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

GLM 5.2 and the coming AI margin collapse

martinalderson.com

Post reply on HN
Print view View original

471–480 of 495 posts

Re: GLM 5.2 and the coming AI margin collapse

#471

Post by gexla »

Add me to the list of skeptics.

1) A lot of companies simply won't trust Chinese models

2) These companies will still trust a direct frontier provider over a non Chinese company hosting Chinese models.

3) Frontier model providers working with cloud hosting providers for data residency.

4) Hosting your own models is expensive (hardware, people to setup and maintain, risk of little to no return.) And smaller companies likely won't go this route.

5) As mentioned in the article, a good but cheap model can still be expensive if it takes longer to do a task. I imagine I could save days doing work with Fable vs using Chinese models.

ETA: And loads of demand. Not sure the pricing situation is yet settled. As Chinese models get more popular, they may have to raise their prices as well. US models may struggle more with meeting demand rather than getting margins squeezed.

Re: GLM 5.2 and the coming AI margin collapse

#472

Post by peab »

saberience wrote: post #269
peab wrote: post #107

Earlier quoted context omitted.

It really depends on what you're doing, but most LLM usage and agentic runs are pretty interchangeable in my experience, and it's usually trivial to switch. If anything, you're better off supporting multiple LLMs as backup because most model providers have been so inconsistent with working all the time

Dude it’s not trivial to switch because the behaviors are different! You’re clearly not building a product based on an LLM. I’m still using various old Anthropic and OpenAI models for products I’ve built and released because I can’t risk the behavior changing in unpredictable ways and the users being pissed. It’s much easier to switch out some deterministic software than an LLM which you’ve spent a ton of time on tes…

I've been building multiple products with LLMs, and they are in fact interchangeable for the most part.

In fact, most benchmarks show this! Most benchmarks have similar performance for the same classes of models.

On top of this, there are tools like open router, or even the openai SDK which trivially allows you to swap endpoints for the LLM!

If you're using the agents SDK from openai or something, then yeah it's not interchangeable but that's you doing it wrong

Re: GLM 5.2 and the coming AI margin collapse

#473

Post by jononor »

sajithdilshan wrote: post #470
jononor wrote: post #456

Earlier quoted context omitted.

EU already has supercomputers and data centers. Some examples: https://www.eurohpc-ju.europa.eu/supercomputers/our-supercom... The goal is to triple the number of data centers over the next years. It is not like there is nothing going on...

Again, super computers are not used to train frontier models. Please do your own research before blobbing just random links. By the time they triple those data centres the AI race is already lost

GB200 and GB300 are decent for training LLMs? H100 can be used also? Though the current European deployments are rather small, even the biggest are just some thousand GPUs.

Re: GLM 5.2 and the coming AI margin collapse

#474

Post by oblio »

ben_w wrote: post #386
irishcoffee wrote: post #331

Earlier quoted context omitted.

Stand up against tech? Cut off your nose to spite your face? Apple sure won’t care. Nor is Apple all encompassing of “tech”

> Apple sure won’t care Europe collectively is about 26.7% of their 2025 revenue, according to SEC filings, so I bet they'd care. https://www.sec.gov/Archives/edgar/data/320193/0000320193250...

It's worse than that. If Apple were banned, someone else would fill their spot.

Re: GLM 5.2 and the coming AI margin collapse

#475

Post by Centigonal »

etdznots wrote: post #251
Centigonal wrote: post #138

Earlier quoted context omitted.

For 2 and 3, office software and OSs have strong network effects and up-stack effects, just like CPU instruction sets. Also, I'm sorry, but OSS office suites compete with Office and GSuite the way grocery store frozen pizza competes with Domino's and Papa John's. Quality and completeness of execution matter a lot in that category.

Lol, your example of the other side of grocery store pizza is Domino’s and Papa John’s? I guess in offices where M$ products are used the people there think mmm yumm dominos and hold up their noses at digiornos lol.

That was intentional - I think Office (or, uh, Microsoft 365 Copilot as it is now called) is frustrating mass-market mediocre software! That said, it consistently lets me do my job, while LibreOffice is often unusable for my purposes.

