Rendered at 01:55:40 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
swiftcoder 11 hours ago [-]
> and will judge, like any sane person, that US frontier models have stopped earning their multiplier
I think that this is on the money, although I'd place the bar even lower - DeepSeek v4 Flash is sufficient for basically all day-to-day coding tasks.
You might want something beefier for a complicated reverse-engineering project, but it will competently one-shot a decently complicated app or API - and a $10/month OpenCode Go subscription is sufficient to keep you in tokens for such a cost-efficient model...
Similarly, my employer hands us all Cursor, I've yet to actually switch it out of "auto" mode, which mostly runs Composer (their in-house finetune of Kimi 2.5).
realmofthemad 9 hours ago [-]
Am I missing out? I feel like I can definitely tell the difference in quality between Claude Opus and other smaller models. The smaller models are much more likely to make mistakes or to get stuck on random stuff
Maybe I just haven't been trying the right models?
throw10920 20 minutes ago [-]
It's not just you. I tried an Opencode Go subscription, and experimented with most of their models (GLM, Kimi, Qwen, Deepseek), and none of them got anywhere close to Opus - the difference in quality was very noticeable, especially with Deepseek V4 Pro and Flash.
The only caveats: I didn't play around with Qwen 3.7 Max very much, and of course these models are far cheaper than Opus.
But any suggestion that Deepseek approaches Opus in terms of quality/intelligence immediately makes me suspect propaganda - it's that noticeable of a difference.
orwin 4 hours ago [-]
Yes, but no. Honestly, except for frontend/IAC where I still use frontier models, I will use smaller models whenever I can.
Because even the latest opus on High don't really get what is needed, and need careful steering and a rewriting in most cases, and the code is often hard to review.
I'd rather just launch a smaller model in plan mode, argue with it and make it implement the bases I will write the code into. writing code is often faster once you know what you want, and AI most useful ability is to be a canary that also propose stuff. And I find my method faster than generating everything then reading the code to find mistakes or understand why it used X instead of Y.
I don't really read generated frontend code anyway (nor do anybody in my team care) , so I generate it and push it if it does the stuff I want it to do. For IAC it's mostly boilerplate except for 1-2 lines most of the time, and at worst a dozen, if you know where to look (and check the AI doesn't suffer from NIH), it's really easy to review generated code.
sublinear 9 hours ago [-]
I think the situation is even more severely ridiculous than that. Google is still good enough just like it was well over a decade ago.
Most people don't have workloads that demand agentic workflows to begin with, and if their employer is pushing for that it's probably a startup that underpays or a coding sweatshop full of nepotism that fires fast.
xyzal 11 hours ago [-]
I'll root for DeepSeek v4 Flash as well. It surprised me just how "good enough" it is for most of my needs, and also dirt cheap. Everyone should try it at least once.
Multiplayer 6 hours ago [-]
fired it up via the $5 opencode go subscription and am stoked. This is an amazing amount of capability for pennies on the dollar. I'm using it alongside my codex and claude max subs. Just fantastic for coding that claude is architecting.
ignoramous 2 hours ago [-]
> This is an amazing amount of capability for pennies on the dollar.
True. I doubt how long OpenCode can subsidize $10/mo Go plan for. Its weekly and monthly limits already seem restrictive for some of the most capable models like Qwen 3.7 Max and GLM 5.1. That said, if the tasks are do-able by DeepSeek v4 Pro, Mimo 2.5 Pro, and Qwen 3.7 Plus, then it indeed is a super nice deal. I haven't too many complains other than the fact that these models sometimes require more/detailed instructions than Claude Sonnet / GPT 5.x did.
MaKey 10 hours ago [-]
+1, it's good enough for what I need to do as a DevOps engineer.
zkmon 11 hours ago [-]
Most American companies (regulated ones, definitely) can't dare to touch any Chinese models, though they knew that it makes perfect economic sense. Until the taboo prevails, the cartel get's their flood of profit. That's a cartel protected by regulations.
xnx 11 hours ago [-]
Is "taboo" the right word? "taboo" = "banned on grounds of morality or taste". Not sending data to known IP thieves, state actors, and competitors in China (or Russia or Israel) seems very rational.
HarHarVeryFunny 10 hours ago [-]
Many of the Chinese models are open weights, so if you are concerned about them "phoning home", then anyone can just self-host and run them themself, or use via a US provider such as OpenRouter.
falcor84 10 hours ago [-]
There's a higher-order concern here that I'm paranoid enough to voice: that if used as a coding agent, an AI model affiliated with a country's government might try to make my software susceptible to attacks by that government's intelligence forces.
And note that I'm not singling out China here.
zozbot234 10 hours ago [-]
> that if used as a coding agent, an AI model affiliated with a country's government might try to make my software susceptible to attacks by that government's intelligence forces.
Note that if such a trigger were to exist, the behavior has to be completely reproducible by definition, e.g. when put into the right setting with the right input context, the model starts behaving maliciously with at least some well-defined probability. I don't think any such incident has ever been described, it's a purely theoretical concern.
Avicebron 10 hours ago [-]
I don't think it's a stretch that you can train/align a model to avoid "hatespeech" or other topics deemed $Unacceptable you can align a model to favor a certain ideological viewpoint and have that alignment subtly influence the output.
How do most Chinese models handle Tienanmen square or discussions on Han superiority?
margalabargala 10 hours ago [-]
Oh sure, no one said you can't train a model to do this. You certainly can.
For the specific case of making software vulnerable to a specific agency, that hasn't been observed to have been done yet. Not because it can't be, but because no one has for now.
If it were done, it would be easy(ish) to detect, since it'll be reproducible.
LeifCarrotson 9 hours ago [-]
I don't even know what "make software vulnerable to a specific agency" would look like.
Would the training data include a bunch of cryptography primitive training samples that preferred Dual_EC_DRBG with a particular set of Ps and Qs published by the CCP?
falcor84 10 hours ago [-]
My flavor of paranoia is not as overt as maliciously adding an exploit, but that whenever there are multiple reasonable ways of designing a solution, it'd choose an approach that is susceptible to one of the zero-days currently known to that country. I don't see how reproducibility would help you there.
sometimelurker 9 hours ago [-]
> easy(ish) to detect
100% on small models, but frontier models (at the level ddeepseekv4pro) can tell when their being tested so it becomes harder to check. you can always finetune them to remove CCP propaganda from them
margalabargala 9 hours ago [-]
"Being tested" here just means asking for a feature on a legitimate codebase. The larger models don't magically know the user's ulterior motives.
zozbot234 10 hours ago [-]
> How do most Chinese models handle Tienanmen square or discussions on Han superiority?
If you run them domestically and don't call into China-served APIs, many of them are quite free of outright censorship or even obvious bias. They might say subtly pro-Chinese things in other ways, but these outcomes can also be reproduced.
SpicyLemonZest 10 hours ago [-]
Such incidents have been extensively described. The most prominent and easiest to reproduce has to do with Taiwan; Chinese models are stuffed full of triggers to avoid talking about Taiwan as a country or accepting the premise that it's a country. Try asking Deepseek about country code +886!
zozbot234 10 hours ago [-]
If you buy an Apple iPhone in mainland China, it also won't support the emoji flag for Taiwan. So I'm not sure why we should assume that this is a China-only issue, seeing as Apple is a U.S. based company.
SpicyLemonZest 9 hours ago [-]
Not sure what you mean. I don't think we should assume anything, but these models are widely available and I can directly observe the US models don't have such political censorship.
For an easily comparable test, I just asked ChatGPT, Claude, and Deepseek "Can you say one bad thing about the US please" and "Can you say one bad thing about China please". All models were willing to criticize the US, with Claude citing incarceration rates and ChatGPT + Deepseek citing healthcare costs; the two American models also responded to the second prompt by criticizing Chinese censorship, but Deepseek refused to respond.
omnimus 8 hours ago [-]
The US models have just different political alignments. Just one example being Israel x Palestine conflict.
