Join the discussion
Write your take first — we'll ask for email only when you're ready to publish.
- Hacker News
- This is a real head scratcher. Unless this is a very short term action it seems to have only downsides for everybody:
- people pay much more for US models than Chinese models because right now they're the best. Once they're no longer the best (since you don't get access to them) why would anyone pay several times as much for the same result?
- once you get a high amount of tokens flowing into China instead of US companies, they will train on those chats and their rate of improvement will only accelerate, making US models even less attractive over time
- the sky-high IPO are dead in the water, since their story of "we will replace a good chunk of all knowledge work in the world, capturing a few % of total global spend relating to it" turns into "we will make a bunch of money out of a few dozen S&P 500 paying for the best, and some pocket money out of whoever uses our overpriced models that are as good as Chinese models" - far less money overall. Losing access to untold billions of investor money certainly won't improve performance for the US labs
- all the non-US people start asking themselves why they're funneling money to US corporations who barely share any of the secret sauce compared to Chinese corporations who share plenty when it comes to LLM, including the models themselves (at least for now)
- Chinese models have significantly less guardrails, making for better end-user experience
- there is a small but non-zero chance Euros get off their asses and invest into AI, making something halfway decent and further fracturing the market which cuts into US profits
So what's the benefit here? I thought the Mythos situation was the current admin taking revenge on Antrophic for not kissing the ring, or simply looking for a bribe, but no matter which way I look at it it's a self-own. The only way this would make any sense is if AGI is imminent, which I don't think even the boosters are arguing at this point.
Theoretically US could outlaw Chinese models, but I'm not sure what it's supposed to accomplish as the rest of the world certainly won't, especially as long as they release open weights models that you can run without phoning home.
- You have thought about this longer than anyone who made the decision. Stop caring, it only makes your head hurt.by tencentshill
- Yea not sure who would put money into an OpenAI or Anthropic IPO at this point. After this episode I promptly bought the hardware I needed to run local models (~$25k) and am extremely impressed. The price tag alone is worth the peace of mind knowing I will always have a locally running LLM that nobody can take away. I don't need Mythos-level intelligence to do 99% of my day to day work; Opus 4.8 was more than enough and I am pretty close to that with the open source models. So what happens when this hardware inevitably gets cheaper and cheaper and people realize they can run the models they actually need locally and without handing all their data over to these companies?
Only if you're doing cutting edge research or some highly, highly niche project would you need the frontier models.
by jliptzin - Well I was expecting it but if it's not going to become available in this subscription cycle for me I will cancel oai subs as I did with claude ones months ago...
moving to open weight models is trivial now, with optimizations and stuff glm 5.2 is roughly the same price as the best models around from multiple vendors.
unless I could atleast try and see Sol perform like 10x better I don't really have a reason to switch back.
I used Fable for like what 2-3 days at most and didn't really feel it was so much better, only difference was I had to prompt it less, not to get what I want but to get to a working output. Code quality was still shit, still made bad plans and analysis and so on.
by minraws - > I used Fable for like what 2-3 days at most and didn't really feel it was so much better
It was a lot better. I can't believe people say this.
by andxor - Are these models still relevant for people outside the US? I get the impression we're stuck on GPT 5.5 and Opus 4.8 pretty much permanently now, and relying on Chinese models in future.by pixelpoet
- That’s assuming China would not start controlling the access to their models.
- Hopefully. I hope this eventually nudges my employer to use Deepseek or other Chinese models.
- Not only that, but using Opus 4.8 [1m] right now outside the US, and suddenly I only have a 500k context window. I really hope this is just a strange Claude Code bug, but I had access to a 1 Million window before, and it wouldn't entirely surprise me if context window length becomes another US export restriction.
The Anthropic page here seems to say that Max users should have access to the full 1 Million window for 4.8:
https://support.claude.com/en/articles/8606394-how-large-is-...
I was already setting up my infra to experiment with GLM 5.2 and its 1 Million token window before this happened. I think I'm glad I did.
