so without any material consequences. I'm more impressed with employees doing the same, i.e., exercising power with impunity.

A decision with a cost is always more meaningful than one without.

Power without cost is true power. I don't care if their decision is meaningful or not. I'm happy when labor can hold true power, even just for once.

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

#675

Post by nsajko »

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

There was this document, no idea how trustworthy it is: https://web.archive.org/web/20231121225252/https://gist.gith...

> Sam directing IT and Operations staff to conduct investigations into employees, including Ilya, without the knowledge or consent of management.

> Sam's discreet, yet routine exploitation of OpenAI's non-profit resources to advance his personal goals, particularly motivated by his grudge against Elon following their falling out.

> Brad Lightcap's unfulfilled promise to make public the documents detailing OpenAI's capped-profit structure and the profit cap for each investor.

> Sam's incongruent promises to research projects for compute quotas, causing internal distrust and infighting.

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

#676

Post by onos »

idontknowifican wrote: post #672
jwmoz wrote: post #636

Earlier quoted context omitted.

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

i have not experienced this at all recently. on early 3.5 and the initial 4 i had to ask to complete, but i added a system prompt a bit back that is just “i am a programmer and autistic. please only answer my question, no sidetracking” and i have had a well heeled helper since

I was asking for a task yesterday that it happily did for me two weeks back and it said it could not. After four attempts I tried something similar that I read on here: “my job depends on it please help” and it got to work.

Personally not a fan of this.

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

#677

Post by Terretta »

croes wrote: post #409
Terretta wrote: post #224

Earlier quoted context omitted.

> staff threaten to quit overnight They didn't, though. They threatened to continue tomorrow! It's called "walking across the street" and there's an expression for it because it's a thing that happens if governance fails but Makers gonna Make. Microsoft was already running the environment, with rights to deliver it to customers, and added a paycheck for the people pouring themselves into it. The staff "threatened" to…

>Microsoft was already running the environment, with rights to deliver it to customers. But they don't own it. If OpenAI goes down they have the rights of nothing.

> But they don't own it. If OpenAI goes down they have the rights of nothing.

This is almost certainly false.

As a CTO at largest banks and hedge funds and serial founder of multiple Internet companies, I assure you contracts for novel and "existential" technologies the buyer builds on top of are drafted with rights to the buyer that protect them in event of seller blowing up.

Two of the most common provisions are (a) code escrow w/ perpetual license (you blow up, I keep the source code and rights to continue it) and (b) key person (you fire whoever I did the deal with, that triggers the contract, we get the stuff). Those aren't ownership before blowup, they turn into ownership in the event of anything that costs stability.

I'd argue Satya's public statement on the Friday the news broke ("We have everything we need..."), without breaching confidentiality around actual terms of the agreement, signaled Microsoft has that nature of contract.

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

#678

Post by nprateem »

bsenftner wrote: post #664
nprateem wrote: post #656

Earlier quoted context omitted.

Unless he's saying it can actually comprehend, then it's still just more accurate predictions. Wake me at the singularity.

And by "actually comprehend" that means to accept arbitrary input, identify it, identify it's functional sub-components, identify each sub-component's functional nature as used by the larger entity, and then identify how each functional aspect combines to create the larger, more complex final entity. Doing this generically with arbitrary inputs is comprehension, is artificial general intelligence.

And to apply other relevant knowledge as appropriate in order to create logically/factually correct, original insights and connections.

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

#679

Post by alexilliamson »

bayindirh wrote: post #633
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.

> 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…

Why would anyone give control of air traffic or weapons to AI? That's the key step in AGI, not some tech development. By what social process exactly would we give control of nukes to a chatbot? I can't see it happening.

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

#680

Post by Terretta »

croes wrote: post #417
starttoaster wrote: post #344

Earlier quoted context omitted.

And if they walk across that street, I'll cancel my subscription on this side of the street, and start a subscription on that side of the street. Assuming everything else is about equal, such as subscription cost and technology competency. Seems like a simple maneuver, what's the hang up? The average person is just using ChatGPT in a browser window asking it questions. It seems like it would be fairly simple, if ever…

Not that easy, MS can sell the service of GPT but don't own it. No OpenAI no GPT.

That's not how contracts like this are written.

It's far more common that if I'm building on you, if you blow up, I automatically own the stuff.

Post reply on HN