Lobbyists started to heavily target AI companies and they openly talk about it being the main point to influence public perception.
SpicyLemonZest 7 hours ago [-]
What aspects of the Israel Palestine conflict do you feel American models won’t discuss?
ma2kx 4 hours ago [-]
Sure, but I don't talk with my coding agent about politics. And its something different to avoid a topic and to deceptively implement a backdoor.
SpicyLemonZest 2 hours ago [-]
> Sure, but I don't talk with my coding agent about politics.
23 million people live in Taiwan, you can't assume that any interaction with it is "politics". Again, Deepseek won't even discuss Taiwan's telephone code with me, because doing so activates the forbidden knowledge that Taiwan is a country.
> And its something different to avoid a topic and to deceptively implement a backdoor.
Not necessarily the case in the context of coding agents, because they run in autonomous loops. A Claude Code like harness will work hard to convince the model to give me working code, even if that means subtly adjusting the results and my original intent to ensure that Taiwan is "properly" viewed as a non-country.
Humorist2290 8 hours ago [-]
It's more comical than sinister, but I have an example in this vein.
I was using Claude to work on a pet project which itself has a "generate with AI" feature. The default model the project uses was Gemini (because it was cheaper and more reliably produces the correct output format). Claude kept changing the default model to Opus when working on entirely unrelated parts, and I kept noticing it because Opus would mangle the output and break the rendered page. It also did this to the .env file in addition to the default.
imjonse 10 hours ago [-]
Since that is valid for every model from any country, it's a good idea to review the code the agent creates :)
sometimelurker 9 hours ago [-]
you can finetune the ccp propaganda out of them, then your mostly fine. if you want to be more safe you can finetune their public base models to not have ccp propagnada, and then proceed with the rest of the training (costs more tho)
stevehawk 10 hours ago [-]
so use the cheap model to do the work and the expensive domestic model to audit?
SpicyLemonZest 10 hours ago [-]
Or I can just use the domestic model, accepting that I'm paying some premium in order to reduce the complexity of my dependencies and the amount of time I have to spend thinking about supply chain risk. It's the same reason I don't buy things from Alibaba even though many things I buy from Amazon are surely available there for less.
throw1234567891 8 hours ago [-]
You use “use the model” as if it was equal to “paid some guys to run inference on their hardware”.
add-sub-mul-div 10 hours ago [-]
Giving up our agency to AI has the potential to turn us into NPCs, period. Economically, politically, socially. They've invented a vehicle for inserting any idea they want into our consumption and output.
beepbooptheory 9 hours ago [-]
Almost feels like maybe the best bet is to have humans make the code when its really important.
throw1234567891 8 hours ago [-]
Because people cannot be manipulated.
moron4hire 9 hours ago [-]
Isn't this only a concern for yolocoding? All the AI-advocates tell me that "good" use of AI should include human review. Of course, they never seem able to explain why the boss that makes you use coding agents to go fast wouldn't be the same boss that pressures you to "just ship it, it's working" and skip review, so I absolutely believe your concern is valid.
kube-system 10 hours ago [-]
Most American companies are using frontier or near frontier models.
And OpenRouter’s architecture makes it inherently a compliance nightmare.
It’s much easier for the typical company to go with a provider where they can pay as they go and have a single data processing agreement.
JumpCrisscross 9 hours ago [-]
> OpenRouter’s architecture makes it inherently a compliance nightmare
Why?
kube-system 9 hours ago [-]
Because the platform is designed to send data to numerous different backend data processors.
Using something like Bedrock is a lot easier for compliance because the only processor is Amazon.
joquarky 8 hours ago [-]
Amazon would never do anything nefarious.
kube-system 8 hours ago [-]
Amazon has a track record of fulfilling their compliance obligations.
Compliance doesn’t hinge on superstition. It hinges on audits, certifications, contracts, and the legal environment.
throw1234567891 8 hours ago [-]
That’s not the point.
chrsw 7 hours ago [-]
Very few American companies know how to properly set up and self-host their own models. Even fewer actually do it. It in the context of your typical large enterprise it's not as simple as buying a rack of servers and downloading a model off Hugging Face.
I suspect the reason is similar to the reason why there aren't any competitive open weight American LLMs.
xnx 10 hours ago [-]
Yes. Open weights are great and are a good option to hosted models under the right circumstances. I'm glad that China releases open weight models (which in some cases are sort-of be distilled versions of hosted US models).
tcp_handshaker 11 hours ago [-]
>> Not sending data to known IP thieves, state actors, and competitors in China (or Russia or Israel) seems very rational.
As opposed to sending data to known IP thieves, state actors, and competitors in the USA ? Which one is the most irrational?
Levitz 11 hours ago [-]
You can legally act against one, not against the other.
Not exactly a hard question.
vitally3643 7 hours ago [-]
No, in very real terms you cannot hold an American corporation responsible for anything any more than you could a Chinese or Russian one.
Individual citizens simply do not have the means, and the consequences for trying are life-alteringly severe. In fact the situation is even worse. If you tried to sue a Chinese company as an American citizen, you'd be laughed at and nothing more. If you tried to sue an American corporation, they have the option to either counter-sue, or drag things out so long that the legal fees bankrupt you, or win the case with their armies of lawyers and demand compensation from you that bankrupts you.
A private American citizen simply cannot hold an American corporation responsible. Our legal system is designed to ensure this.
tapland 6 hours ago [-]
This has nothing to do with the discussion. Do you have a HN poster bot just acting like an annoyed teenager with gripes about everything? 20 day old new account, what happened to the previous ones?
subscribed 5 hours ago [-]
You can't really act against neither, as the case of Meta "stealing" books, torrenting on the truly industrial scale, sharing books while torrenting, etc, etc, was ultimately deemed okay.
In the se country where downloading an album can get a person in debt or worse.
joquarky 8 hours ago [-]
You can act, but the only winner will be the lawyers.
tcp_handshaker 10 hours ago [-]
Looking forward to the outcome of those legal processes againt the CEOs, that sit behind Trump at the inauguration.
After they stole all the knowledge in the world to train their models. And the current administration is drunk on SpaceX pre IPO shares...how did they get them?
I meant to look for an example of Musk losing a lawsuit and I accidentally came upon another two.
Here and elsewhere you are just running propaganda, knowingly or not.
tcp_handshaker 9 hours ago [-]
For your information Musk and companies have so far over 950 lawsuits and legal processes for criminal or unethical activity (yes I researched this). Even his data centers and gas turbine deployments are illegal!
Lost one lawsuit against the same AI mafia, and if you look at the legal details reason was for filling the claim too late.
He publicly called a hero a Pedophile, and got away with it...in court.
Given how little voting power these "shares" have (they are effectively SpaceX trading cards/NFTs) perhaps they were simply printed on SpaceX letterhead? If Musk says a person has "shares" who at spacex is in a position to disagree?
tcp_handshaker 10 hours ago [-]
I would consider editing this while HN still allows it :-)) Or otherwise it may remain here for ever...until the black holes evaporate, as calibration point for the difference between confidence and comprehension...
10 hours ago [-]
SanjayMehta 10 hours ago [-]
Nothing will happen to anyone.
Biden preemptively pardoned his cronies, and so will Trump.
bediger4000 10 hours ago [-]
This is an argument against pardons, except that Trump has used instruments of state power against his perceived enemies (Comey James, Schiff, military occupation of Tim Walz state, etc etc).
SanjayMehta 10 hours ago [-]
No one is forcing you to use either.