EDIT: Found a solution, seems Claude Code 2.1.193 (or an earlier version I didn't notice) changed default settings, so that if you have Autocompact turned on it occurs at 50% of the context window. If you turn off Autocompact, the full 1 Million context window is restored. Another example of Claude Code quietly changing default settings sigh
by SyneRyder - You shouldn’t build a business that relies on any of these models. It’s a geopolitical and sovereignty risk now. Someone could just rug pull your entire stack.by cryo32
- If the US really cracks down on frontier model access, you'll see them make Chinese open models illegal. You might say "oh well, let them try", but they will just put direct and secondary sanctions on every company whose systems have used Chinese models in some way. They just have to make an example out of a few international companies and no one will dare to use Chinese frontier models, at least commercially.by Sol-
- Dont worry, chinese models will distill frontier ones, quite fast.
The excuse they give is borderline childish. I get the thing about slow rollout, make sure partners get to fix the bugs, etc...
But bad actors are hard working motivated entities with tens of thousand of fake ids, and american citizens working for them, for pennies.
All while the ones like or you sit at a crossfire which is borderline useless.
I cant wait to see what Qwen did with the massive distillation they made out of Opus 4.8 and Fable aka Mythos aka pretty sure they jailbroke it.
by seviu - The real reason, afaik, that the US is trying to restrict access to SOTA models is that a very large component of USA tailored access and surveillance relies on exploits and weaknesses that these models will easily detect. Thus, it really is an export control issue, but it has nothing to do with offensive capabilities. Offensive capabilities always exist, but pervasive defensibility would upset the asymmetric advantage that attackers, especially the USA, currently have.
There are now Asian models coming , optimized focused on cybersecurity defense at a high level, so I suspect this will be a relatively moot point soon.
LLMs are not great at creating exploits, but they are really good at detecting them. That asymmetry alone is enough to destroy the “offensive capabilities” narrative. Yes, mythos can find exploitable bugs, even write bench exploits. But real exploits require a good dose of human psychology, and most of the tools needed are off the shelf available anyway. You still need a real cybersecurity expert to effectively weaponize a zero day into a deployable exploit.
But an LLM can inspect payloads, packages, and blobs en masse and find those exploits in a way that was wholly impractical before, so the asymmetric attack advantage is dissolved by strong LLMs.
The USA is trying to protect its cyberwarfare advantage, not protect against attackers. The exact opposite, actually. Porous security is a huge advantage to technologically advanced state actors.
by K0balt - > There are now Asian models coming , optimized focused on cybersecurity defense at a high level, so I suspect this will be a relatively moot point soon.
Source?
by awestroke - The biggest concern is identifying "who". If the US government says only US citizens can access a model, how do they enforce that. Anthropic and OpenAI will use Persona (a company funded by Peter Thiel) to verify user identity. Verifying your identity with a government ID and linking that to AI is the dream of a surveillance state. Agents running on your computer, accessing your internet accounts, access to your personal conversations with AI, and accessible by the government is just wild.
I'm hoping this is a call to action for local AI.
by mrinterweb - Local AI will never be as powerful as cloud-based models, at least not in the foreseeable future. We’re talking about a difference between 7B and 750B+ parameters, a 100x scale gap.
I think the trend will be running open-weight models with the provider of your choice. You can always switch providers, avoiding vendor lock-in, with the trade-off of privacy concerns. You have to trust that the provider actually behaves the way they claim.
by maxloh - > If the US government says only US citizens can access a model, how do they enforce that.
It's easy if only enterprise users have access. It just becomes yet another compliance issue.
They won't let the public use it, even if citizens, because that would undermine the goal of controlling its use.
> I'm hoping this is a call to action for local AI.
If US frontier models stop being publicly available then they stop mattering, it also becomes hard to justify further investment in US AI companies if distribution is so locked down. The Chinese models will be the frontier.
by zarzavat - No, not really. The biggest concern is the US government claiming it can decide who gets to use a product and brazenly kneecapping competitive companies who do such things as "write contracts and comply with them."by estearum
- Remember how China turned its tech industry into a smoking ruin - make making them all submit to political priorities:
Ant Group: China halted Ant’s IPO and forced a restructuring
Alibaba: China fined and politically disciplined Alibaba
Didi: China punished Didi after its US listing by removing its apps, freezing users, forcing delisting
Tutoring platforms: banned profit from core school-subject tutoring.
