I wonder what are the economics driving these pricing decisions? Are the Chinese companies just subsidizing their models to a greater degree than the US, or is this an emergent property of energy policy between countries?
Throwing out another factor: Chinese companies have been banned and/or limited from buying nvidia, and turned to local companies for their hardware. I haven't actually seen pricing/benchmarks comparing Chinese AI accelerators, but it wouldn't surprise me if that also worked out in their favor as well.
Their models are much smaller: 1T vs 5T for the frontier models. 1T is Sonnet/Google Flash size, not Opus size.
The $0.87/M tokens price for Mimo Pro is probably subsidized.
Mimo models aren't widely available on western providers, but Kimi and Deepseek are similar sizes and cost about the same to run. They are priced $3-$4/M tokens (which is right were Google's very confused range of Flash models are priced at: between $0.40/M tokens and $9/M tokens depending on exactly which model - and you don't want the $9 one!).
Anthropic overprices Sonnet (probably because of their capacity issues). GPT 5.4 mini is $4.50/M tokens.
For one, they invested in infrastructure. They can build fast and efficiently. They can provide power, they can provide cooling. Even if you just make roads better you make everything more efficient. Plus level of standard education. It all compounds.
On HN China is seen as a cheap labor copycat. This used to be a fair approximation at some point in the past. In my opinion China is getting ahead of everyone else much more than US used to be.
SF is a beautiful thing in the US, vast power and wealth comes from there. Smart people collaborating communicating and building fast and with excitement. China did SF kind of thing for many different sectors in many different places.
The Chinese economics: possibly the USA's experience.
It was pretty clear the USA won World War 2 because it out produced and out innovated everyone else. Probably with that in mind, after World War 2 the USA adopted the "Vannevar Bush" model, summarised in this picture: https://www.researchgate.net/figure/annevar-Bushs-Science-th... The idea is to jump start R&D through public funding. The hoped for outcome was that R&D feed private enterprise, leading to a productivity boom.
The boom happened, and the USA did seem to out-compete everybody else in R&D, science, and the products they delivered for decades after that.
That way of doing things seems to have faded over time in the USA. The decline seemed to coincide with the rise of Neo-econmics, and now of course it's been obliterated by Trump. He's very keen to fund Intel to produce chips in a year or two's time (which is something the stock market and banks do perfectly well), but funding basic science is getting drastic cuts.
Still other countries noticed the rise of the USA, and some adopted similar funding models for basic R&D. China seems to have picked it up with gusto, both subsidising R&D and STEM training, leading to huge numbers of engineers and scientists. Whether it will lead to an economic boom remains unknown, but acceleration of ideas and innovations coming out of China seems undeniable. More recently, Ukraine showered its local engineering garages with funds in the hopes of getting a similar outcome to the USA in WW2. It looks like it worked. If the Iran war continues, it's entirely possible arms trade will reverse: the USA could well start buying drones off Ukraine.
Lower cost of labor, lots of under the hood optimizations (e.g. cache hits for DS), many of these companies have existing infra (fewer upfront costs for deployment), etc
China isn't that cheap for labor. And if you think the guys in Z.ai or xiaoxiao aren't the exact same guys from Tsinghua, Peking, MIT, Stanford, CMU, etc. and pulling in amazing salaries you'd be wrong.
I'm pretty sure Xi is also a sociopath, but he differs from Trump in that he's competent. And maybe that's a good thing for American democracy--if we had a competent dictator who could manifest massive infrastructure projects maybe the pro-democracy backlash would be significantly attenuated?
Python never met a footgun it didn’t need to adopt. In this case, however, it’s not equality checks, but operator overloading. I was a Python developer for a decade before switching to Go and life on this side is so much better.
Operator overloading has never been an issue for me, but terminating a line with a comma creating a tuple, or white space (including new lines) between strings to concatenate have cost me days of work over the years.
I understand why those exist, but they’re pure evil.
IIRC, SQLAlchemy overloads this to return an object that represents an equality check in SQL. Because it was returning an object, it was always evaluating to True, because of another of Python’s footguns: truthiness/falsiness. This was a decade ago, and these particular footguns were not even remotely the biggest culprits in our bug backlogs (another honorable mention includes accidentally calling a sync function in an async context, causing timeouts in unrelated endpoints and leading to cascading system failure).
Some countervailing forces off the top of my head:
* Hardware improvements will reduce costs
* Model training improvements (read: more efficient model training) will reduce costs
* Better models will reduce costs (more inference for less hardware time while keeping quality constant)
* Tooling and platform will stabilize—less need to dump money into applications and backend systems because they will become mature—also improvements in AI efficiency and quality will lower the cost of maintenance and future feature development
* Energy buildout will stabilize (we will eventually have enough energy supply to meet AI demand)
* Chips market will stabilize (chip supply will catch up to AI demand, lowering the hardware costs)
What's your time estimation for the last 2 points? Last I heard TSMC is not willing to commit dozens of billions to build new fabs for what might be a fad. Granted, they're not theonly foundry, but that's a signal nonetheless. Given the current craziness around hardware, I doubt the stabilization will come soon. Probably not before token costs soar.