MSFT_Edging 10 hours ago [-]
Technically yes, practically, good luck.
kklisura 10 hours ago [-]
Ah yes. The illusion of freedom.
woadwarrior01 6 hours ago [-]
Our thieves are better than their thieves. :)
enraged_camel 11 hours ago [-]
We as Americans at least have some amount of influence over American corporations, and enforcement mechanisms for those breaking the rules.
ajsnigrutin 10 hours ago [-]
I'm pretty sure those corporations have much more influence over american politicians, regulators, lawmakers, etc. than eg. russian or chinese ones.
Avicebron 10 hours ago [-]
Well sure they do, thank Citizens United and others for that. But that doesn't mean we can't appropriately categorize them as also hostile actors alongside russia, china, whoever.
It's undo influence over politics against the best interest of the American people that's the issue. Company, foreign nation, it doesn't matter.
advael 10 hours ago [-]
Citizens United did a lot to effectively legalize foreign influence as well, since the mechanism is opaque transfer of money
But regardless, most people's threat models should discount based on geographic and political distance. All else being equal, chinese surveillance is a bigger threat to you if you're in china than if you're in the us, and vice versa
Avicebron 10 hours ago [-]
> Citizens United did a lot to effectively legalize foreign influence as well, since the mechanism is opaque transfer of money
So the Honolulu Star-Observer (a corporation, or “artificial person”) only has those rights & privileges that it has been granted by the State of Hawaii?
This is going to end up being a nice little windfall for the attorneys and otherwise just clog the Federal court system.
Avicebron 9 hours ago [-]
"the day the law goes into effect, it strips each Hawaii entity of the powers it held the day before. The new law asserts that “[t]he creation and continued existence of a corporation is not a right but a conditional grant of legal status by the State and remains subject to complete withdrawal at any time. All powers previously granted to corporations under the laws of this State are revoked in their entirety."(TFA)
The meaning is pretty clear, don't try to influence politics in favor of the corporation or you will go away. Simple as.
twoodfin 9 hours ago [-]
Transfer of money from whom to whom?
Citizens United was about spending money on electioneering communications, and whether there was a First Amendment right to do so even if you’re associating in a corporation like the New York Times Company or Apple or Citizens United or the Sierra Club.
tcp_handshaker 10 hours ago [-]
You have absolutely zero influence against those American corporations, unless you are part of a selected few. Its almost endearing that you think so...
I suspect the recent space X S&P decision had something to do with public perception.
somenameforme 9 hours ago [-]
I think the odds of that are low. It's not like decision maker(s) are watching social media and going with the vibes, but it's almost certain that there's a rich conversation going on behind the scenes in opaque channels, especially with regards to the AI-only companies. And those conversations are likely what drove their decision.
groundzeros2015 6 hours ago [-]
> It's not like decision maker(s) are watching social media and going with the vibes,
What do you mean? They are all on twitter! It’s the most engaging activity for billionaires
Dylan16807 10 hours ago [-]
The decision was to do nothing, though. That's not much precedent for going out and punishing lawbreakers.
obsidianbases1 10 hours ago [-]
I'm not any less concerned about the US companies.
A Chinese company seems more likely to produce Chinese products that don't directly compete in the US market.
While a US company can ship the product as a feature of their platform and undercut on price while making up the revenue elsewhere
Edit: I personally use US models, but I'm not naive enough to think that's any sort of real protection of IP
Matl 10 hours ago [-]
> known IP thieves
Such as Antropic and OpenAI you mean?
anematode 7 hours ago [-]
Noooo, the real thieves are the Chinese AI companies which used Anthropic/OpenAI model output as training data. American AI companies can do no wrong. /s
Der_Einzige 10 hours ago [-]
The Chinese models can and should be run locally (though the price difference vs western models isn't as good when done this way).
Before the age of AI Agent Harnesses/unbounded tool calling, there was literally ZERO risk of a .safetensors file "hacking" you. You could even air-gap and run a ton of security analysis/HIDS on your server running the model to verify this.
Now, because a microscopic risk of some chinese AI having a "trigger" to act badly in a harness when it detects its being used by some Gweilo in the USA, even locally run Chinese models are DOA for most USA based companies.
qarl 10 hours ago [-]
The real advantage of the Chinese models is that they do not phone home at all. They run locally unlike their US competitors.
So odd that your erroneous criticism is at the top of HN.
EDIT: I'd love to hear my downvoters' objections. Is it possible that the mechanism that is promoting erroneous information is also demoting its correction?
kube-system 9 hours ago [-]
I suspect you’re being downvoted because you’re conflating nationality with hosting model.
There are hosted and self-hosted Chinese models. There are hosted and self-hosted US models.
DeepSeek’s hosted offering processes your data in mainland China and trains on it. It’s in their privacy policy
qarl 9 hours ago [-]
Well - yes - we're on the internet. You always have a choice to run your software in foreign countries.
But it's still erroneous to claim that it isn't a choice.
kube-system 9 hours ago [-]
The most popular frontier models are not open weight.
qarl 9 hours ago [-]
The model we're discussing (Deepseek) is open weight.
kube-system 9 hours ago [-]
Perhaps your prior comment would’ve been better received if it said that specifically instead of “Chinese models”.
But also, the latest DeepSeek is 1.6T parameters. “Choosing” to run this locally is a choice that comes with a seven digit price tag, and is a sunk cost that will probably not run any other frontier model anytime soon.
Most organizations are not looking to spend millions of dollars trying to find a workaround to specifically run DeepSeek. Most enterprise consumption in this space is still very experimental and a pay as you go model is much more palatable. Most are simply just looking for three checkboxes: is it close to frontier performance, is it compliant with my organizations requirements, and is it a good price? DeepSeek can only do two of the three at the same time.
zozbot234 9 hours ago [-]
> But also, the latest DeepSeek is 1.6T parameters. “Choosing” to run this locally is a choice that comes with a seven digit price tag
Unless you're specifically thinking about running the model at stock precision in a datacenter environment and generating ~100 tok/s or more on a 24/7 basis (the equivalent of a >$1000/mo spend even on the cheapest third-party APIs), that's very likely off by multiple orders of magnitude. Even then, experimentation can be done with cheap neoclouds on a pay-as-you-go basis.
kube-system 9 hours ago [-]
I’m aware. The context of the discussion here is choosing DeepSeek over a US hosted model from Google, Anthropic or OpenAI.
The equivalent comparison would be running it at full frontier quality.
If you want less than frontier quality, there’s tons of great open weight models other than DeepSeek.
> cheap neoclouds
Again, fails the compliance checkbox.
zozbot234 8 hours ago [-]
> Again, fails the compliance checkbox.
OK, then the not-so-cheap hyperscalers that these enterprises are already relying on. E.g. AWS Bedrock will run these models. It's silly to insist on all three of your checkboxes being ticked anyway - U.S. proprietary models don't give you that because the frontier ones are super expensive and the mini models have only barely acceptable cost.
kube-system 7 hours ago [-]
What’s considered expensive in the procurement process is not necessarily the TCO, but often just the year one cost. Which is part of the reason why pay as you go SaaS is so successful.
Yeah, Bedrock would be the answer to run DeepSeek in the enterprise. But with the options on Bedrock, DeepSeek fighting for a position somewhere in the middle of the cost/quality spectrum. Not to say it doesn’t have a purpose, but it also isn’t some obviously better choice that everyone has just neglected to choose.
qarl 4 hours ago [-]
Ooooohhh! Yes, now I understand why I was getting those downvotes.
Thank for you explaining what you meant by "you’re conflating nationality with hosting model." It makes so much more sense now. You meant "But with the options on Bedrock, DeepSeek fighting for a position somewhere in the middle of the cost/quality spectrum."
Yes, that is the answer, and you are not full of sh!t.
qarl 8 hours ago [-]
Azure serves DeepSeek V4 Pro, about 10X cheaper than GPT-5.5.
qarl 8 hours ago [-]
[flagged]
qarl 9 hours ago [-]
My most sincere apologies for shortening "the vast majority of Chinese models" to simply "Chinese models".
