Discussion
Simon Willison’s Weblog
airstrike: [delayed]
zoba: I tried the new qwen model in Codex CLI and in Roo Code and I found it to be pretty bad. For instance I told it I wanted a new vite app and it just started writing all the files from scratch (which didn’t work) rather than using the vite CLI tool.Is there a better agentic coding harness people are using for these models? Based on my experience I can definitely believe the claims that these models are overfit to Evals and not broadly capable.
vonneumannstan: Were they kneecapped by Anthropic blocking their distillation attempts?
softwaredoug: I wonder how a US lab hasn't dumped truckloads of cash into various laps to ensure these researchers have a place at their lab
bilbo0s: They probably have tried, but you have to have more cash than those researchers feel they can get starting their own lab. When you think about the fact that their new startup lab would have the entire nation of China as essentially a captive market, you start to see how almost any amount of money would be too little to convince them not to do a startup.I think Alibaba needs to just give these guys a blank check. Let them fill it in themselves. Absent that, I'm pretty sure they'll make their own startup.I do think it'd be a big loss for the rest of the world though if they close whatever model their startup comes up with.
sosodev: I've noticed that open weight models tend to hesitate to use tools or commands unless they appeared often in the training or you tell them very explicitly to do so in your AGENTS.md or prompt.They're also struggle at translating very broad requirements to a set of steps that I find acceptable. Planning helps a lot.Regarding the harness, I have no idea how much they differ but I seem to have more luck with https://pi.dev than OpenCode. I think the minimalism of Pi meshes better with the limited capabilities of open models.
skeeter2020: Getting a bit of whiplash goin from AI is replacing people, to AI is dead without (these specific) people. Surely we're far enough ahead that AI can take it from here?Wild times!
mhitza: We've gone from AGI goals to short-term thinking via Ads. That puts things better in perspective, I think.
multisport: inb4 qwen is less of a supply chain risk than anthropic
butILoveLife: >I’m hearing positive noises about the 27B and 35B models for coding tasks that still fit on a 32GB/64GB MacIsnt it interesting that you never see someone say "I used this on my Mac and it was useful"Instead we get "you could put this on your Mac" or "I tried it, and it worked but it was too slow"I feel like these people are performing an evil when they are making suggestions that cause a waste of money.
kamranjon: I use Qwen 3 Coder Next daily on my mac as my main coding agent. It is incredibly capable and its strange how you are painting this picture as if its a fringe use case, there are whole communities that have popped up around running local models.
butILoveLife: Can I doubt your claim? I have had such terrible luck with AI coding on <400B models. Not to mention, I imagine your codebase is tiny. Or you are working for some company that isnt keeping track of your productivity.I am trying super hard to use cheap models, and outside SOTA models, they have been more trouble than they are worth.
simonw: The thing I'm most excited about is the moment that I run a model on my 64GB M2 that can usefully drive a coding agent harness.Maybe Qwen3.5-35B-A3B is that model? This comment reports good results: https://news.ycombinator.com/item?id=47249343#47249782I need to put that through its paces.
JLO64: Yesterday I test ran Qwen3.5-35B-A3B on my M3 Pro with 36GB via LM Studio and OpenCode. I didn’t have it write code but instead use Rodney (thanks for making it btw!) to take screenshots and write documentation using them. Overall I was pretty impressed at how well it handled the harness and completed the task locally. In the past I would’ve had Haiku do this, but I might switch to doing it locally from now on.
velcrovan: What the US has done is dumped truckloads of cash to make it likely that as a legal immigrant you will be abducted and sent to a camp.
xrd: I suppose this shows my laziness because I'm sure you have written extensively about it, but what orchestrator (like opencode) do you use with local models?
simonw: I've not really settled on one yet. I've tried OpenCode and Codex CLI, but I know I should give Pi a proper go.So far none of them have be useful enough at first glance with a local model for me to stick with them and dig in further.
sosodev: I really hope this doesn't hinder development too much. As Simon says, Qwen3.5 is very impressive.I've been testing Qwen3.5-35B-A3B over the past couple of days and it's a very impressive model. It's the most capable agentic coding model I've tested at that size by far. I've had it writing Rust and Elixir via the Pi harness and found that it's very capable of handling well defined tasks with minimal steering from me. I tell it to write tests and it writes sane ones ensuring they pass without cheating. It handles the loop of responding to test and compiler errors while pushing towards its goal very well.
a3b_unknown: What is the meaning of 'A3B'?
paoliniluis: what's your take between Qwen3.5-35B-A3B and Qwen3-Coder-Next?
kamranjon: In my experience qwen 3 coder next is better. I ran quite a few tests yesterday and it was much better at utilizing tool calls properly and understanding complex code. For its size though 3.5 35B was very impressive. coder next is an 80b model so i think its just a size thing - also for whatever reason coder next is faster on my machine. Only model that is competitive in speed is GLM 4.7 flash
xrd: What do you use as the orchestrator? By this I mean opencode, or the like. Is that the right term?
simonw: I use the term "harness" for those - or just "coding agent". I think orchestrator is more appropriate for systems that try to coordinate multiple agents running at the same time.This terminology is still very much undefined though, so my version may not be the winning definition.
