HN user
Maybe it's silly, but from someone who is ignorant on this topics, what are the consequences of this kind of "discoveries"?
Is it something "revolutionary" or just another small brick that will pile up until something really "revolutionary" will happen?
Here we go.
It's one of the 2 endgames of all companies involved in the AI race:
- ADS
- make the users that pay for tokens find a PMF for a product/service and then make the product/service their self
We are still in time for the "revolution" ... but it will take longer and won't look like than most thought.
We still need a bubble pop to drive out all nonsense ideas and paths.
I guess it has to do with the fact that everything is greyscale nowadays ... did you notice there are no more coloured cars, and the few that are there are all faded.
And what if they use your data tò generate syntethic data to train on?
Coreweave and Nebius think this is a great business model.
It's irrelevant.
If their business model thinks they can make a profit doing it this way, why stop them?
I don't think someone needs to stop them, but there are some legit questions that need an answer:
- what happens to all these companies when growth decelerate or stop?
- what happens to nvidia stock when it has to buy back unused gpus?
- what are the risk that a sectorial financial crisis turn into a major economic crisis?
To make money on financial markets it's not just about "if" but also "when" ... and even if I think it's problem (which I didn't say) i for sure don't know "when"!
My understanding is that it's not about the money itself but the model:
- you fund a new company and sign long terms contracts with it - this new company uses the money you gave it and a lot of debt (backed by long term contracts) to build datacenters and buy a lot of GPU - your figures look great
What happens when they run out of debt or funds? If they reach some kind of profitability it's not a big deal, but if not ...
EDIT
Forget to mention the buyback of unused capacity problem: what happens to your figures when you have to buy back tons of unused GPUs?
I think it's the first time ever we don't see the dominant model being surpassed by new released concurent models.
Did anthropic found their moat or we hit a Wall?
If it's uptime it's definitively not much!
Wait to see what they are using for emails and for most of their internal docs (containing any kind of secrets)!
I know companies that will tell you "I'm not gonna put any of my data in cloud, especially not American ones" but they are perfectly fine using any major cloud based office suite (mail, docs, chat/video apps, ecc ecc) where they voluntarily and deliberately load any kind of data.
I agree that non-tech companies seem to be taking longer to see results from AI
Rewriting your app in rust won't increase your revenue, it will cost you more in terms of tokens and risk increasing.
especially from the AI labs
IMO this is one of the endgame for big ai LABS, they will allow and subsidize users to test and validate on their behalf and once there is a PMF they will step in.
Although the Council emphasizes that the *scans will be limited to the absolutely necessary extent* and that no general, indiscriminate surveillance will take place
I'm 100% sure that this is the case and about the good intentions of the proposers.
/s
It was more a critique to the crowd saying "Zitron is making wrong predictions since 2024" rather than Zitron itself.
There's at least $50B of LLM usage being paid for in 2026
Just out of curiosity, where did you get the 50B figure?
Hundreds of thousands of developers have already changed the way they were working for decades
I agree on this, but it doesn't mean that there is an automatic benefit on business side ... and business is what is paying our wages & tokens!
We are still in the discovery phase, but we don't know yet if there will be enough return to repay those hundreds of billions already invested and other few trillions that will be invested in the near future.
Just keep in mind that Zitron is in the business of content creation. It doesn't mean he's wrong[1] but his job isn't making successful predictions.
[1] Actually I think it's right on the overinvestments and ROI claims
It's just because not enough people had this very specific problem before.
This article will be part of the next model training set, and probably it will be able to solve it despite not understanding anything about world or not studying or thinking.
Actually it's almost 2 months /s
Even if "is just the same old shit" they have much more data and of a much higher quality to scale the RL pipeline.
Because you provide them with the "problem" and the "solution" and once you have both you can scale your RL pipeline.
Both parties are using the same upstream components so it ends up looking the same.
I had a look at the 4 screenshots in the post, and definitively it's not just using the "same upstream components", it's a verbatim copy.
Don't know about the rest of the app and the actual code, but I won't be surprised to find out it's basically the same.
how many other people have encountered a problem close enough to yours and solved it somewhere on the open internet
I'm 100% sure that all our web, cc, codex or whatsoever sessions are used in the training, RL or either both.
This makes the size of the universe models know about at least one order of magnitude bigger than the open internet.
Did they acquire also BMC?
Probably the first >20k$ macbook ever ...
Debt is directly tied with the ability to repay it, if the cash flow is enough to keep paying it, it's not a big issue, but thing can go horribly pretty fast if someone start having cash problems.
Mythos is a service you buy from a US company ... So this can be enforced better without running into thorny speech issues
Do you think AWS or Microsoft would be so dominant in cloud services and office productivity if the previous administrations had banned them for non-us firms?
In my opinion no, and all across the world we would be using non-us tech.
Precisely because it's a service, the US have the interest make the rest of the world continue to send their data to them.
So a little bit better than GPT 5.5 but 4 times more expensive (but obviously we don't know the true cost of these models).
PS: I was curious also about GLM 5.2.
AI-assisted coding is accelerating skill atrophy for developers who haven't built the foundations yet
I'd argue that this is true also for seasoned developers.
One thing I noticed (and annoys me a lot) after almost 1 year of ai-assisted coding on daily basis is that it's incredibly tiring to do the job yourself if the tool doesn't do what you want or how you want.