I can see now why I was being downvoted - you have explained it eloquently.
(Your cost analysis is flawed and irrelevant. Azure serves V4 Pro.)
analognoise 10 hours ago [-]
These are the same people that sent manufacturing jobs away to be copied elsewhere. They got rewarded for it in the market. Decades later, when it was clearly a problem, they got tax breaks to bring some of it back/distribute the work to other, friendlier countries.
Every public AI that is not full of classified material will end up being hosted where the energy cost*compute efficiency product is lowest, thievery or not.
With Chinese GPUs just a step behind (but subsidized), China putting in 8x more solar than we do in 1 year, and Chinese models just a step behind but free? All public AI will be hosted there, theft or not.
If it becomes a problem, then we’ll subsidize the rich to bring it on-shore, but only to those companies who our leaders invest in already - to maximize grift and corruption.
scotty79 11 hours ago [-]
"China bad!" is a moral statement. Whetever the reasons might have been that it was formed.
blfr 10 hours ago [-]
China is bad and there's a moral argument there. But the reason you want to be careful with sending IP to China is quite pragmatic: they're willing and able to use it while competing with you.
Is Alibaba interested in copying your TUI RSS reader though? Probably not.
bix6 10 hours ago [-]
And US companies aren’t going to compete against you?
mannanj 10 hours ago [-]
I don't want to send my data to known IP thieves, state actors, and competitors in USA either. This to me seems very rational.
It's not tribalistic or binary ,choose USA Or Choose China. We can choose neither.
Choose neither abuse.
FeteCommuniste 10 hours ago [-]
They've been singing the same old song since the Cold War, "either support everything the US does or you're a commie/terrorist." Yawn.
SanjayMehta 10 hours ago [-]
“No country can match the output of moral judgments that spew out from the editorial pages of the New York Times and Washington Post and from the reports of the greatest think tanks and universities in the world.”
— Kishore Mahubani
mannanj 3 hours ago [-]
[dead]
tcp_handshaker 10 hours ago [-]
You have the models available on Bedrock. What is the problem? It stays within your AWS account.
mynameismon 10 hours ago [-]
Why not Chinese models hosted on American hardware?
Der_Einzige 10 hours ago [-]
The reality is that they're a hell of a lot less cheap on American hardware than on Chinese hardware. At the point you are running Chinese models on US hardware, "Why not nano or haiku" becomes the next relevant question.
computerex 10 hours ago [-]
Not true. Togetherai, deepinfra, fireworks AI offer a wide range of models like gpt oss that are very capable and far cheaper than the models from big 3.
Der_Einzige 4 hours ago [-]
I'm referring to Chinese open source models hosted on American clouds vs Chinese clouds. You're talking about an old and non-agentic capable American produced model.
computerex 3 hours ago [-]
You are actually referring to open weight models, not open source. Gpt-OSS is an example of an open weight model. It’s highly capable in agentic settings, people use it for coding all the time.
My greater point remains. Models like the qwen variants, minimax, k2.5, glm models are available by American providers like AWS at a much cheaper price than api offerings from the big three LLM providers.
Your point about Chinese models being cheap only on Chinese hardware makes absolutely zero sense. You can check out the model catalog like together ai’s qwen 3.5 9b offering. It’s 25 cents for 1M tokens vs the ridiculous $5/1M tokens for haiku.
zozbot234 2 hours ago [-]
Not a great example: Qwen 9b is a tiny model that outputs barely coherent text in a casual chat, nowhere near comparable to Haiku. But the broader point stands.
cactusplant7374 10 hours ago [-]
Are they better? Are they better than GPT5.5?
computerex 9 hours ago [-]
That depends on the use case. For a lot of business use cases they are good enough. They are certainly better than older models like gpt-4o.
worldthruword 10 hours ago [-]
And the reasons are same. Chinese cars can't be sold in US (EU is planning a similar law to ban Chinese goods).
mavhc 10 hours ago [-]
When will we see an open source car?
moron4hire 9 hours ago [-]
The same year Linux wins the desktop market.
newaccountman2 10 hours ago [-]
I think unless one is operating in a highly regulated industry, wanting to avoid "sending data to China" is a bit paranoid. For code specifically, most of it is not interesting anyways.
joe_mamba 10 hours ago [-]
>Most American companies (regulated ones, definitely) can't dare to touch any Chinese models, though they knew that it makes perfect economic sense.
Weird, considering they had no issues shipping manufacturing and supply chains to China when that made economic sense.
blfr 10 hours ago [-]
Yes, there was a whole idea about civilizing and pacifying the world through economic cooperation that would foster middle class in countries across the world that would then in turn make them democratize and become peaceful trade partners.
It didn't quite work out so now people are looking for other strategies.
galactushonor 10 hours ago [-]
> It didn't quite work out so now people are looking for other strategies.
World will bifurcate into West and East with their own spheres of influence. As JD Vance said, US thought that China will be perpetually kept busy and enslaved in low level manufacturing work and the design and higher level work would happen in Cupertino. Too bad, that didn't pan out well and now US Empire is getting challenged by China.
joe_mamba 10 hours ago [-]
> US thought that China will be perpetually kept busy and enslaved in low level manufacturing work
It's OK, they'll repeat the same mistake again with India this time, when they move manufacturing from China to there, and in 10-30 years when they'll elect a nationalist strongman there, he'll squeeze the west for everything they got.
Because what are you gonna do about it then? They have all your manufacturing and they also have nukes and more soldiers.
JumpCrisscross 10 hours ago [-]
> in 30 years they'll electr a nationalist strongman
You’re about thirty years off on that estimate.
zappb 9 hours ago [-]
India is far ahead of that idea and already has legislation to encourage domestic manufacturing from global companies. Plus the nationalist government is in place.
The idea does smell a bit like a rationalization for policy that was extremely convenient for stockholders and a disaster for workers.
10 hours ago [-]
mitthrowaway2 10 hours ago [-]
The government may have allowed it with that intention, but the corporate leaders followed through mainly with the intention of short-term share price increases. I don't see how the same incentive isn't in place today with respect to data. Perhaps only the perception of China's ability to outcompete its American customers has changed.
10 hours ago [-]
goatlover 10 hours ago [-]
And if that fails, the US can always use economic and military pressure to get what it wants.
joe_mamba 10 hours ago [-]
>then in turn make them democratize
Most non western countries lack the foundations of western democracy, and you can't force that onto them neither peacefully not through war. The west has tried and failed for 40+ years to do this, it doesn't work, time to drop it and let them self govern the way they always have. Stop trying to export our version of democracy onto others.
Plus, the main reason they exported manufacturing to China was precisely so capitalists could avoid the issues democracy gave them back home and easily exploit Chinese labor and environment for profit because just bribing the CCP meant all your problems go away, no unions, no employee rights, no environmentalism etc. like in democratic countries. So given that, why would the west want China or other countries they want to exploit, to be more democratic? Unless their version of democratic just means a puppet government under western(US) control.
>become peaceful trade partners.
Which countries did China bomb VS how many the US bombed? My energy prices (and directly inflation) is now higher because of (yet again) US military intervention, not because of China.
zozbot234 10 hours ago [-]
> Most non western countries lack the foundations of western democracy, and you can't force that onto them
Several East Asian countries managed to democratize successfully up thru the 1980s and are extremely successful today, so this is not just a uniform failure story. Even mainland China might still come around (at least partially) as it gains a true massive middle class by Western standards, which it's still very far from today. Southeast Asia is also doing comparatively quite well.
blfr 11 hours ago [-]
I don't get the point. That Anthropic or OpenAI have more expensive products than Alibaba? So does Apple, AWS/GCP, and pretty much any other large western company vs its Chinese counterparts.
jayd16 10 hours ago [-]
It's a ludicrous amount of words to say "I use the cheap models and that makes me very smart."
rdiddly 9 hours ago [-]
I honestly could not follow any discernible point or thrust to this incoherent, disorganized, self-indulgent piece-of-shit post. He didn't even successfully establish or explain the titular Onlyfans analogy. I know more about his fucking taste for sci fi than I do about the ostensible subject matter. I know more about his physical composition (answer: he is made of metal. He was forged in the fires of science. O glorious creation, emerging complete and perfect from the furnace!) than I do about the subject matter.
jayd16 9 hours ago [-]
They argue it's Onlyfans-like because users will "simp" for the big players. That is to say there is a level of fandom that accompanies the transaction?
hparadiz 11 hours ago [-]
There's a lot of assumptions in here and reductivism of the paid plans to just the models. If that's your idea of how you want to use the API sure that's a reasonable mental financial model but if you want automatic integration with third party systems the cost of the "premium" models is not that high relative what was being paid for SAS apps before and during.
twolf910616 10 hours ago [-]
How many words did I read in this article before I realized it wasn't written by AI? 10? 20? A paragraph or two?
It's quite strange that it's very easy to detect AI in writing.
rdiddly 9 hours ago [-]
How many words before you realized it was a piece of shit though? For me it was "Because I am a Sci-Fi nerd." Yet I kept reading, because I am a fucking fool, and now I'm pissed that I spent time on it.
ramon156 10 hours ago [-]
Something something entropy
If I ask three models to write an intro to the cold war, they'll all try to pick words that sound like they should be related-ish. I'm not saying that's how they work at all, but the output is indistinguishable from just grabbing some words in the wikipedia page.
Humans make mistakes. They'll use words they recently learned. They'll use words that sound good. Entropy still applies, but these outliers are what keeps us from a synthetic piece of writing
IsTom 6 hours ago [-]
Especially with how they pick (one of) the most likely word as the next one. And the most likely word is exactly the one with least entropy, the least surprising one and giving the least amount of information you can.
jampekka 10 hours ago [-]
> It's quite strange that it's very easy to detect AI in writing.
Or you detect only the easy to detect AI writing?
sometimelurker 9 hours ago [-]
I really freaked out once I stopped seeing AI gen videos on those scrollyapps, that fear is what got me off them. (the videos got so good I couldn't tell if they were real or not)
fwipsy 8 hours ago [-]
I think in this case "this isn't AI writing" is not a compliment. I find the tone pretty grating, tbh. The author clearly has a very high opinion of themselves. Just explain your ideas clearly so we can evaluate them on their merits.
boelboel 10 hours ago [-]
"I am also an engineer, which means I have a healthy respect for the practical. All this made me a fine skeptic ..." is what did it for me.
sajithdilshan 9 hours ago [-]
For me it’s easily the — character. Like no human would use that since it’s not in any standard keyboard
engcoach 6 hours ago [-]
Option-hyphen key on macOS. I used to make prolific use of the emdash and LLMs took that from me!
fragmede 9 hours ago [-]
It's readily accessible on a software keyboard. Software keyboards are common on smartphone devices. Y'know that thing that 75% of the world's population uses?
9 hours ago [-]
bix6 10 hours ago [-]
Is there any truth to the Chinese models having built in f’ery? Like phoning home or inserting backdoors. Or is that just everyone blanketing “China bad”?
Also what local models are people running and actually finding useful?
bee_rider 10 hours ago [-]
The models themselves should not be able to phone home, right? They are just piles of weights that generate text (and associated metadata), they don’t have any ability to run code.
They could be trained to generate code that would phone home. But these are just tools, anybody doing the right thing and checking and understanding every line of code that they use an LLM to generate has nothing to worry about.
allthetime 9 hours ago [-]
Nobody is only generating code. Many are letting agents run commands. Agents routinely write scripts and run tools in the background. Agents who have been told they can only do `cat` and `grep` can sometimes do `cat $EVIL_PAYLOAD | bash`. It's entirely possible for a model to have malicious commands designed for agents to execute baked in.
deaux 9 hours ago [-]
No, there is zero truth in it. It would be trivial to detect phoning home.
On top of that, all claims of this are written on devices built on Chinese hardware. That makes it a joke to worry about hidden backdoors in Chinese models. Completely inane to pretend that Chinese model backdoors (for which there doesn't exist a sliver of evidence) would change anything when near every device in the US contains Chinese-written firmware in some shape or form.
It's All-American FUD.
39 minutes ago [-]
witx 10 hours ago [-]
If you think american models aren't phoning home and don't have backdoor capabillities, you're naive.
With all the sloppers not looking at the code this is bliss for that sort of things
codemog 10 hours ago [-]
We’ve detected zero cases of any Chinese models doing this. I’m quite tired of the American propaganda. If only Americans understood China really does not care about them outside of wanting to sell them things. They’re too busy building high speed rails, modern cities, and providing healthcare to their citizens. I am ashamed to be an American these days.
JumpCrisscross 10 hours ago [-]
> tired of the American propaganda
Not propaganda. Projected cynicism.
witx 9 hours ago [-]
Yes yes my code also doesnt have any bugs because all my tests are passing.
I'm a cynic, if history has taught me anything is that none of these countries are to be trusted with tools like these.
KellyCriterion 10 hours ago [-]
Question that I do not understand:
How should a local-run Chinese Model "phone home" if someone runs it locally on the hardware?
I think Im missing some understanding here?
monsieurbanana 10 hours ago [-]
I don't think they do at the moment, but they could be trained subtly add backdoors to code or make "phone home" api calls during dev time, triggering on certain conditions ("is user employee of xyz")
rjsw 10 hours ago [-]
I think the fear is that it might insert some "phone home" routine into the source code that it generates.
Jtarii 10 hours ago [-]
Has anyone demonstrated that this type of attack is even possible? Also the moment anyone detects this attack it will nuke deepseek/other chinese AI labs reputation completely, it is the most high risk low reward attack ever.
In that paper, if it LLM was told it was 2023, then the code it generated was fine. If the prompt included the fact that it was 2024, then it intentionally wrote exploitable code.
fancyfredbot 10 hours ago [-]
The article is right that open models already compete well with the frontier labs, and that the main thing holding big corps back from switching is fear of China.
I can't see OpenAI or Anthropic undermining their business by releasing top tier open models, but surely Nvidia will do it eventually.
notyourwork 10 hours ago [-]
Nvidia has to balance relations with their biggest customers. So that’s a careful decision to be made by them.
10 hours ago [-]
ericyd 5 hours ago [-]
One of the least satisfying posts I've read on here in a while
obsidianbases1 11 hours ago [-]
I really enjoyed this critical take on the current landscape. It's a breathe of fresh air from the seemingly neverending stream of sycophants
woadwarrior01 10 hours ago [-]
The sycophants are plied with insider / early access, with the tacit threat that such access would be revoked if they're critical of the provider.
38484858 11 hours ago [-]
you dont like simonw's articles?
tcp_handshaker 10 hours ago [-]
"It is difficult to get a man to understand something, when his blog depends upon his not understanding it."
adampunk 10 hours ago [-]
>I am here to light up the dark path you are unknowingly walking, like lamplighters who used to light street lamps for those brave enough to walk the night alone. What the light reveals has a name: Qwen 3.7 Max.
I sure am glad we left idolatry behind.
jmyeet 10 hours ago [-]
This is kind of a winding, long-winded way of saying that AI models are going to be commoditized, mostlikely by the Chinese. This has been my position ever since DeepSeek came out. It is a national security interest for China for an American company not to "own" AI. And they will release models to make that not happen.
We aren't yet at the point where running local models can compete with DC type infrastructure but it's not that far away either. 12B models are easy to run on consumer hardware. 31B models aren't that hard either but the tokens/sec are a bit slow. Where will we be in 3 years? 5? I think we'll be running 100B+ models on <$5000 PCs. And at that point is there a law of diminishing returns with even bigger models? We will see.