Twirrim: I've been testing the same with some rust, and it's has spent a fair bit of time going through an infinite seeming loop before finally unjamming itself. It seems a little more likely to jam up than some other models I've experimented with.It's also driving itself crazy with deadpool & deadpool-r2d2 that it chose during planning phase.That said, it does seem to be doing a very good job in general, the code it has created is mostly sane other than this fuss over the database layer, which I suspect I'll have to intervene on. It's certainly doing a better job than other models I'm able to self-host so far.
Aurornis: > it's has spent a fair bit of time going through an infinite seeming loop before finally unjamming itself.I think this is part of the model’s success. It’s cheap enough that we’re all willing to let it run for extremely long times. It takes advantage of that by being tenacious. In my experience it will just keep trying things relentlessly until eventually something works.The downside is that it’s more likely to arrive at a solution that solves the problem I asked but does it in a terribly hacky way. It reminds me of some of the junior devs I’ve worked with who trial and error their way into tests passing.I frequently have to reset it and start it over with extra guidance. It’s not going to be touching any of my serious projects for these reasons but it’s fun to play with on the side.
simgt: > I do think it'd be a big loss for the rest of the world though if they close whatever model their startup comes up with.That's very likely to happen once the gap with OpenAI/Anthropic has been closed and they managed to pop the bubble.
mft_: Indeed; or, Europe badly needs a competitive model to hedge against US political nonsense.
tiahura: Competitive models are illegal in the EU.
riddlemethat: This is FUD. The US has dumped truckloads of cash to make it likely that masked men with no cameras and little training will parade around abducting anyone they even suspect of being an illegal immigrant, after even Yale admitted it's likely that more than 22M+ people came here illegally. https://insights.som.yale.edu/insights/yale-study-finds-twic...It'd be good if Congress could do something to remove the masks, put cameras on these agents, and for the local governments to stop fighting removal of all people who are here illegally so we can pretend we have borders again.
abhikul0: Are you running it locally with llama.cpp? If so, is it working without any tweaking of the chat template? The tool calls fail for me when using the default chat template, however it seems to work a whole lot better with this: https://huggingface.co/Qwen/Qwen3.5-35B-A3B/discussions/9#69...
gaoshan: ICE has been detaining Chinese people in my area (and going door to door in at least one neighborhood where a lot of Chinese and Indians live). I was hearing about this just last week as word spread amongst the Chinese community here (Ohio) to make sure you have some legal documentation beyond just your driver's license on you at all times for protection. People will hear about this through the grapevine and it has a massive (and rightly so) chilling effect. US labs can try but with US government behaving like it is I don't think they will have much luck.
sourcegrift: Yes. Yes, so true. And the phd types building these models are probably even scared in China that ICE will fly there to deport them.
gordonhart: Surely you know that this is an extreme misrepresentation? There are >35 million legal immigrants in the US. It's far from "likely" that as one of them you're abducted and sent to a camp.
autoexec: Give them time, they've only just started. They do waste a lot time abducting random US citizens though.
nu11ptr: What hardware do you have it running on? Do you feel you could replace the frontier models with it for everyday coding? Would/will you?
bigyabai: I'm getting ~30 tok/s on the A3B model with my 3070 Ti and 32k context.> Do you feel you could replace the frontier models with it for everyday coding? Would/will you?Probably not yet, but it's really good at composing shell commands. For scripting or one-liner generation, the A3B is really good. The web development skills are markedly better than Qwen's prior models in this parameter range, too.
mattnewton: I feel like we would disagree on the role of immigration in the US but I really appreciate you calling out how the current administration’s approach is only effective at making viral clips online. Meta comment, but it’s refreshing to talk with people who have different goals while still referencing a shared reality. Removing the masks and adding cameras shouldn’t be controversial unless your goal really is to make a paramilitary force for the president.
ljsprague: Are the people being detained in the country illegally?
hwers: My conspiracy theory hat is that somehow investors with a stake in openai as well is sabotaging, like they did when kicking emad out of stabilityai
liuliu: apples v.s. oranges. The later is true, Emad did get sabotaged (for not being able to raise money in time, about 8-month before he's leaving). Junyang didn't have that long arc of incidents.
ivan_gammel: Offering „You are welcome“ relocation package to Anthropic might be a good idea.
