Hacker Newsnew | past | comments | ask | show | jobs | submit | ursuscamp's commentslogin

I don't necessarily buy that API prices represent "real" prices of the models.

Well of course they're higher than marginal cost. But these providers have to also generate a fat return on investment.

I moved one of our daily workflows over to Kimi K3 on Fireworks. It replaces two sales assistant positions, and was about $50/day for 14 million tokens total (in/out). They surely are not subsidizing this price as it is just inference only and other providers are even less money.

It's probably a safe assumption that openrouter prices for Kimi K3 are "real".

If they're hosted in China next to a coal power plant and maintained by people getting paid 1/10th of a US engineer sure.

I doubt opanai and anthropic will go that road


Not in the slightest as there are providers selling K3 for half of what OR has them listed for. (And maintaining profitability)

What do you think the real price is? Subscriptions are heavily subsidised, I don't know anyone who would deny that.

The general model for subscriptions is that power users are subsidized by subscriptions of casual users, like a gym membership or whatever.

This is a little dicier in post-agent AI, because it's easier for casual users to automate power-user consumption, but the providers have done decently in discouraging that.


There's people here saying they're obviously subsidized, there's people here saying they're obviously profitable. I think they're probably subsidized, but I would hold back on saying it's obvious.

Yes, they are ludicrously profitable - even when future training costs are taken into account!

[citation needed]

I don't want to host my code on a platform that has activist goals beyond open source advocacy, but I'm glad it exists because it's a useful filter, like the X/Bluesky divide.

I agree, and I’m hopeful that the people who leave will consider truly decentralized alternatives like Radicle. Especially for projects that might be more vulnerable/at risk to the whims of gatekeepers.

I really don't think you want to make an example of the "free speech app" that militantly censors pro-Palestine accounts while allowing AI-generated CSAM

Can someone explain to me the difference between this approach and using planning with a larger model, then just switching to a small model for implementation without clearing the context? I understand that it specifically does the first edit as well either way the larger model. Is there some other difference I am missing here?

When a frontier makes a succesfull edit based on the plan that it made, it leaves an procedural trace in turn biases the NEXT model, low cost model, straight into procedural action. The cheaper model doesn't need to reread everything again because it has enough information from the frontier model to complete the task. A simple "plan" of what needs to be done does not carry this information.

If I understand correctly, switching to a small model makes the small model read the context again.

You're in for a great deal of pain in the coming years.

Nasal phenylephrine is a miracle when I am trying to sleep with a stopped up nose. A spray in each nostril and my nose is clearer than even normal within a few minutes.


Bitcoin is so dead that jamesob is posting about AI.


Bitcoin booster -> AI slopper pipeline


Too late, damage is already done. We have been on the wrong timeline for years at this point.


SpaceX is planning to do data centers in orbit. Engineering issues aside, that might dovetail with Cursor nicely if timings work out.


I cannot believe full grown adult are taking that seriously


As long as the stock goes up, everyone is happy, right?


Iran gained international credibility by adhering strictly to the JCPOA, even long after the Trump admin broke it. I doubt they will squander that by not adhering to whatever deal they negotiate next.


Following the rules got them bombed, so why would they make that mistake again?


It's a good question, but Iran has a new weapon now: The Strait of Hormuz. Maybe that's enough leverage that they retain that they will stop their nuclear program for a while.


The Straight of Hormuz isn't new, though. IIUC Iran did exactly what war gamers would have expected them to do.


It seems to me that demonstrated leverage actually is different from theoretical leverage. I do think that is a crucial difference.


The readme says it's 88,000 lines of "hand-written" C, and yet there's only 41 commits all from the last day, and they're all co-authored by Claude.

I have no problem with AI code, but it should not be advertised as hand-written.


I think Claude wrote that part, referring to itself having hand-written them. (It seems to like that phrase a lot.)


Machine generated lying.


“Machine-generated lying”, or “the machine generated those lies”.


I saw these referenced in the repo. Maybe the hand written parts are here?

https://github.com/nordstjernen-web/lexbor

https://github.com/nordstjernen-web/quickjs


No, as those are just forks of already existing projects and the forks don't contain any additional commits compared to upstream.


Consider applying for YC's Fall 2026 batch! Applications are open till July 27.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: