Comment by tcgv | Hacker News Reader

Comment by tcgv | Hacker News Reader

This post is a lame PR stunt, which will only add fuel to the fire. It tries to portray openAI as this great bennefactor and custodian of virtue:

> “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...”

> I've seen you (and Sam and other OpenAI people) doing a lot of interviews recently extolling the virtues of open sourcing AI, but (...) There are many good arguments as to why the approach you are taking is actually very dangerous and in fact may increase the risk to the world

How lucky are we that openAI, in its infinite wisdom, has decided to shield us from harm by locking away its source code! Truly, our digital salvation rests in the hands of corporate benevolence!

It also employs outdated internal communication (selectively) to depict the other party as a pitiful loser who failed to secure openAI control and is now bent on seeking revenge:

> As we discussed a for-profit structure in order to further the mission, Elon wanted us to merge with Tesla or he wanted full control.

> 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”. [4]

If their case to defend against Elon's action relies on his "Yup"from 2016, and justifications for being able to compete with Google, it's not a strong one.

I also thought this was kind of entertaining in their method to try and tar and feather. The highly selective (much dated) e-mail comms in a very roughly packaged statement. If that's how they are trying to protect their public image it doesn't sell their position strongly if anything it makes them look worse. It looks very amateurish almost childish to be honest.

A bunch of rich people who are all full of shit pissing on each other. Nothing to see here, keep scrolling.

Assume good intentions and nothing of substance being hidden: Is there any way to be transparent here that would have satisfied you or are you essentially asking for them to just keep to themselves or something else?

Oh, honey. Never assume good intentions when lawyers are involved

That wasn't the point of the question. The question was a hypothetical to test if there was any possible response that would've satisfied the original poster.

They're not suggesting to assume good intentions about the parties forever. They're just asking for that assumption for the purposes of the question that was asked

There is no satisfying answer if your actions before are not satisfying. The question implies that the original poster cannot be satisfied, and thus shift the blame implicitly. The problem remain not what the answer is, or how it is worded. The answer only portrays the actions, which are by themselves unsatifsying.

The answer is no. Companies don’t do things out of good intentions when lawsuits are involved.

A great way to be transparent would be to admit that some enormous egos prevented work that should be open from being open, and counterfactually opening them up. Sure, it may piss off microsoft, but statistically things that piss off microsoft also has been great for the world at large.

But that will never happen, will it?

I thought it was the fact that they needed a lot of money to train AGI.

Everyone seems to agree that's the case. Do you have evidence that it's not?

What does “train AGI” even mean dude

It means to run the back-prop algorithm to train neural networks to solve the problem(s) of AGI.

Dude.

Unless the problem-in-question is Adjusted Gross Income, I don't think you're talking about a well-defined issue in the first place. That's kinda the problem; even half-definitions like "meets or exceeds human performance in all tasks" doesn't specify what those tasks are or what human performance represents.

In other words, targeting "AGI" is a ill-defined goalpost that you can move anywhere you want, and shareholders will be forced to follow. Today it's about teaching LLMs to speak French, next week they'll need 7 billion dollars to teach it to flip pancakes.

GOOD GRIEF. You're right, AGI is not well defined. But it is perfectly well defined for the purposes of this conversation.

Working in this space (however you want to say it) costs a lot of money. Everyone knows this. You nitpicking definitions does not change that.

> Working in this space (however you want to say it) costs a lot of money.

Not necessarily? OpenAI deliberately chooses to scale their models vertically, which incumbents like Google and Meta had largely avoided. Their "big innovation" was GPT-4, an LLM of unprecedented scale that brute-forced it's way to the top. The Open Source 72b+ models of today generally give it a run for it's money, the party trick is over.

Nitpicking the definitions is important, because for OpenAI it's unclear what comes next. Consumers don't know what to expect, portions of the board appear to be in open rebellion, and our goal of AGI remains yet-undefined. I simply don't trust OpenAI (or Elon, for that matter) to do the right thing here.

> Not necessarily?

Well then, I expect we'll see a bunch of small projects that beat the big players any day now. I won't hold my breath.

Again you haven’t a clue of the point here. They were created as a research lab and instead after taking a paper created at Google they went closed source and are simply scaling that instead of working to create new and improved models that can actually run on reasonable hardware.

Small transformers being able to beat the same models but scaled up is unrelated to anything being discussed and you just seem like a fanboy at this point

> Again you haven’t a clue of the point here

I guess we're just going to have to agree to disagree.

We can’t agree to disagree about something that is a fact and you’re wrong about.

LOL.

You're basically describing what happened to GPT-2 when T5-flan came out. Not to mention, the incumbent model at the time (BERT) was extremely competitive for it's smaller size. Pretty much nobody was talking about OpenAI's models back then because they were too large to be realistically useful.

So yeah, I do actually anticipate smaller projects that devalue turtle-tower model scaling.

Cool. Send me a note when that happens.

No you missed the point in that the definition being vague, and an implication of magical tooling, was my point when I asked what that even means. By saying this they can now right off any criticism and people like you clearly eat it up.

No it’s a scapegoat to justify doing whatever and using a meme word so that sci-fi fans will accept that whatever. The fact you’re eating it up here is pure cringe, grow up. These people took a non-profit with an egalitarian mission and reversed course once they saw they could make fuck you money. The “AGI” excuse is one only the immature are buying. Dude.

I don't remember reading an exchange like this on HN ever. Either I wasn't paying attention enough or it's just demographic changing? I don't want to start an argument with either of you but it's painful to see US literally being split in half, even when you watch a movie, you can reasonably guess which side the filmmaker is standing depending on the story, narrative, and perhaps the ending. Same for any other outlet that you can think of, including comments on HN, methinks.

I don't like being insulted and respond in kind. Tit-for-tat is a very effective game theory strategy.

No one insulted you, you used a meme as an excuse for a company that completely went against its founding principles and I called you out on it.

We're just going to have to agree to disagree, friend.

You take care now.

Well I’ve been here a few years longer than you and I can assure you topics like this always get this heated.

The parent comment isn't entirely wrong, though. There are reasonable, safe and productive degrees of curiosity, and their are unreasonable, unsafe and antiproductive degrees too.

AI itself is not worthless, the goal of advancing machine-learning and computer vision has long-since proved it's worth. Heralding LLMs as the burgeoning messiah of "AGI" (or worse yet, crowning OpenAI) is a bald-faced hype machine. This is not about space exploration or advancing some particular field of physics along a well-defined line of research. It's madness plus marketing, and there's no reasonable defense of their current valuation. At least Tesla has property worth something if they liquidate, OpenAI is worth nothing if they fail. Investing in their success is like playing at a roulette wheel you know is rigged.

In this context (Elon v. OpenAI), I don't see how this seems like a lame PR stunt. They are defending their stance against Elon's BS. He's been spewing BS over the past couple of years saying he funded them to make the open source while he's always wanted the tech for his company and always meant to make it a closed tech with the intention to compete against Google. Elon literally mentioned that OpenAI should merge with Tesla and use it as a cash cow. After reading all of this, you think OpenAI's response is a PR stunt? What about Elon's lies so far? If anything, this drama just details out Elon's hypocrisy.

So, where is the source code of Grok, the LLM that Elon is building?

Did Twitter become a nonprofit while I wasn't paying attention?

Well, they aren't profitable to Elon at least

Did Elon ever announce, or even imply, that Grok would be open source?

Reply on news.ycombinator.com