Imustaskforhelp: Given how American govt. has treated Anthropic, I think you might be right. EU truly has a remarkable opportunity to make Anthropic/Claude European.
hintymad: There has been tension between Qwen's research team and Alibaba's product team, say the Qwen App. And recently, Alibaba tried to impose DAU as a KPI. It's understandable that a company like Alibaba would force a change of product strategy for any number of reasons. What puzzled me is why they would push out the key members of their research team. Didn't the industry have a shortage of model researchers and builders?
jwolfe: This thread is about bringing these people to the US.
dude250711: Claude is incapable of producing a native application for itself, and is bad enough with web ones to justify Anthropic acquiring Bun.
cmrdporcupine: Anthropic has gone out of their way to make a point about how much they love and admire the US state and its defense sector. Only drawing the line at a very far point and even when they drew the line it was with a big thing about how they believe in the American defense sector blah blah blah.In any case, there's no way Anthropic's investors in Silicon Valley would countenance such a move.Also, I'm biased the logical place is Canada, not Europe. Much of the fundamental/foundational research on LLMs, and a large part of the talent, came from universities in Canada anyways.
cmrdporcupine: The unstated but obvious (to me?) goal of what ICE is doing is not to get large numbers of people out of the country, but to drive costs down for migrant labor by further disenfranchising them, making them scared, marginal, etc.If they actually thoroughly evicted non-status migrant workers they'd have a outright revolt on their hands from farmers and other businesses that depend on them.Instead those businesses can now take further advantage of the fear of harassment and/or deportation to drive down compensation and rights.Contrast with countries like Canada that have a legal temporary foreign agriculture worker program that provides a regulated source of seasonal migrant farm worker labour under a non-citizen temporary status, but with some rights (still often abused). It's notable to me as a Canadian that I don't see this being advocated on any large scale by either party in the US.Anyways, all this just to say that the jackboot clown theater is the point, not a side effect.
reactordev: They do have a revolt on their hands from farmers… go watch some of their pleas for help.
ecshafer: China is also giving them dump trucks full of cash though. Plus you have to content with the nationalism reason (unfortunately this has died off in America for too many). The idea of building your country is valued for most Chinese I have met. Plus China is incredibly nice to live in, especially if you have lots of money and/or connections. So you can work in China, get paid lots of money, feel like you are doing good. Or In America you can get paid lots of money, and get yelled at by people online because the Government wants to use your model.
petcat: > Or In America you can get paid lots of money, and get yelled at by people online because the Government wants to use your model.Isn't it just straight-up illegal in China to refuse the government from using your model? USA isn't perfect, but at least it has active discourse.
ecshafer: I would imagine if it isn't illegal its a very bad idea not to. But regardless, I would bet large amounts of money that you would never get any flack for doing anything for the government. If I went on X, Threads, Bluesky, TikTok and said "Hey I am a software engineer selling awesome new technology to the government and military!" I am going to get Americans attacking me for supporting Trump / ICE / FBI whatever the current issue of the day is. If I did the same on Douyin or Weibo the response would be able making China strong, and there would be no criticism of that choice.
mmaunder: Yeah that was my first thought is it’s a tit for tat poach. They got the Gemini researcher so google responded in kind.
cmrdporcupine: It's nothing like it would be if ICE was actually doing substantially more than fascist theater.There'd be no food on the tables, frankly. And people in Silicon Valley would have messy houses and algae in their pools.
Jcampuzano2: The reality is - it doesn't matter. The fact that they have had as many false positives as they have and the way they treat people in general causes it to have rippling effects even for people who are legally here, or are considering legally immigrating.The risk and level of publicity is just too high for many people to even consider, especially people already intelligent/capable enough to immigrate anywhere else that doesn't have these issues or stay in their own country.
reactordev: Honestly think it’s just a matter of resources and they would rather play theater for their leader than actually do the job. However, the effect has been felt.Soybean farmers are screwed.
neves: At least it has been decades since China Gov bombed innocent people in other countries. A peaceful and responsible government.
misnome: Who cares when you get a bonus per person either way?
plorkyeran: Limiting the supply of migrant labor drives costs up, not down, and the ICE raids have had a significant negative effect on businesses reliant on illegal immigrants.