I'm far from an expert, but I'm pretty sure the chips bottleneck at the moment is memory chips which has little to do with TSMC and the memory industry is _dumping_ capital into increasing manufacturing capacity. I'm not sure when energy prices will stabilize.
If a particle was dropped into the sun’s gravity (not with “horizontal” motion that might cause it to orbit), is it time dilation that causes it to accelerate toward the sun somehow?
My guess: In the reference frame where the particle is not moving, the sun would be either a) moving (with a perpendicular component) and be ever so slightly moving toward the particle or b) not moving but a third body would be, moving both the sun and the particle at different “strengths” (different mass and distance, different time dilation) thus the particle and the sun would appear to move closer to each other. That means in either case (sun in the middle or particle in the middle) the third body moving closer must make it look like the particle and sun are gravitationally pulling each other. If we then shift reference frame back to the third body being stationary and the particle and sun moving, we should see that. It would be really cool if we could simulate this to test it but I believe that would require solving the 3 body problem.
> Most people would probably say that’s a good thing, if I read the tea leaves correctly.
I'm very pessimistic that this is about AI safety. I think it's probably more about giving the Trump administration leverage over AI companies. It will be able to coerce them into e.g. propagandizing or surveilling or similar or else they will risk the same kind of "regulatory oversight" that caused television networks to fire comedians who made jokes the regime didn't like.
I can almost excuse Apple for not being concerned about the relatively niche “mac as a server” use case. The thing that boggles my mind is how their keyboard and autocorrect experience get steadily worse with each release. This is the primary way to interact with their flagship device—the thing that generates an enormous share of their revenue. Why go out of your way to make that worse?
100% I’ve used iPhone voice dictation for years. My voice has r changed, but its speech to text makes me sound like a stroke victim. AutoCorrect is not quite as bad, but it’s definitely regressed over the years.
I would happily jump ship for any competitor that offers solid AI inference benchmarks at a competitive power efficiency, but as far as I can tell Apple owns that market by a pretty big margin. I’m sure someone will point out if I’m wrong.
Owning a human is much more problematic, and anyway a human’s peak efficiency might be higher, but humans have to sleep and take breaks so overall probably not.
Humans also tend to hallucinate a lot, but it’s not polite to use that word for them. With humans we say they were wrong, because offending a human reduces performance and you can’t reset their contexts.
I just have a desktop at home that I run inference off of. It is a great setup and I don't find myself wanting to inference models directly on my laptop.
That’s what I would do too, but I haven’t found a desktop build that can rival a Mac Mini or Mac Studio on performance per watt. I haven’t looked super hard, but it seems like Mac is in a different ballpark.
I mean, once it's a desktop, watts are pretty cheap, so it's a bit strange to optimize on that factor for the desktop form factor. For laptops it makes a ton of sense.
If I was optimizing solely for cost I wouldn't use a desktop in the first place. Watts are cheap, but they result in pollution and noise (e.g., fan) or you need some costly alternative. Also I just value efficiency.
Illinois online forms are egregiously bad. A couple examples off the top of my head:
I was traveling through the state and got on a toll road by accident without cash to pay the tolls, the toll booth employee said I could just pay online, pointing to a sign that said "missed a toll? pay online at <url>", so we continued our trip intending to pay the tolls online. At the end of the trip, I logged on to pay our tolls--the application insisted that we enter the specific toll IDs of each toll that we missed, even though we didn't know we were supposed to be recording any toll IDs. If you didn't know the toll IDs, you could use a map application to look up the toll IDs, but the map application would crash within a second of opening it and even if you managed to get a screenshot the resolution was so low that the text was indecipherable. When I called the support number, they told me that I would be fined triple the cost of the tolls if I didn't pay within a week, but I would be able to pay without knowing the specific toll IDs that I missed. When I asked the agent to tell me what tolls I missed (clearly they knew if they were going to fine me), they told me they couldn't tell me for another week. I pointed out that this would be after the toll deadline and they relentlessly tried to avoid acknowledging that simple fact. Eventually I sent them my best guess about what I owed with a letter stating what I had attempted and that I would contact a lawyer about any fines and never heard from them again.
Some years later, after moving to Chicago, I had to file state taxes for the first time. The state issued me a driver's license with a 12 digit number, but the state tax form only allowed me to authenticate with an 8 digit Illinois driver's license number (the other acceptable forms of identification didn't apply to me for reasons I no longer remember).
Illinois is one of the most administratively fucked up and corrupt states in the country. Four of the last ten governors have been so bad they actually served prison time and the current governor appears to be cut from the same cloth.
Illinois has a history of corruption for sure, but illinoispolicy.org is right-wing propaganda. Also, it’s corrupt and bureaucratic for a purple state but it’s still much better than the median red state (my native Iowa went from purple to solid red a decade ago, and we nosedived on everything from education to economy to deficit spending to public health to corruption).
I feel like half of these are quickly resolved in court when you explain what happened and the judge calls the state a bunch of fucking morons and dismisses the case. But the other half of the time the judge won't use common sense. Also the state is relying on you not wanting to show up in front of a judge out of your own state to contest what I presume is a relatively small amount of dollars.
reply