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

OpenAI and Elon Musk

openai.com

Post reply on HN
Print view View original

521–530 of 1001 posts

Re: OpenAI and Elon Musk

#521

Post by lupusreal »

FeepingCreature wrote: post #399
throwaway11460 wrote: post #369

Earlier quoted context omitted.

Reporting requirements are not going to save you from Chinese, North Korean, Iranian or Russian programmers just doing it. Or some US/EU based hackers that don't care or actively go against the law. You can rent large botnets or various pieces of cloud for few dollars today, doesn't even have to be a DC that you could monitor.

Sure, but China is already honestly more careful than America: the CCP really doesn't want competitors to power. They're very open to slowdown agreements. And NK, Iran and Russia honestly have nothing. The day we have to worry about NK ASI takeoff, it'll already long have happened in some American basement. So we just need active monitoring for US/EU data centers. That's a big ask to be sure, and definitely an invasi…

> Sure, but China is already honestly more careful than America: the CCP really doesn't want competitors to power. They're very open to slowdown agreements.

Don't be naive. If the PRC can get America/etc to agree to slowdowns then the PRC can privately ignore those agreements and take the lead. Agreements like that are worse than meaningless when there's no reliable and trustworthy auditing to keep people honest. Do you really think the PRC would allow American inspectors to crawl all over their country looking for data centers and examining all the code running there? Of course not. Nor would America permit Chinese inspectors to do this in America. The only point of such an agreement is to hope the other party is stupid enough to be honest and earnestly abide by it.

Re: OpenAI and Elon Musk

#522

Post by sgift »

davej wrote: post #486
magnio wrote: post #6

> Elon understood the mission did not imply open-sourcing AGI. As Ilya told Elon: “As we get closer to building AI, it will make sense to start being less open. The Open in openAI means that everyone should benefit from the fruits of AI after its built, but it's totally OK to not share the science...”, to which Elon replied: “Yup”. Wow the most open OpenAI has ever been is when someone sue them. On the other hand, th…

This is misleading. Please read the actual source in context rather than just the excerpt (it's at the bottom of the blog). They are talking about AI safety and not maximizing profit. Here's the previous sentence for context: > “a safe AI is harder to build than an unsafe one, then by opensorucing everything, we make it easy for someone unscrupulous with access to overwhelming amount of hardware to build an unsafe AI…

I find it good that no one until now provided this bullshit lie as context. The "we need it to be closed for it to be safe" claptrap is the same reasoning people used for feudalism: You see, only certain people are great enough to use AGI responsibly/can be trusted to govern, so we cannot just give the power freely to the unwashed masses, which would use it irresponsibly. You can certainly trust us to use it to help humanity - pinky swear.

Re: OpenAI and Elon Musk

#523

Post by gandutraveler »

baobabKoodaa wrote: post #370
Falimonda wrote: post #364

Earlier quoted context omitted.

How about the possibility that OpenAI and the Albanian government - along with many other governments - have a relationship of which you're unaware?

Uh huh. And they arrived at this 5.5 number how? It's bullshit.

[flagged]

Re: OpenAI and Elon Musk

#524

Post by eastbound »

Nevermark wrote: post #518
elif wrote: post #56

Earlier quoted context omitted.

>they may be right that open sourcing AGI would lead to too much danger. I think this part is proving itself to be an understandable but false perspective. The hazard we are experiencing with LLM right now is not how freely accessible and powerfully truthy it's content is, but it is precisely the controls upon it which are trying to be injected by the large model operators which are generating mistrust and a poor und…

> it is precisely the controls upon it which are trying to be injected by the large model operators which are generating mistrust I would prefer things were open, but I don’t think this is the best argument for that Yes, operators trying to tame their models for public consumption inevitably involves trade offs and missteps But having hundreds or thousands of equivalent models being tuned to every narrow mindset is t…

> Yes, operators trying to tame their models for public consumption

*Falsify reality.

Re: OpenAI and Elon Musk

#525

Post by pests »

seydor wrote: post #394

The timestamps of the emails are interesting. apparently they work at 3:40 AM but also at midday and at 8 AM

Life becomes work for these types of people, I'm sure there's no hard line when at work or at home, esp during the early days or the early frenzy.

Re: OpenAI and Elon Musk

#526

Post by gandutraveler »

advael wrote: post #113

Like most airing out of drama, this doesn't really materially change anything aside from making everyone involved look extremely bad But also, all entities involved looked bad already to observers who weren't so deeply rooting for them already that nothing would change their minds, so this is basically tabloid fodder drama in terms of importance as far as I can tell

Agree. But this showed Elon to be a much worse cry baby.

Re: OpenAI and Elon Musk

#527

Post by ben_w »

scarygliders wrote: post #483
ben_w wrote: post #433

Earlier quoted context omitted.

> Counter-counterpoint: absolutely nobody who has unguardrailed Stable Diffusion installed at home for private use has ever asked for more guardrails. Not so. I have it at home, I make nice wholesome pictures of raccoons and tigers sitting down for Christmas dinner etc., but I also see stories like this and hope they're ineffective: https://www.bbc.com/news/world-us-canada-68440150

Unfortunately you've been misled by the BBC. Please read this: https://order-order.com/2024/03/05/bbc-panoramas-disinformat... Those AI generated photos are from a Twitter/X parody account @Trump_History45 , not from the Trump campaign as the BBC mistakenly (or misleadingly) claim.

> Those AI generated photos are from a Twitter/X parody account @Trump_History45 , not from the Trump campaign as the BBC mistakenly (or misleadingly) claim.

They specifically said who they came from, and that it wasn't the Trump campaign. They even had a photo of one of the creators, whom they interviewed in that specific piece I linked to, and tried to get interviews with others.

Re: OpenAI and Elon Musk

#528

Post by honeybadger1 »

enriquec wrote: post #208
ajisiekskaiek wrote: post #153

Earlier quoted context omitted.

You are advocating here for an unresolvable set of ethics, which just happens to be one that conveniently leaves abuse of AI on the table. You take as an axiom of your ethical system the absolute right to create and propagate in public these AI technologies regardless of any externalities and social pressures created. It is of course an ethical system primarily and exclusively interested in advancing the individual a…

[flagged]

[flagged]

Re: OpenAI and Elon Musk

#529

Post by lupusreal »

ben_w wrote: post #434
BriggyDwiggs42 wrote: post #308

Earlier quoted context omitted.

What kind of damage can you do with a current day llm? I’m guessing targeted scams or something? They aren’t even good hackers yet.

The moment they are good hackers, everyone has a trivially cheap hacker. Hard to predict what that would look like, but I suspect it is a world where nobody is employing software developers because a LLM that can hack can probably also write good code. So, do you want future LLMs to be restricted, or unlimited? And remember, to prevent this outcome you have to predict model capabilities in advance, including "tricks"…

Use the hacking LLM to verify your code before pushing to prod. EZ
Post reply on HN