cmrdporcupine: Do you have numbers on how many migrant farm workers have actually been deported or detained?Because going around and harassing and deporting other or non-essential non-status immigrants would drive labor costs down because of the chill it would put through those who are grudgingly tolerated.And besides, given the quality of personality ICE seems to be employing even (especially) at its highest levels, I simply assume there's corruption such that if I'm a large orchard or whatever I simply pay ICE to stay away.
kamranjon: I'm basically using the agentic features of the Zed editor: https://zed.dev/agenticIt's really easy to setup with any OpenAI compatible API and I self host Qwen Coder 3 Next on my personal MBP using LM Studio and just dial in from my work laptop with Zed and tailscale so i can connect from wherever i might be. It's able to do all sorts of things like run linting checks and tests and look for issues and refactor code and create files and things like this. I'm definitely still learning, but it's a pretty exciting jump from just talking to a chat bot and copying and pasting things manually.
quantum_state: I would second that Qwen3.5 is exceptionally good. In a calibration, it (35b variant) was running locally with Ada NextGen 24GB to do the same things with easy-llm-cli in comparison with gemini-cli + Gemini 3 Pro, they were at par … really impressive it ran pretty fast …
vardalab: q4 quant gives you 175 tg and 7K pp, beats most cloud providers
gaoshan: No, all of the specific cases I heard about were Chinese people that were naturalized citizens (some for decades) who were cuffed and detained for a few hours before being released. As others have said it doesn't really matter, though. It's the sentiment that counts.
zozbot234: What Anthropic was complaining about was training on mass-elicited chat logs. It is very much a ToS violation (you aren't allowed to exploit the service for the purpose of building a competitor) so the complaint is well-founded but (1) it's not "distillation" properly understood; it can only feasibly extract the same kind of narrow knowledge you'd read out from chat logs, perhaps including primitive "let's think step by step" output (which are not true fine-tuned reasoning tokens); because you have no access to the actual weights; and (2) it's something Western AI firms are very much believed to do to one another and to Chinese models all the time anyway. Hence the brouhaha about Western models claiming to be DeepSeek when they answer in Chinese.
leptons: Chinese people are very racist towards non-Chinese. It might seem like a happy utopia, but if you aren't Chinese, then you may not really enjoy your time there. It may not be quite as bad as being black in rural US south, but being black (or anything non-Chinese) in China is still not going to be a good time.
lacoolj: I wonder if an american company poached one/all of them. They've been pretty much bleeding edge of open models and would not surprise me if Amazon or Google snatched them up
janalsncm: Anthropic has one nine of uptime right now. One.https://status.claude.com/If AI could effectively replace people, you wouldn’t need CEOs to keep trying to convince people.
mungoman2: Not sure what the uptime is meant to signal. People have quite low uptime as well…
jug: Huh? Servers aren't people and thus have completely different expectations, or what am I missing here
aruggirello: I'm getting ~27 tok/s on my 5060 Ti 16Gb (just bought for ~€600) on an old 8th gen Core 7 (2017, with DDR4 memory); I'm using the iGPU for the graphics, so the card is free for inference - I'm using Qwen3.5-27B-UD-IQ3_XXS, and almost 72k context - I went with ik_llama.cpp [0] - Qwen3.5 is very promising, I'm definitely going to use it.I also tried ZSE [1] but experienced some issues, I was hoping I could squeeze in 4bit or 5bit quants[0] https://github.com/ikawrakow/ik_llama.cpp[1] https://news.ycombinator.com/item?id=47160526
px43: Wild to call 1.42 billion people racist despite having met very few of them.
storus: More likely some high ranking party member's nepobaby from Gemini sniffed success with Qwen and the original folks just walked away as their reward disappeared.
ahmadyan: source?
misnome: I've been playing with 3.5:122b on a GH200 the past few days for rust/react/ts, and while it's clearly sub-Sonnet, with tight descriptions it can get small-medium tasks done OK - as well as Sonnet if the scope is small.The main quirk I've found is that it has a tendency to decide halfway through following my detailed instructions that it would be "simpler" to just... not do what I asked, and I find it has stripped all the preliminary support infrastructure for the new feature out of the code.
shaan7: > that it would be "simpler" to just... not do what I askedThat sounds too close to what I feel on some days xD
sheepscreek: That sounds awfully similar to what Opus 4.6 does on my tasks sometimes.> Blah blah blah (second guesses its own reasoning half a dozen times then goes). Actually, it would be a simpler to just ...Specifically on Antigravity, I've noticed it doing that trying to "save time" to stay within some artificial deadline.It might have something to do with the system messages and the reinforcement/realignment messages that are interwoven into the context (but never displayed to end-users) to keep the agents on task.