Re: GLM 5.2 and the coming AI margin collapse

#476

Post by npodbielski »

Certhas wrote: post #374
cpursley wrote: post #252

Earlier quoted context omitted.

Because nobody’s paying for tokens, they’re paying for monthly plans and right now those are still a better bargain.

Individuals, sure. For enterprise you can't get monthly plans. You have to pay per token. It's a bit like saying "nobody pays for Microsoft Office". I certainly don't know anyone personally who has. Students get a free Education License and then your employer provides one for you...

Strange. This is what they got at my company. I do not remember anyone mentioning paying for tokens. Maybe because it is fairly small, couple of hundred people in IT.

Re: GLM 5.2 and the coming AI margin collapse

#477

Post by gnarbarian »

yogthos wrote: post #71
gnarbarian wrote: post #66

the economics of this are a little counterintuitive. is there a market saturation point for intelligence? how about for software? it seems like the more you have the more you want because you're trying to do more things. as the models get smarter I get busier because I'm doing more things...

There's definitely a saturation point depending on the complexity of the problem you're solving. For example, any model can write a small shell script to resize a video with ffmpeg for you right now, so it doesn't matter whether you're using a local Qwen model, GLM, or Fable. They'll all do a roughly comparable job and you'll end up with a working script that does what you need. Then you have things like CRUD apps, w…

if you give me a smarter model I will find a more complex problem for it to work on that will require an exponential number of sub-agents.

the smarter the model the more capable it is to marshal resources to accomplish the goal.

this means you don't need a smart person to be able to take advantage of this. a smart model will be able to do the same.

Re: GLM 5.2 and the coming AI margin collapse

#478

Post by Certhas »

npodbielski wrote: post #476
Certhas wrote: post #374

Earlier quoted context omitted.

Individuals, sure. For enterprise you can't get monthly plans. You have to pay per token. It's a bit like saying "nobody pays for Microsoft Office". I certainly don't know anyone personally who has. Students get a free Education License and then your employer provides one for you...

Strange. This is what they got at my company. I do not remember anyone mentioning paying for tokens. Maybe because it is fairly small, couple of hundred people in IT.

Officially it's not available anymore. But there are team plans that are not enterprise, so if you are small enough and fine with the data protection included in those maybe that is what you're working with? And I do know NGOs were offered seat based plans after they were officially not available anymore.

Also: This change came in in March so if you got your contract before then this will only bite once you renew.

https://support.claude.com/en/articles/9797531-what-is-the-e...

Re: GLM 5.2 and the coming AI margin collapse

#479

Post by ywvcbk »

pavlov wrote: post #313
logicchains wrote: post #277

Earlier quoted context omitted.

Europe became powerful before it was unified, and ever since the creation of the EU it's been becoming less and less important on the world stage.

So you’re saying that the war-ravaged Europe of 1946 that was split by the Iron Curtain and needed Marshall Aid was more powerful and important than today’s EU? Insane take. But somehow people will go to any lengths to disparage the EU.

Globally in relative terms? Kind of. Sure the gap to the US was huge but it still controlled pretty much entire Africa and much of Asia (with China and Japan obviously not doing great..)

Re: GLM 5.2 and the coming AI margin collapse

#480

Post by cbg0 »

echelon wrote: post #388
cbg0 wrote: post #258

Earlier quoted context omitted.

> We're accelerating at lightning speed now. Accelerating how much slop you can output? A better model will still produce slop for your feature factory that pumps out software which nobody is interested in buying.

I'm at a 3M run rate in 4 months. Because I'm solving real problems. You don't seem like an entrepreneur. Why are you on HN? YC wants people to build AI startups. You're here shitting on them. Half of this community is. You're all a bunch of old men grumpy at the new tools. I'll offer my own analysis: if you're not using AI very effectively, you won't have a career in computing in a few years.

I'm neither an old man, nor am I grumpy at the new tools. I use AI daily, I'm just not deluded into thinking AI is more capable than it is, or myself for that matter.
Post reply on HN