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

I love LLMs, I hate hype

geohot.github.io

Post reply on HN
Print view View original

81–90 of 340 posts

Re: I love LLMs, I hate hype

#81

Post by SwellJoe »

This line: "this is my main argument against the valuation of frontier labs. It’s not that AI won’t create that much value, it’s that they won’t capture it."

That is a very astute and concise way to explain everything about how the frontier labs are behaving and how they're trying to push more people to pay token rates for the best models. At the current subscription prices ($100 or $200 a month for a generous, though bounded, amount of tokens), frontier models are a no-brainer, most folks and companies will use them. But, at token rates, 10x or 100x the cost of open models or what I was spending on the frontier models a month ago? That is a harder question to answer "yes" to. I certainly wouldn't spend $1000 a month for the best model, much less $10,000; my employer might pay $1000/month, but definitely not $10,000. The frontier labs need everyone to answer "yes" to spending 100x what they currently spend to justify the valuations, and it's just not going to happen as long as everyone knows how to make these models.

Both OpenAI and Anthropic are trying to figure that out now. Anthropic, in particular, has their finger on the trigger...they want to push people to usage-based billing for Fable. But, OpenAI released 5.6 Sol, competitive with Fable (or close enough), and it's available via subscription (even the $20 subscription!), and there's no moat keeping someone from switching. If Anthropic really does end Fable access on the subscription plans in a few days, I predict a large market move back toward OpenAI.

The market isn't going to bear the cost of making the frontiers investment make sense.

Re: I love LLMs, I hate hype

#82

Post by jagenabler2 »

Razengan wrote: post #40
jagenabler2 wrote: post #35

Earlier quoted context omitted.

Do you feel like the ideas you’re getting from brainstorming these days are the same level of quality as in the past? I’ve been doing some of the same, but I’ve also been feeling like the downtime where I’m genuinely stuck is where my most innovative solutions come to light. I’m not going as deep into problem spaces anymore. I’ve also lost my ability to self-filter. In the past, I’d write down an idea and if I was st…

> Do you feel like the ideas you’re getting from brainstorming these days are the same level of quality as in the past? You have to be careful and "remain yourself": Like I've been trying to think of a generic save/load system for my game framework, but the ideas given by Codex so far don't suit my desired design/interface, BUT it makes me certain of how I DON'T want to do it heh If I got lazy and just blindly took t…

Thanks, that makes sense! I’m realizing that’s how I’ve been learning to approach it too and seeing the best results.

Re: I love LLMs, I hate hype

#83

Post by fragmede »

andy99 wrote: post #48

He says he might have been too harsh in his “eternal sloptember” post from may: https://geohot.github.io/blog/jekyll/update/2026/05/24/the-e... I wonder what he thinks was too harsh, still seems pretty bang on, I think it’s going to age well.

> the adoption of AI agents into software development will be one of the most costly mistakes in the field’s history. Agents cannot program, and it’s taking longer and longer to realize that they can’t.

I think he now thinks agents can maybe program a little bit.

Re: I love LLMs, I hate hype

#84

Post by fragmede »

pphysch wrote: post #73
kordlessagain wrote: post #2

There's good reason to hate the merchants and their marketing. But builders are not merchants. They build with whatever tool is available.

Geohot is one of the (attempted) merchants, but maybe that is not going so well and he is changing his tune.

Not sure why you had to add the (attempted) qualifier. He started a company and is selling a box. That makes him a merchant. How successful that venture is, is a different question, but he absolutely is a merchant in this arena.

Re: I love LLMs, I hate hype

#85

Post by wizzwizz4 »

ghthor wrote: post #77
fwlr wrote: post #58

I get it, I want to agree, I really do like the “this is a new tool in the toolkit of the professional software craftsperson” argument… …but consider: the Q-tip. “Don’t use it to clean your ears”, but for most people that’s all they want to do with it, and empirical observation indicates that this dynamic results in either “using Q-tips irresponsibly” or “not using Q-tips”, with “uses Q-tips properly” being a small-t…

Qtips are made for cleaning your ears. It says not to do that so they are NOT sued every time some idiot fucks up their ear with one.