wood_spirit: Yeah that happened to me with Claude code opus 4.6 1M for the first time today. I had to check the model hadn’t changed. It was weird. I was imagining that maybe anthropic have a way of deciding how much resource a user actually gets and they had downgraded me suddenly or something.
anana_: I've had even better results using the dense 27B model -- less looping and churning on problems
VWWHFSfQ: > China is incredibly nice to live inI'm sure it's a very nice place to live if you're content to just stay quiet in society and never put a political sign in your yard or even just talk about the wrong thing with your friend in a WeChat.
cyberax: This is an exaggeration. Nobody in China cares about what you speak with each other privately, and people talk about stupid policies all the time. The government cares about _public_ actions.In practical terms, if you're not kind of person who would want to run for an office in the US, China is incredibly comfortable. Cities are safe, with barely any violent crime. Public drug use is nonexistent. And with the US-level AI researcher income, you'd be in the top 0.1% earners.
WarmWash: Racism in even the worse parts of America doesn't even begin to touch the racism present in monocultural/monoracial countries.
red2awn: The "distillation attacks" are mostly using Claude as LLM-as-a-judge. They are not training on the reasoning chains in a SFT fashion.
zozbot234: So they're paying expensive input tokens to extract at best a tiny amount of information ("judgment") per request? That's even less like "distillation" than the other claim of them trying to figure out reasoning by asking the model to think step by step.
kylemaxwell: Everything on that page has two nines, so not sure what you're trying to say here.
WarmWash: There is no source. But the party in China does have ultimate control.There would never be an Anthropic/Pentagon situation in China, because in China there isn't actually separation between the military and any given AI company. The party is fully in control.
lreeves: In my experience Qwen3.5/Qwen3-Coder-Next perform best in their own harness, Qwen-Code. You can also crib the system prompt and tool definitions from there though. Though caveat, despite the Qwen models being the state of the art for local models they are like a year behind anything you can pay for commercially so asking for it to build a new app from scratch might be a bit much.
WarmWash: Making viral clips is exactly what they want.Their goal is for every one person violently detained, 10 decide to leave on their own, and 100 decide to not come in the first place.
KerrAvon: "Goal" implies there's a plan instead of just wanton cruelty for the sake of cruelty.
maxglute: > get yelled at by people online because the Government wants to use your modelWell duh, as recently demonstrated, an US model used by the US gov will 100% end up murdering actual children sooner than later, in this case less than a calendar year in some far flung war that many Americans do not support. Alternatively PRC model used by CCP might kill in some hypothetical future but for national reunification/rejuvenation that many Chinese support. At the end of the day, researchers and population on one side sleeps more soundly.
w10-1: It sounds like the lead was demoted to attract new talent, quit as a result, and the rest of the team also resigned to force management to change their minds.If so, I'm happy that the team held together, and I hope that endogenous tech leads get to control their own career and tech destiny after hard work leads to great products. (It's almost as inspiring as tank man, and the tank commanders who tried to avoid harming him...)(ducking the downvote for challenging the primacy of equity...)
mijoharas: It'd be great if they went to Mistral!
0x3f: Have they had a lot of false positives? Almost every story I see seems to fall apart on further investigation. To be clear, I'm sure they have some false positives, but do they have a lot of them relative to any other immigration system?
petcat: > nobody in China cares about what you speak with each other privately, and people talk about stupid policies all the time. The government cares about _public_ actions.https://news.ycombinator.com/item?id=47252833My comment and the linked video says otherwise. The guy was in a private group chat and said some nasty things about the police for confiscating his motorcycle. Now he's arrested and in the Tiger Chair.How are we explaining this?
maxglute: Group with 75 people. That's a crowd, doesn't matter if gated behind QR code invites. Shit talk cops and gov with the bois is fine. Shit talk / soapbox in a crowd (virtual or real) and get caught or reported = drink tea on the menu.