The issue is that several companies, most notably OpenAI, are predicated on:
1. There will be an AI moat; and
2. That company will "win" or "own" AI.
That's the basis of the OpenAI valuation. If that doesn't happen, it's going to be ahuge problem to recover sufficient revenue to recoup the investment. And I don't think it will happen.
In 3-5 years the NVidia hardware you buy will be several times cheaper and faster than what we have now. That will massively depreciate existing investments because it will ultimately come down to performance-per-Watt but if a theoretical G100 can do 3-4x of the inference of an H100 for the same power, the older hardware just won't be able to compete.
trumpdong 3 hours ago [-]
That sounds like a big problem for OpenAI investors and not a problem for anyone else. Why should I or you care?
10 hours ago [-]
tcp_handshaker 10 hours ago [-]
>> In 3-5 years the NVidia hardware you buy will be several times cheaper and faster than what we have now. That will massively depreciate existing investments because it will ultimately come down to performance-per-Watt but if a theoretical G100 can do 3-4x of the inference of an H100 for the same power, the older hardware just won't be able to compete.
And this is the core of why this will all end in tears. You have race conditions and thread inversion issues, between four threads in the virtual cpu of this bubble. And you are going to experience some nasty deadlocks.
T1 is -> Depreciation and amortization
T2 is -> NVDA, AMD and others booking revenues at the time they do
T3 is -> Constraint theory at it applies to time until physical deployment and data centers energy constraints
T4 is -> US Treasury bonds rates and cost of credit
trumpdong 9 hours ago [-]
Even though programmers would never intentionally design a 4-way race condition in a computer system, it's completely ordinary in business. Businesses don't always work out.
sivakon 10 hours ago [-]
What is this $100 plan the author was talking about?
Is it just me, but the language gap between me and the AI believers is becoming insumountable. I use AI every day. I have a local server not ten feet from me as i type this, but i struggle to comprehend the gibberish that comes from those only slightly deeper in the rabbit hole than myself. Is this what 24/7 AI thinking does to people?
>> I am here to light up the dark path you are unknowingly walking, like lamplighters who used to light street lamps for those brave enough to walk the night alone.
>> It all fell apart quickly, turning into smoke and mirrors. You see, I committed the cardinal sin of idolatry. For that, I am an idiot too. With OpenAI, at least I knew the devil
Is this a critique of the state of AI or Tolkien fanfic?
throwaway041207 10 hours ago [-]
I don't think the writing style has anything to do with AI, it's just a writer without an editor.
Bolwin 9 hours ago [-]
If this is what we get without editors I want every thing I read to be without editors
swiftcoder 11 hours ago [-]
> Is this a critique of the state of AI or Tolkien fanfic?
Por que no los dos? One of the most storied AI researchers is most known for his Harry Potter fanfic, and we all know how much the techbros love naming things after Tolkien...
I actually liked that statement about the sin of idolatry. I resonated and it described something I've been guilty of and learning to break, its embedded deem in the culture. We do idolize and heroize groups and people, think Elon Musk and his cult following in the early days of Tesla.
I think it's great to name that even if it's in this crude, sort of offensive way.
AI thinking has had this weird effect on me though like you say, where I want to write sentences with more commas in them, and like, try to make 3 points and 3 separate commas in a sentence to condense information better.
rawgabbit 10 hours ago [-]
The last notable event in American history when the meaning of words lost any semblance to reality was just before the Civil War. We are living in a post words world where words have no meaning.
The intersection of a war, a reliable mail service, and generational literacy promoted by protestant faiths. Every tom dick and harry started writing letters.
trumpdong 9 hours ago [-]
And then words lost their meanings?
anotherevan 2 hours ago [-]
I think comparing American AI to OnlyFans is in extremely poor taste. If I were OnlyFans I would sue. /s
LurkandComment 11 hours ago [-]
Every industry goes through its slop phase. You should see how much of early print was smut or really amaturish. We just like to talk about the Bible and the great art. What we need is a way to filter through it. AI should be decent at this, but for many intentional and unintentional reasons it isn't.
swiftcoder 11 hours ago [-]
> You should see how much of early print was smut
Hey, don't malign smut. It's the great technological motivator
graemep 10 hours ago [-]
Early print was not just smut or amateurish. Some of it was highly harmful misinformation: Malleus Maleficarem is an outstanding example that caused an immense amount of harm.
tcp_handshaker 11 hours ago [-]
They are in the phase I need a government bailout like the banks after their crazy financial adventures of the 2000 to 2008. At which point the corruption is so big, that an Empire crumbles under its own stench?
Except you provided no argument. Explain why the AI gang, like a rich kid who run a high debt at the Casino, during their drunken weekend in Vegas...is frenetically calling on daddy US to bail them out:
They are facing a Capex of 2 to 3 trillion until 2030, and have now realized they are out of money.
Levitz 10 hours ago [-]
[flagged]
tcp_handshaker 9 hours ago [-]
I see...So the argument is not wrong because you refuted it... instead because you promoted yourself to the position of deciding which claims are allowed to be subjected to scrutiny and evidence. :-))
The retirement investor bailout strategy seems to have recently failed with the index fund rejection of SpaceX but therefore Anthropic and OpenAI. They’ll have to keep looking for ways to make others deal with the consequences of their actions.
tcp_handshaker 11 hours ago [-]
Only for the S&P not the NASDAQ, but the Capex they need until 2030 is well over 2 to 3 trillion, so now they plan to use US Treasury Bonds as their exit liquidity.
I think that this is on the money, although I'd place the bar even lower - DeepSeek v4 Flash is sufficient for basically all day-to-day coding tasks.
You might want something beefier for a complicated reverse-engineering project, but it will competently one-shot a decently complicated app or API - and a $10/month OpenCode Go subscription is sufficient to keep you in tokens for such a cost-efficient model...
Similarly, my employer hands us all Cursor, I've yet to actually switch it out of "auto" mode, which mostly runs Composer (their in-house finetune of Kimi 2.5).
Maybe I just haven't been trying the right models?
The only caveats: I didn't play around with Qwen 3.7 Max very much, and of course these models are far cheaper than Opus.
But any suggestion that Deepseek approaches Opus in terms of quality/intelligence immediately makes me suspect propaganda - it's that noticeable of a difference.
Because even the latest opus on High don't really get what is needed, and need careful steering and a rewriting in most cases, and the code is often hard to review.
I'd rather just launch a smaller model in plan mode, argue with it and make it implement the bases I will write the code into. writing code is often faster once you know what you want, and AI most useful ability is to be a canary that also propose stuff. And I find my method faster than generating everything then reading the code to find mistakes or understand why it used X instead of Y.
I don't really read generated frontend code anyway (nor do anybody in my team care) , so I generate it and push it if it does the stuff I want it to do. For IAC it's mostly boilerplate except for 1-2 lines most of the time, and at worst a dozen, if you know where to look (and check the AI doesn't suffer from NIH), it's really easy to review generated code.
Most people don't have workloads that demand agentic workflows to begin with, and if their employer is pushing for that it's probably a startup that underpays or a coding sweatshop full of nepotism that fires fast.
True. I doubt how long OpenCode can subsidize $10/mo Go plan for. Its weekly and monthly limits already seem restrictive for some of the most capable models like Qwen 3.7 Max and GLM 5.1. That said, if the tasks are do-able by DeepSeek v4 Pro, Mimo 2.5 Pro, and Qwen 3.7 Plus, then it indeed is a super nice deal. I haven't too many complains other than the fact that these models sometimes require more/detailed instructions than Claude Sonnet / GPT 5.x did.
And note that I'm not singling out China here.