But the part of the ear that needs cleaning can be reached without a cotton bud. This is like shoving a sponge down your windpipe to remove mucus.

Re: I love LLMs, I hate hype

#86

Post by sixtyj »

hamandcheese wrote: post #10

> where’s all this new magical software that the productivity improvements should imply? It's running, privately, in my homelab. I think we are entering what I call the "have it your way" era. If an open source project doesn't do exactly what you want it to do, fork it, or create a new version. It's too easy. This makes me a bit concerned about the future of open source. Upstreaming used to be worth it, since maintai…

Creating a fork of an active project only makes sense if you are its sole user (of the fork) and you really need exactly the modification you've been dreaming of.

I have seen so many unnecessary forks of popular projects that I think it's better to stick with the original, even if that means it won't be perfect.

Re: I love LLMs, I hate hype

#87

Post by arscan »

SwellJoe wrote: post #81

This line: "this is my main argument against the valuation of frontier labs. It’s not that AI won’t create that much value, it’s that they won’t capture it." That is a very astute and concise way to explain everything about how the frontier labs are behaving and how they're trying to push more people to pay token rates for the best models. At the current subscription prices ($100 or $200 a month for a generous, thoug…

Who is going to end up capturing all this value being generated is going to be very interesting. Back in 1980, who’d have thought MS would capture the majority of the value from PCs over the next 3 decades, and not IBM?

Re: I love LLMs, I hate hype

#88

Post by jacobgold »

lukan wrote: post #70
jacobgold wrote: post #54

Earlier quoted context omitted.

Yes, LLM agents are "magic" in the sense that "any sufficiently advanced technology is indistinguishable from magic"[1] But it's not actually magic. Technical people understand that it's just software running on computers. 1. https://en.wikipedia.org/wiki/Clarke%27s_three_laws

Sure, nothing is magic. You can go look how a simple LLM works and build your understanding from there. But calling it "just software" is trivializing it in my opinion. I can write software, but I cannot write software that writes software.

> But calling it "just software" is trivializing it in my opinion

The bigger mistake would be trivializing the rest of the technology involved just because LLMs are the newest piece. LLMs are only "magic" because they're built on a stack that was already "magic" without them.

LLMs are impossible without:

- operating systems

- programming languages

- compilers

- data centers / power grids / air conditioning

- servers / switches / routers

- CPUs / RAM / GPUs / SSDs

- fiber networks

- etc

Re: I love LLMs, I hate hype

#89

Post by emp17344 »

simianthoughts wrote: post #46

This guy is sooooo annoying with his stale takes. This is what he wrote before. > I’m calling it now, the adoption of AI agents into software development will be one of the most costly mistakes in the field’s history. Agents cannot program, and it’s taking longer and longer to realize that they can’t. Now he's writng > I love the progress. I’m so excited for the new LLMs, self driving cars, video generation models, a…

You act as a hype peddler in practically every LLM-adjacent thread. No wonder you’re taking this article so personally. Touch grass.

Re: I love LLMs, I hate hype

#90

Post by reinitctxoffset »

password54321 wrote: post #20

Yeah I don't think any of the labs have some secret sauce for intelligence either. It seems most of the advancements are still coming from hardware, making LLMs more efficient and throwing more compute and data at problems. And even those problems still require a lot of prompt engineering: https://cdn.openai.com/pdf/04d1d1e4-bc75-476a-97cf-49055cd98...

I'm pretty sure at this point that Anthropic is training mixture models (at least in the heavy pre-train) and deploying them dense with explicit loss on thinking trace coherence.

Having a thinking trace that is legible, coherent, and immediately implies the explicit turn output and/or tool use seems difficult if not impossible to reliably get from mixture models.

I predict MoE is a transitional technology, it's got too many problems and the benefits are...kinda grandfathered into the dogma at this point.

Post reply on HN