4RealFreedom: "especially people already intelligent/capable enough to immigrate anywhere else that doesn't have these issues or stay in their own country" Isn't that the point? Come here legally or don't come at all.
cmrdporcupine: Sure, but the difference is that while the Chinese state is measurably awful on all sorts of human rights things within their own borders... they're not currently dropping bombs on foreign cities, starving a neighbour of critical petroleum shipments, or heavily funding an ally to slowly exterminate a population.
fc417fc802: What point are you trying to make here? Are government abuses somehow inherently better or worse depending on where they happen?Do you imagine an invasion of Taiwan won't involve dropping bombs?I feel like we should be able to agree that providing authoritarian regimes with high tech tools is immoral in the general case.
impossiblefork: I'm not sure goals are totally aligned though. The current models are created by enormous expense. We know that many stages are done incorrectly. I am confident that they can be replicated without any unique US knowledge.At the moment my impression is instead that the issue is computational resources. It's important to stay near the frontier though, and to build up ones capacity to train large models.Consequently I don't think we need Anthropic. It wouldn't be terrible if they came. Especially if they picked a nice location. Barcelona would be very nice, for example.
sosodev: Around 20ish tokens a second with 6-bit quant at very long context lengths on my AMD AI Max 395+I’m trying to use local models whenever possible. Still need to lean on the frontier models sometimes.
bdangubic: try to protest in america and see how that works out for you long-term. or say protest against genocide in gaza at an uni or generally in public…
cyberax: Sigh. Let's not invent things? You can protest anything in the US just fine, with generally no consequences. Heck, our local _high_ _school_ students go out and protest everything to weasel out of classes.
cheema33: Trump admin did put people in prison and then deported them, for doing nothing more than protesting.Not as bad as China sure, but not as good as other civilized nations.
fc417fc802: Let's just clarify that visitors don't have the same rights as citizens. Whether or not you agree with the current administration's policies hopefully we can agree that it is entirely reasonable for them to deport foreign political dissidents more or less at their discretion.If you want to put this to the test try crossing the Canadian border and when they ask you the purpose of your visit respond that it's to attend a protest.
expedition32: If memory serves the father of the Chinese bomb studied in America and went back. It may be inconceivable to Americans but Chinese patriotism exists.Besides you can live a comfortable life in PRC nowadays or live in a racist America.
danny_codes: China city life is amazingly convenient. Trains and subways are just such an enormous quality of life boost. Add to that the relative cleanliness of having nearly zero homelessness and you’ve got something very compelling.I will say we are winning in accessibility. China doesn’t have much of a ramp game
softwaredoug: All very true.I wonder if you max out your options in China. It seems the Party is suspicious of ambition and high profile winners. I'm sure you can live comfortably, but there's a ceiling.
danny_codes: That’s not relevant to normal people. If you’re a billionaire with aspirations of power then it’s probably good there’s a ceiling. Sure beats having Elon randomly firing your public servants while high on ketamine.
cmrdporcupine: My point is as a non-American I feel no allegiance to either state, and current events don't make me sympathetic to the geo-political aims of the USA. So I don't see a strong moral case for this tech being an especial purvey of either party.If you'd asked me two years ago my answer might have been different.And to the original point, yeah, I would feel entirely justified in the critique of engineers in providing tools to the US defense apparatus at this point.
Terr_: Depends, how are we defining "false positive"? Ex:1. Detained the incorrect person2. Detained the correct person, with the correct legal status3. Detained the correct person, with the correct legal status, but in unlawful circumstances4. Detained the correct person, with the correct legal status, in ostensibly-lawful circumstances, but in a way which is unconstitutional or crazyAn example of the final category are the immigrants that spent years being vetted, following the law, and doing expensive paperwork to be citizens. ICE snatched them when they showed up on at the last second as they were to take their citizenship oath. [0] Not because of anything they did, but because today's Republican party has decided that it's OK to hurt people based on their "shithole" country of birth.[0] https://www.theguardian.com/us-news/2025/dec/30/us-citizensh...
0x3f: These are all forms of false positives but the most popular news stories seem to be where they detain the correct person, correct legal status, lawfully, and the story happens to gloss over the facts about the legal status and focuses on the hardship. Yeah, it's a hardship to be split from your family, I can't deny that. But I'm not aware that most countries are very sympathetic to illegal immigrants.If anything I find the stories featuring white/European people oddly racist because they seem to assume that I, the reader, will assume a white/European person couldn't possibly be in violation of immigration rules. But all the ones I've read turned out that they were indeed in violation of immigration rules.Overall as a potential immigrant to the US myself, I find the process capricious and that US citizens by birth don't fully appreciate how painful it is or why it shouldn't be that way. But I don't find it notably worse or more onerous than the vast majority of immigration policies of other countries in practice.
platinumrad: I'm not sure what you're talking about. The most popular stories are the ones when they detain US citizens, rough them up, and then dump them on the side of the road somewhere without even apologizing.[1] https://www.theguardian.com/us-news/2026/jan/13/ice-immigrat...[2] https://www.pbs.org/newshour/nation/a-u-s-citizen-says-ice-f...[3] https://www.propublica.org/article/immigration-dhs-american-...
snackerblues: Won't someone think of the poor tech millionaires' pools and their cleaning slaves? If we get rid of the slaves, they'll have to pay at first world cleaning rates! :(
bdangubic: what is the issue with having a ceiling?
