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

It's OpenAI's world, we're just living in it

stratechery.com

Post reply on HN
Print view View original

211–220 of 284 posts

Re: It's OpenAI's world, we're just living in it

#211

Post by dragonwriter »

karel-3d wrote: post #83

I don't like Ed Zitron automatic dismissal of everything AI and the constant profanities in his writing are getting old, and it's usually not very well structured, but that said... I like the perspective he has about the money involved. https://www.wheresyoured.at/the-case-against-generative-ai/ OpenAI needs 1 trillion dollars in next four years just to keep existing. That's more than all currently available capital,…

> OpenAI needs 1 trillion dollars in next four years just to keep existing. That's more than all currently available capital, together, from all private equity.

Recent estimates I've seen of uninvested private equity capital are ~$1.5 trillion and total private equity $5+ trillion, with several hundred million in new pricate equity funds raised annually, so this simply seems incorrect even assuming either only currrent “dry powder" is considered, or only new funds available over the next four years, much less both and/or the rest of private equity.

Re: It's OpenAI's world, we're just living in it

#212

Post by rhetocj23 »

impossiblefork wrote: post #120
mattgreenrocks wrote: post #102

Earlier quoted context omitted.

All that money and they’ve been trying to shy away from talk of AGI recently to lower expectations.

The AI is getting better at things like maths. I recently asked it about iterating the Burrows-Wheeler transform, and it appeared to really understand that. It's not super easy to reason about why it's reversible, etc. and I felt that it got it. This is obviously not AGI, and we're very far from AGI as we can see by trying out these LLMs on things like stories or on analyzing text, or dealing with opposing arguments,…

My answer to this is - so what? What are the effects in the real economy?

Theres probably a good 2-3 years left of runway for substantial progress before things really fall off a cliff. There has to be some semblance of GDP (real) growth.

Re: It's OpenAI's world, we're just living in it

#213

Post by jcranmer »

B56b wrote: post #122
tptacek wrote: post #116

Earlier quoted context omitted.

He's saying more than that the companies are going to collapse; he's making pronouncements about the underlying technology, which are claims that are much harder to defend. I'm not entirely sure he understands the distinction between the companies and the technology, though.

Respectably...what?? Ed at this point is one of the most well read people on Earth for this topic. Of course he knows the difference between the companies and the technology. He goes in depth both on why he think the companies are financially unviable AND why he's unimpressed by LLM's technologically alllll the time.

Even as someone who is generally inclined to agree with his thesis, I find Ed Zitron's discussions as to why AI does not and will never work deeply unconvincing.

Re: It's OpenAI's world, we're just living in it

#214

Post by psychoslave »

yubblegum wrote: post #41
kllrnohj wrote: post #13

Platforms usually deliver significant value that is hard to replicate. OpenAI doesn't have any such thing. It's trivially replaced, and there's many competitors already. OpenAI is ahead of the curve, but they don't seem to have any particular way to do sticky capture. Migrating to a different LLM is an afternoon's work at most, not nearly the complexity of porting an app between OS' or creating a robust hardware driv…

> Migrating to a different LLM is an afternoon's work at most, not nearly the complexity of porting an app between OS' or creating a robust hardware driver model. I question this. Each vendor's offering has its own peculiar prompt quirks, does it not? Theoretically, switching RDBMS vendors (say Ora to Ingress) was also "an afernoon's work" but it never happened. The minutia is sticky with these sort of almost-but-not…

No, changing rdbms is a totally different challenge. They provide predictible reproducible output which are very sensitive and brittle to the slightest minor change in the input, while with LLM you expect the exact opposite.

Changing the LLM backend in some IDE is as complicated as selecting an option in a dropbox for those who integrate such a feature. They are other scenarios where it might be a bit more complicated to transition of course, but that's it.

Re: It's OpenAI's world, we're just living in it

#215

Post by impossiblefork »

rhetocj23 wrote: post #212
impossiblefork wrote: post #120

Earlier quoted context omitted.

The AI is getting better at things like maths. I recently asked it about iterating the Burrows-Wheeler transform, and it appeared to really understand that. It's not super easy to reason about why it's reversible, etc. and I felt that it got it. This is obviously not AGI, and we're very far from AGI as we can see by trying out these LLMs on things like stories or on analyzing text, or dealing with opposing arguments,…

My answer to this is - so what? What are the effects in the real economy? Theres probably a good 2-3 years left of runway for substantial progress before things really fall off a cliff. There has to be some semblance of GDP (real) growth.

I think it might be possible to automate development of short programs. I also think it might be possible to reduce the confusion, misunderstandings and cliches, so that the models can work for longer stretches.

But people probably expect to get the next version for what they pay in subscriptions now, so I can't imagine much more revenue growth for the model companies.

Re: It's OpenAI's world, we're just living in it

#218

Post by beeflet »

llbbdd wrote: post #150
exolymph wrote: post #132

Earlier quoted context omitted.

Homelessness and world hunger are not bottlenecked by money. There's a stronger case for healthcare, but that's also substantially a political issue.

There's a stronger case for world hunger being bottlenecked than healthcare. World hunger is a logistics problem now, but no amount of money lets you print doctors.

You can't just throw money at the world hunger problem, it will end up in some warlord's coffers. Hunger still exists because it is politically useful to keep people hungry.

Re: It's OpenAI's world, we're just living in it

#219

Post by beeflet »

llbbdd wrote: post #176
dgfitz wrote: post #165

Earlier quoted context omitted.

With a little bit of lag time (school) we could have a metric fuckton of doctors. We have a metric fuckton of shitty lawyers. Doctors are artificially gated in the US What’s the joke? “What do you call the person who graduated last in their class from med school? A doctor.”

Sure, I think that helps my point. You can't create "good" doctors out of thin air with money, just "more" doctors, and it takes forever.

Do we really need "good" doctors all of the time? I think the industry is overregulated to restrict the supply of healthcare.

Re: It's OpenAI's world, we're just living in it

#220

Post by beeflet »

vbezhenar wrote: post #110
moistoreos wrote: post #100

Earlier quoted context omitted.

You could literally eliminate homelessness, world hunger, and fully pay health care costs in that same timeframe with much less cash.

Cash is just paper. You can't eat paper.

Time is money. the investment of time and resources spent on AI could be spent producing food or healing people.

Not that I think we should, but the "Cash is just paper" attitude makes no sense. If it's just paper, how is OpenAI training AI using just paper?

Post reply on HN