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

401–410 of 1001 posts

Re: OpenAI and Elon Musk

#401

Post by goatlover »

MrScruff wrote: post #376
lagt_t wrote: post #105

Earlier quoted context omitted.

Everytime they say LLMs are the path to AGI, I cringe a little.

I don’t know if they are or not, but I’m not sure how anyone could be so certain that they’re not that they find the mere idea cringeworthy. Unless you feel you have some specific perspective on it that’s escaped their army of researchers?

Because AI researchers have been on the path to AGI several times before until the hype died down and the limitations became apparent. And because nobody knows what it would take to create AGI. But to put a little more behind that, evolution didn't start with language models. It evolved everything else until humans had the ability to invent language. Current AI is going about it completely backwards from how biology did it. Now maybe robotics is doing a little better on that front.

Re: OpenAI and Elon Musk

#402

Post by jakupovic »

BriggyDwiggs42 wrote: post #322
dkjaudyeqooe wrote: post #133

Earlier quoted context omitted.

The benefit is the science, nothing else matters, and having OpenAI decide what matters for everyone is repugnant. Of course they can give us nothing, but in that case they should start paying taxes and stop claiming they're a public benefit org. My prediction is they'll produce little of value going forward. They're too distracted by their wet dreams about all the cash they're going to make to focus on the job at ha…

I agree with your sentiment but the prediction is very silly. Basically every time openai releases something they beat the state of the art in that area by a large margin.

We have a saying:

There is always someone smarter than you.

There is always someone stronger than you.

There is always someone richer than you.

There is always someon X than Y.

This is applicable to anything, just because OpenAI has a lead now it doesn't mean they will stay X for long rather than Y.

Re: OpenAI and Elon Musk

#403

Post by leereeves »

pama wrote: post #211
mac-attack wrote: post #127

Earlier quoted context omitted.

With all due respect, they are not doing a sleight of hand while selling widgets... they are in the process of reshaping society in a way that may have never been achieved previously. Finding out that their initial motives were a facade doesn't portend well for their morality as they continue to gain power.

I still don’t understand why you think their initial motives were a facade. They have always been trying to get to AGI that will be useable by a large fraction of society. I am not sure this means they need to explain exactly how things work at every step along the way or to help competitors also develop AGI any more than Intel or Nvidia had to publish their tapeouts in order for people to buy their chips or for comp…

At one time their mission statement said:

> OpenAI is a non-profit artificial intelligence research company. Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return. Since our research is free from financial obligations, we can better focus on a positive human impact.

That's obviously changed to "let's make lots of money", which should not be any "non-profit" organization's mission.

Re: OpenAI and Elon Musk

#404

Post by ascorbic »

Barrin92 wrote: post #353
Falimonda wrote: post #309

Earlier quoted context omitted.

Why is that such an insane thing for them to say?

because it's a completely made up claim with a random number thrown in. How do you end up with "5.5" years for a process that doesn't have a timetable, number of estimated emails required for EU accession divided by number of emails ChatGPT can generate per day?

I was pretty skeptical about this too, but looking into it there is some basis in fact. They're using it to help translate and integrate the "Acquis communautaire" - the huge body of EU laws and regulations that need to be enshrined in national laws of candidate countries. This is one of the most time-consuming parts of the process, and usually takes many years. Leaving aside how risky this is (presumably they will have checks in place), I can see how this could save years of work. Saying 5.5 years is ridiculous false precision though.

It won't help with the toughest part though, which is the politics of the other member states.

Re: OpenAI and Elon Musk

#405

Post by FeepingCreature »

goatlover wrote: post #386
FeepingCreature wrote: post #354

Earlier quoted context omitted.

What evidence do we have that it's impossible or even just very unlikely?

We don't have any evidence other than billions of biological intelligences already exist, and they tend to form lots of organizations with lots of resources. Also, AIs exist alongside other AIs and related technologies. It's similar to the gray goo scenario. But why think it's a real possibility given the world is already full of living things, and if gray goo were created, there would already be lots of nanotech tha…

The world we live in is the result of a gray goo scenario causing a global genocide. (Google Oxygen Holocaust.) So it kinda makes a poor argument that sudden global ecosystem collapses are impossible. That said, everything we have in natural biotech, while advanced, are incremental improvements on the initial chemical replicators that arose in a hydrothermal vent billions of years ago. Evolution has massive path dependence; if there was a better way to build a cell from the ground up, but it required one too many incremental steps that were individually nonviable, evolution would never find it. (Example: 3.7 billion years of evolution, and zero animals with a wheel-and-axle!) So the biosphere we have isn't very strong evidence that there isn't an invasive species of non-DNA-based replicators waiting in our future.