WarmWash: Star athletes really hate being told they can't score more than 10 goals in a season because it's unfair to the other weaker players. The players will either leave to go play somewhere else, or they become weaker players themselves.
bdangubic: wowsa - wasn't expecting star athletes and sports to enter this conversation... wild!
bdangubic: this is funny if you are being sarcastic
cyberax: Oh, I fully support their right to protest.It just looks a bit ridiculous when students walk out in protest against things that are far outside the influence of their school, city, or even state.
seanmcdirmid: They already kind of do, but I think anyone who was into US money has already left for it, and the money China is throwing at the problem is pretty good also. You can also have a lot more influence in a Chinese company without having to adopt a weird new American corporate culture.
petcat: > A peaceful and responsible government.People in Hong Kong died. Over 10,000 were arrested and many are still in prison. The rest are permanently disgraced in their social-credit society.Again, USA is not perfect, but let's not dream up some fantasy about the CCP.
cyberax: This "social credit" thing is dead in China.
petcat: As an American, I have no fear of calling the US President a pedo or saying Fuck the Police on my Twitter. Not the case in China. It's horrifying.https://reclaimthenet.org/china-man-chair-interrogation-soci...
Barrin92: > I have no fear of calling the US President a pedo or saying Fuck the Police on my Twitter.Does that matter? In China people don't judge the state of their civilization by how easily you can insult the police but whether you need to be afraid to meet them on the street. "I can insult my pedophile president" (who doesn't care if you do) isn't exactly a flex.It does tell us something though that the evaluation of American life now consists of parasocial interactions with the president on social media. I'm starting to belief Bruno Maçães, ex Portuguese secretary of state, was prescient with his diagnosis that American material society has rotted to the point where life is now entirely defined by virtual interactions. That's the difference between China and the US today.
px43: 9% uptime?
AgentME: One 9 would be 90% (aka 0.9)
red2awn: LLM-as-a-judge is quite effective method to RL a model, similar to RLHF but more objective and scalable. But yes, anthropic is making it more serious than it is. Plus DeepSeek only did it for 125k requests, significantly less than the other labs, but Anthropic still listed them first to create FUD.
simonw: It's the number of active parameters for a Mixture of Experts (misleading name IMO) model.Qwen3.5-35B-A3B means that the model itself consists of 35 billion floating point numbers - very roughly 35GB of data - which are all loaded into memory at once.But... on any given pass through the model weights only 3 billion of those parameters are "active" aka have matrix arithmetic applied against them.This speeds up inference considerably because the computer has to do less operations for each token that is processed. It still needs the full amount of memory though as the 3B active it uses are likely different on every iteration.
zozbot234: It will benefit from a full amount of memory for sure, but AIUI if you use system memory and mmap for your experts you can execute the model with only enough memory for the active parameters, it's just unbearably slow since it has to swap in new experts for every token. So the more memory you have in excess to that, the more inactive but often-used experts can be kept in RAM for better performance.
EnPissant: The ability to stream weights from disk has nothing to do with MoE or not. You can always do this. It will be unusable either way.
zozbot234: Agreed but for a dense model you'd have to stream the whole model for every token, whereas with MoE there's at least the possibility that some experts may be "cold" for any given request and not be streamed in or cached. This will probably become more likely as models get even sparser. (The "it's unusable" judgmemt is correct if you're considering close-to-minimum reauirements, but for just getting a model to fit, caching "almost all of it" in RAM may be an excellent choice.)
e1g: Claude Code recently downgraded the default thinking level to “medium”, so it’s worth checking your settings.
nekitamo: Thank you. The difference was quite noticeable today.
jtonz: As someone that started using Co-work, I feel like I am going insane with the frequency that I have to keep telling it to stay on task.If you ask it to do something laborious like review a bunch of websites for specific content it will constantly give up, providing you information on how you can continue the process yourself to save time. Its maddening.
ekianjo: There is no capital in the EU
OsrsNeedsf2P: That's 99% is two nines?
janalsncm: It was 98.xx this morning when I posted.
jefftk: This is pretty pedantic, but I think it's usually rounded. 1=90%, 2=99%, 3=99.9%. I'd say 98% is "not even two nines" but not "one nine".
iso-logi: If they are illegal citizens, they need to go.