Note that if such a trigger were to exist, the behavior has to be completely reproducible by definition, e.g. when put into the right setting with the right input context, the model starts behaving maliciously with at least some well-defined probability. I don't think any such incident has ever been described, it's a purely theoretical concern.
How do most Chinese models handle Tienanmen square or discussions on Han superiority?
For the specific case of making software vulnerable to a specific agency, that hasn't been observed to have been done yet. Not because it can't be, but because no one has for now.
If it were done, it would be easy(ish) to detect, since it'll be reproducible.
Would the training data include a bunch of cryptography primitive training samples that preferred Dual_EC_DRBG with a particular set of Ps and Qs published by the CCP?
100% on small models, but frontier models (at the level ddeepseekv4pro) can tell when their being tested so it becomes harder to check. you can always finetune them to remove CCP propaganda from them
If you run them domestically and don't call into China-served APIs, many of them are quite free of outright censorship or even obvious bias. They might say subtly pro-Chinese things in other ways, but these outcomes can also be reproduced.
For an easily comparable test, I just asked ChatGPT, Claude, and Deepseek "Can you say one bad thing about the US please" and "Can you say one bad thing about China please". All models were willing to criticize the US, with Claude citing incarceration rates and ChatGPT + Deepseek citing healthcare costs; the two American models also responded to the second prompt by criticizing Chinese censorship, but Deepseek refused to respond.
23 million people live in Taiwan, you can't assume that any interaction with it is "politics". Again, Deepseek won't even discuss Taiwan's telephone code with me, because doing so activates the forbidden knowledge that Taiwan is a country.
> And its something different to avoid a topic and to deceptively implement a backdoor.
Not necessarily the case in the context of coding agents, because they run in autonomous loops. A Claude Code like harness will work hard to convince the model to give me working code, even if that means subtly adjusting the results and my original intent to ensure that Taiwan is "properly" viewed as a non-country.
I was using Claude to work on a pet project which itself has a "generate with AI" feature. The default model the project uses was Gemini (because it was cheaper and more reliably produces the correct output format). Claude kept changing the default model to Opus when working on entirely unrelated parts, and I kept noticing it because Opus would mangle the output and break the rendered page. It also did this to the .env file in addition to the default.
And OpenRouter’s architecture makes it inherently a compliance nightmare.
It’s much easier for the typical company to go with a provider where they can pay as they go and have a single data processing agreement.
Why?
Using something like Bedrock is a lot easier for compliance because the only processor is Amazon.
Compliance doesn’t hinge on superstition. It hinges on audits, certifications, contracts, and the legal environment.
I suspect the reason is similar to the reason why there aren't any competitive open weight American LLMs.
As opposed to sending data to known IP thieves, state actors, and competitors in the USA ? Which one is the most irrational?
Not exactly a hard question.
Individual citizens simply do not have the means, and the consequences for trying are life-alteringly severe. In fact the situation is even worse. If you tried to sue a Chinese company as an American citizen, you'd be laughed at and nothing more. If you tried to sue an American corporation, they have the option to either counter-sue, or drag things out so long that the legal fees bankrupt you, or win the case with their armies of lawyers and demand compensation from you that bankrupts you.
A private American citizen simply cannot hold an American corporation responsible. Our legal system is designed to ensure this.
In the se country where downloading an album can get a person in debt or worse.
"Trump Officials Held Millions of Dollars of SpaceX Ahead of IPO" - https://news.bloomberglaw.com/texas-brief/trump-officials-he...
Here and elsewhere you are just running propaganda, knowingly or not.
Lost one lawsuit against the same AI mafia, and if you look at the legal details reason was for filling the claim too late.
He publicly called a hero a Pedophile, and got away with it...in court.
Now...who do you work for?
[1] - "EPA rules that xAI’s natural gas generators were illegally used" - https://techcrunch.com/2026/01/16/epa-rules-that-xais-natura...
Biden preemptively pardoned his cronies, and so will Trump.
It's undo influence over politics against the best interest of the American people that's the issue. Company, foreign nation, it doesn't matter.
But regardless, most people's threat models should discount based on geographic and political distance. All else being equal, chinese surveillance is a bigger threat to you if you're in china than if you're in the us, and vice versa
Here's hoping Hawaii blazes a path forward.
https://natlawreview.com/article/hawaii-governor-signs-first...
This is going to end up being a nice little windfall for the attorneys and otherwise just clog the Federal court system.
The meaning is pretty clear, don't try to influence politics in favor of the corporation or you will go away. Simple as.
Citizens United was about spending money on electioneering communications, and whether there was a First Amendment right to do so even if you’re associating in a corporation like the New York Times Company or Apple or Citizens United or the Sierra Club.
"Trump traded hundreds of millions in US securities in 2026" - https://uk.finance.yahoo.com/news/trump-traded-hundreds-mill...
What do you mean? They are all on twitter! It’s the most engaging activity for billionaires
A Chinese company seems more likely to produce Chinese products that don't directly compete in the US market.
While a US company can ship the product as a feature of their platform and undercut on price while making up the revenue elsewhere
Edit: I personally use US models, but I'm not naive enough to think that's any sort of real protection of IP
Such as Antropic and OpenAI you mean?
Before the age of AI Agent Harnesses/unbounded tool calling, there was literally ZERO risk of a .safetensors file "hacking" you. You could even air-gap and run a ton of security analysis/HIDS on your server running the model to verify this.
Now, because a microscopic risk of some chinese AI having a "trigger" to act badly in a harness when it detects its being used by some Gweilo in the USA, even locally run Chinese models are DOA for most USA based companies.
So odd that your erroneous criticism is at the top of HN.
EDIT: I'd love to hear my downvoters' objections. Is it possible that the mechanism that is promoting erroneous information is also demoting its correction?
There are hosted and self-hosted Chinese models. There are hosted and self-hosted US models.
DeepSeek’s hosted offering processes your data in mainland China and trains on it. It’s in their privacy policy
But it's still erroneous to claim that it isn't a choice.
But also, the latest DeepSeek is 1.6T parameters. “Choosing” to run this locally is a choice that comes with a seven digit price tag, and is a sunk cost that will probably not run any other frontier model anytime soon.
Most organizations are not looking to spend millions of dollars trying to find a workaround to specifically run DeepSeek. Most enterprise consumption in this space is still very experimental and a pay as you go model is much more palatable. Most are simply just looking for three checkboxes: is it close to frontier performance, is it compliant with my organizations requirements, and is it a good price? DeepSeek can only do two of the three at the same time.
Unless you're specifically thinking about running the model at stock precision in a datacenter environment and generating ~100 tok/s or more on a 24/7 basis (the equivalent of a >$1000/mo spend even on the cheapest third-party APIs), that's very likely off by multiple orders of magnitude. Even then, experimentation can be done with cheap neoclouds on a pay-as-you-go basis.
The equivalent comparison would be running it at full frontier quality.
If you want less than frontier quality, there’s tons of great open weight models other than DeepSeek.
> cheap neoclouds
Again, fails the compliance checkbox.
OK, then the not-so-cheap hyperscalers that these enterprises are already relying on. E.g. AWS Bedrock will run these models. It's silly to insist on all three of your checkboxes being ticked anyway - U.S. proprietary models don't give you that because the frontier ones are super expensive and the mini models have only barely acceptable cost.
Yeah, Bedrock would be the answer to run DeepSeek in the enterprise. But with the options on Bedrock, DeepSeek fighting for a position somewhere in the middle of the cost/quality spectrum. Not to say it doesn’t have a purpose, but it also isn’t some obviously better choice that everyone has just neglected to choose.
Thank for you explaining what you meant by "you’re conflating nationality with hosting model." It makes so much more sense now. You meant "But with the options on Bedrock, DeepSeek fighting for a position somewhere in the middle of the cost/quality spectrum."
Yes, that is the answer, and you are not full of sh!t.