That said, if I was an ASI and I wanted to kill every human, I wouldn't make nanotech, I'd mod a new Covid strain that waits a few months and then synthesizes botox. Humans are not safe in the presence of a sufficiently smart adversary. (As with playing against Magnus Carlsen, you don't know how you lose, but you know that you will.)

Re: OpenAI and Elon Musk

#406

Post by Falimonda »

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.

You cannot fathom how a government official may have told them that they've managed to accelerate their EU accession through the help of LLMs by a specific number of months?

Here's a speculative scenario that doesn't seem so insane:

Albania has a roadmap for EU accession. The roadmap is broken up into discrete tasks. The tasks have estimates for time to completion. They've been able to quickly hack away at 5.5 years worth of tasks using LLMs.

Your problem with the statement is that they didn't provide a source. Maybe express interest in the facts that might support that instead of freaking out over perceived insanity.

Re: OpenAI and Elon Musk

#407

Post by BlueTemplar »

flohofwoe wrote: post #382
DrSiemer wrote: post #357

Earlier quoted context omitted.

So you still believe the open internet has any chance of surviving what is coming? I admire your optimism. Reliable information and communication will soon be opt-in only. The "open internet" will become an eclectic collection of random experiences, where any real human output will be a pleasant, rare surprise, that stirs the pot for a short blip before it is assimilated and buried under the flood.

The internet will be fine, social media platforms will be flooded and killed by AI trash, but I can't see anything bad with that outcome. An actually 'open internet' for exchanging ideas with random strangers was a nice utopia from the early 90's that had been killed long ago (or arguably never existed).

E-mail has *not* been killed long ago (aside some issues trying to run your own server and not getting blocked by gmail/hotmail).

It is under threat now, due to the increased sophistication of spam.

Re: OpenAI and Elon Musk

#408

Post by sigmoid10 »

leereeves wrote: post #388
sigmoid10 wrote: post #313

Earlier quoted context omitted.

There is very little in terms of rigorous mathematics on the theoretical side of this. All we have are empirics, but everything we have seen so far points to the fact that more compute equals more capabilities. That's what they are referring to in the blog post. This is particularly true for the current generation of models, but if you look at the whole history of modern computing, the law roughly holds up over the l…

More compute also requires more data - scaling equally with model size, according to the Chinchilla paper. How much more data is available that hasn't already been swept up by AI companies? And will that data continue to be available as laws change to protect copyright holders from AI companies?

It's not just the volume of original data that matters here. From empirics we know performance scales roughly like (model parameters)*(training data)*(epochs). If you increase any one of those, you can be certain to improve your model. In the short term, training data volume and quality has given a lot of improvements (especially recently), but in the long run it was always model size and total time spent training that saw improvements. In other words: It doesn't matter how you allocate your extra compute budget as long as you spend it.

Re: OpenAI and Elon Musk

#409

Post by ben_w »

eightman wrote: post #374
TaylorAlexander wrote: post #275

Earlier quoted context omitted.

The biggest real problem I’m experiencing right now isn’t controls on the AI, it’s weird spam emails that bypass spam filters because they look real enough, but are just cold email marketing bullshit: https://x.com/tlalexander/status/1765122572067434857 These systems are making the internet a worse place to be.

The use case for AI was, is and always will be spam.

For language models, spam creation/detection is kinda a GAN even when it isn't specifically designed to be: a faker and a discriminator each training on the other.

But when that GAN passes the human threshold, suddenly you can use the faker to create interesting things and not just use the discriminator to reject fakes.

Re: OpenAI and Elon Musk

#410

Post by eftychis »

nprateem wrote: post #306
eftychis wrote: post #222

Reading the messages: This is a marketing step of course, no sane lawyer would agree to this. And that is because, I don't think they show what they think and want to show. That at some point Elon had an opinion that a lot of money is needed or that OpenAI maybe had no future? That does not change the duty or obligations of a non-profit to the mission. Also, it is clear some important information has been blacked out…

And what would happen? 2 minutes after being dissolved, ClosedAI is founded and hires all the employees.

Perhaps even before the dissolution.

But the science and IP become public and open or under a non-profit that is tasked with opening them. And for-profit segments are stripped of any exclusive rights that arose from the OpenAI.

The irony is that the dissolution, overseeing by third special referee or permanent injunction (e.g. the Musk suit) of OpenAI is the only ways OpenAI is "opening their AI."

Post reply on HN