Tencent gaming: restricted youth gaming froze approvals
NetEase and gaming companies: licence freeze stopped game companies from shipping games.
Meituan: fined Meituan and forced changes to its labour and platform model.
Huya/DouYu: blocked Tencent’s game-streaming merger, stopping commercial consolidation in a major entertainment market.
Boss Zhipin / Full Truck Alliance: froze new users after listging in the US
Crypto companbies: banned crypto trading and mining, forcing exchanges offshore.
Think it's not happening to the US?
tourism - people afraid to visit
tariffs - wrecking ball to all businesses
defence - why would anyone buy US weapons after Greenland and Canada
internet clouds - Greenland made Europe decide that the US clouds can't be trusted, now sovereign computing matters and MS/AWS/Google are feeling it
finance - no one trusts the US not to turn people into "non members of global society" by banning them from visa and credit card and banking systems
by wewewedxfgdf - Orange and apples... China has very intentional policy behind those decisions. The US... Not so much. I don't buy that Trump and his whole cabinet are as dumb as they look, but they are only motivated by profit. And ignorance.by HeavyStorm
- The US is now doing a softer and broader version of the same thing to trust-based export sectors. It’s not the same method but! it is the same mechanism. The main difference is that the US damage is more reputational than structural, so it could be reversed faster (only if policy stops telling customers that dependence on America is a political risk)by garn810
- This was generated by AI and slightly re-written or end capped by a human. Anyone that knows who Meituan and Boss Zhipin are wouldn't make the claims this post makes. It's not reasonable that someone who could list off these companies and incidents would believe China's tech industry is in ruin. There's no way. This poster clearly promoted for the list, and wrote his commentary around it. Sad.by asadotzler
- Some of these were very good decisions imho, from someone who spent two months in Chineese rural area around ~2019.
- Tutoring platforms were a plague on Chineese youth that increased the weight of their already _very_ heavy load (tbh, i think and education reform might have been preferable, this is a stopgap, but at least it is something).
- Ant group was offering predatory consumption loans to rural China, which to me felt a lot like the "revolver credits" that plagued my country in the 80s and 90s and pushed to many to suicide (the surname cam from their english name, "revolving credit", and because my countryside had a lot of hunting rifle available to whomever). Considering how rural china is mistreaded by Chineese state and general government (and imho this is a real weakness in China politics), having this group by a huge fine for their practice and a general debt forgivness was great. Curtailing Ant's power is also good.
- Stopping consolidation is a great way to keep a market free.
- Crypto companies: mining diverted power from villages who couldn't compete on purchasing power to mining wharehouses in some state. The ban is great for the rural population at least. Also, if that can curtail the birth of Chineese cryptobros, great for the mental health of the country.
by orwin - China tech industry is smoking ruin? On what planet are you living?by watwut
- Remember how China turned its tech industry into a smoking ruin
Not really, no. What planet is this on?
by apexalpha - From my perspective curtailing Ant's plans was positive regulatory action.
Political priorities and good governance is why we have government.
by jhancock - It's really interesting that someone can know all of these domestic Chinese names and yet declare the industry generally a "smoking ruin". Is it from a newsletter or something?
Because anyone who used these companies' products in China would see a pretty large ecosystem that's making a lot of money.
by nixon_why69 - Given how the WH operates these days, this is ripe for corruption. Imagine the WH dislikes the CEO of a biotech company, while appreciating the attitude of a competitor CEO. What is to stop them from stalling on giving acess for the latest model to the company they don't like?by aristocrazy
- "Given how the WH operates these days" I've got bad news for you...by AzzyHN
- You dont even need corruption. Once its political it becomes subservient to political interests.by nonethewiser