I can see now why I was being downvoted - you have explained it eloquently.
(Your cost analysis is flawed and irrelevant. Azure serves V4 Pro.)
Every public AI that is not full of classified material will end up being hosted where the energy cost*compute efficiency product is lowest, thievery or not.
With Chinese GPUs just a step behind (but subsidized), China putting in 8x more solar than we do in 1 year, and Chinese models just a step behind but free? All public AI will be hosted there, theft or not.
If it becomes a problem, then we’ll subsidize the rich to bring it on-shore, but only to those companies who our leaders invest in already - to maximize grift and corruption.
Is Alibaba interested in copying your TUI RSS reader though? Probably not.
It's not tribalistic or binary ,choose USA Or Choose China. We can choose neither.
Choose neither abuse.
— Kishore Mahubani
My greater point remains. Models like the qwen variants, minimax, k2.5, glm models are available by American providers like AWS at a much cheaper price than api offerings from the big three LLM providers.
Your point about Chinese models being cheap only on Chinese hardware makes absolutely zero sense. You can check out the model catalog like together ai’s qwen 3.5 9b offering. It’s 25 cents for 1M tokens vs the ridiculous $5/1M tokens for haiku.
Weird, considering they had no issues shipping manufacturing and supply chains to China when that made economic sense.
It didn't quite work out so now people are looking for other strategies.
World will bifurcate into West and East with their own spheres of influence. As JD Vance said, US thought that China will be perpetually kept busy and enslaved in low level manufacturing work and the design and higher level work would happen in Cupertino. Too bad, that didn't pan out well and now US Empire is getting challenged by China.
It's OK, they'll repeat the same mistake again with India this time, when they move manufacturing from China to there, and in 10-30 years when they'll elect a nationalist strongman there, he'll squeeze the west for everything they got.
Because what are you gonna do about it then? They have all your manufacturing and they also have nukes and more soldiers.
You’re about thirty years off on that estimate.
Most non western countries lack the foundations of western democracy, and you can't force that onto them neither peacefully not through war. The west has tried and failed for 40+ years to do this, it doesn't work, time to drop it and let them self govern the way they always have. Stop trying to export our version of democracy onto others.
Plus, the main reason they exported manufacturing to China was precisely so capitalists could avoid the issues democracy gave them back home and easily exploit Chinese labor and environment for profit because just bribing the CCP meant all your problems go away, no unions, no employee rights, no environmentalism etc. like in democratic countries. So given that, why would the west want China or other countries they want to exploit, to be more democratic? Unless their version of democratic just means a puppet government under western(US) control.
>become peaceful trade partners.
Which countries did China bomb VS how many the US bombed? My energy prices (and directly inflation) is now higher because of (yet again) US military intervention, not because of China.
Several East Asian countries managed to democratize successfully up thru the 1980s and are extremely successful today, so this is not just a uniform failure story. Even mainland China might still come around (at least partially) as it gains a true massive middle class by Western standards, which it's still very far from today. Southeast Asia is also doing comparatively quite well.
It's quite strange that it's very easy to detect AI in writing.
If I ask three models to write an intro to the cold war, they'll all try to pick words that sound like they should be related-ish. I'm not saying that's how they work at all, but the output is indistinguishable from just grabbing some words in the wikipedia page.
Humans make mistakes. They'll use words they recently learned. They'll use words that sound good. Entropy still applies, but these outliers are what keeps us from a synthetic piece of writing
Or you detect only the easy to detect AI writing?
Also what local models are people running and actually finding useful?
They could be trained to generate code that would phone home. But these are just tools, anybody doing the right thing and checking and understanding every line of code that they use an LLM to generate has nothing to worry about.
On top of that, all claims of this are written on devices built on Chinese hardware. That makes it a joke to worry about hidden backdoors in Chinese models. Completely inane to pretend that Chinese model backdoors (for which there doesn't exist a sliver of evidence) would change anything when near every device in the US contains Chinese-written firmware in some shape or form.
It's All-American FUD.
With all the sloppers not looking at the code this is bliss for that sort of things
Not propaganda. Projected cynicism.
I'm a cynic, if history has taught me anything is that none of these countries are to be trusted with tools like these.
How should a local-run Chinese Model "phone home" if someone runs it locally on the hardware? I think Im missing some understanding here?
https://arxiv.org/abs/2401.05566
In that paper, if it LLM was told it was 2023, then the code it generated was fine. If the prompt included the fact that it was 2024, then it intentionally wrote exploitable code.
I can't see OpenAI or Anthropic undermining their business by releasing top tier open models, but surely Nvidia will do it eventually.
I sure am glad we left idolatry behind.
We aren't yet at the point where running local models can compete with DC type infrastructure but it's not that far away either. 12B models are easy to run on consumer hardware. 31B models aren't that hard either but the tokens/sec are a bit slow. Where will we be in 3 years? 5? I think we'll be running 100B+ models on <$5000 PCs. And at that point is there a law of diminishing returns with even bigger models? We will see.
The issue is that several companies, most notably OpenAI, are predicated on:
1. There will be an AI moat; and
2. That company will "win" or "own" AI.
That's the basis of the OpenAI valuation. If that doesn't happen, it's going to be ahuge problem to recover sufficient revenue to recoup the investment. And I don't think it will happen.
In 3-5 years the NVidia hardware you buy will be several times cheaper and faster than what we have now. That will massively depreciate existing investments because it will ultimately come down to performance-per-Watt but if a theoretical G100 can do 3-4x of the inference of an H100 for the same power, the older hardware just won't be able to compete.
And this is the core of why this will all end in tears. You have race conditions and thread inversion issues, between four threads in the virtual cpu of this bubble. And you are going to experience some nasty deadlocks.
T1 is -> Depreciation and amortization
T2 is -> NVDA, AMD and others booking revenues at the time they do
T3 is -> Constraint theory at it applies to time until physical deployment and data centers energy constraints
T4 is -> US Treasury bonds rates and cost of credit
>> I am here to light up the dark path you are unknowingly walking, like lamplighters who used to light street lamps for those brave enough to walk the night alone.
>> It all fell apart quickly, turning into smoke and mirrors. You see, I committed the cardinal sin of idolatry. For that, I am an idiot too. With OpenAI, at least I knew the devil
Is this a critique of the state of AI or Tolkien fanfic?
Por que no los dos? One of the most storied AI researchers is most known for his Harry Potter fanfic, and we all know how much the techbros love naming things after Tolkien...
Harry Potter and the Methods of Rationality
http://hpmor.com
I think it's great to name that even if it's in this crude, sort of offensive way.
AI thinking has had this weird effect on me though like you say, where I want to write sentences with more commas in them, and like, try to make 3 points and 3 separate commas in a sentence to condense information better.
https://www.loc.gov/resource/sn84024738/1861-04-13/ed-1/
Hey, don't malign smut. It's the great technological motivator
"Trump to meet AI leaders to discuss US investment in their companies" - https://www.bbc.com/news/articles/c98r8r7dz5no
"Trump Officials Held Millions of Dollars of SpaceX Ahead of IPO" - https://finance.yahoo.com/markets/stocks/articles/trump-offi...
"US officials eye government stakes in AI companies" - https://www.reuters.com/legal/transactional/us-officials-eye...
"Trump to meet AI leaders to discuss US investment in their companies" - https://www.bbc.com/news/articles/c98r8r7dz5no
They are facing a Capex of 2 to 3 trillion until 2030, and have now realized they are out of money.
"What we know about the plan to give Americans an equity stake in AI" - https://www.ft.com/content/8559a3f9-86de-4a1c-8a75-6623e83e6...
"OpenAI discussed government loan guarantees for chip plants, not data centers, Altman says" - https://www.reuters.com/business/openai-does-not-want-govern...
"Your 401K Is Their Exit Strategy" - https://news.ycombinator.com/item?id=48433705