refulgentis: It’s beautiful at 37 to still see new phrases sometimes, illegal citizens is a quite beautiful one, lol. (also, note the post is clearly about, to put it in your terms, legal citizens)
janalsncm: 9% would be 0.09 which is no nines.
strangegecko: Constant military drills around Taiwan isn't peaceful or responsible.China is bullying lots of countries in the SCS (ramming Philippine coast guard ships, building military installations in the SCS, ...). Not peaceful or responsible.
nozzlegear: Legalize illegal citizens
xienze: > Trump admin did put people in prison and then deported them, for doing nothing more than protesting.Link? I’m guessing we’re going to see that this definition of “protesting” involves being aggressive and directly in the face of law enforcement officers, not merely holding a sign at a distance.
zzrrt: That’s pretty funny when compared with the rhetoric like “AI doesn’t get tired like humans.” No, it doesn’t, but it roleplays like it does. I guess there is too much reference to human concerns like fatigue and saving effort in the training.
martin-t: This is what happens when a bunch of billionaires convince people autocomplete is AI.Don't get me wrong, it's very good autocomplete and if you run it in a loop with good tooling around it, you can get interesting, even useful results. But by its nature it is still autocomplete and it always just predicts text. Specifically, text which is usually about humans and/or by humans.
maxglute: AKA defending itself against separatists and sovereignty intrusions from much less powerful aggressors with unreasonable amount of restraint. One would argue overly peaceful, and irresponsible to the point of detrimental peace disease. BTW PRC settled most border disputes in recorded history with most concessions, majority over 50%, that objectively makes PRC the most peaceful rising power in recent history. Even in SCS PRC was second last to militarize, the other disputees started land reclamations and militarization first (apart from Brunei), aka a fucked around and find out situation. Even then all PRC did was build a bigger island, instead of glassing theirs, PRC coast guard last to weaponize as well.
bobthepanda: Yeah, the Hyundai factory fiasco kind of dashed the idea that the enforcement would spare people working in favored industries setting up in the US.
genxy: The Hyundai factory "enforcement" wasn't even legal. Those workers were here to train US workers and the Hyundai employees had proper visas for this.https://apnews.com/article/immigration-raid-hyundai-korea-ic...https://www.koreatimes.co.kr/foreignaffairs/20251112/hundred...https://www.pbs.org/newshour/nation/attorney-says-detained-k...The regime is powered by racism and doesn't think through things.
limagnolia: Allegedly, though the local labor unions seem to disagree. I guess we'll have to wait for the facts to come out in court.
bandrami: It really is like having an intern, then
androiddrew: Which dense model are you referring to? The dense model isn’t finetuned for code instruction according to the model card.
anana_: https://huggingface.co/Qwen/Qwen3.5-27BI wasn't aware of that, which page mentions that?
phatfish: If China was serious about a military solution for Taiwan they would be invading right now while the US is unloading into the desert.
CamperBob2: There's a process. If the government won't follow it, why should the "illegals?"
jongjong: Interesting reading this. It reminds me of my time in cryptocurrency sector. I suspected that some team members were paid by Ethereum folks to sabotage our project. Why do I suspect Ethereum? Because our project founders ended up switching to the Ethereum ecosystem and ignored/suppressed better solutions from their own ecosystem. I think there's something about tech hype which attracts these kinds of people who like to play dirty.
jiggawatts: "Papers, please." comes to the US of A.
nomel: Every administration since the foundation of ICE has funded ICE and immigration policy/border operations [1].[1] Removals by president: https://www.migrationpolicy.org/article/biden-deportation-re...
segmondy: There's a very big difference between xenophobia and racism. Racism is much worse.
rprend: Anthropic is a great case study in why uptime doesn’t matter. The service is so valuable that you can have one nine uptime and add $9bil ARR in 3 months.
root_axis: Yep. The veil of coherence extends convincingly far by means of absurd statistical power, but the artifacts of next token prediction become far more obvious when you're running models that can work on commodity hardware
cheema33: > Let's just clarify that visitors don't have the same rights as citizens.Yunseo Chung was not a visitor. She came to the United States from South Korea at age 7. She was arrested last year for peacefully protesting. Charges against her were dropped but the govt. canceled her green card.The govt. has been trying to deport her since then, but the courts keep blocking it.https://humanrightsfirst.org/yunseo-chung-v-trump-administra...While the legality of these actions are being debated in courts, I think most of us can agree that this is reprehensible behavior on part of the Trump admin.
sosodev: I’ve been running it via llama-server with no issues. Running the latest Bartowski 6-bit quant
roywiggins: what's an "illegal citizen"
cheema33: > Link? I’m guessing we’re going to see that this definition of “protesting” involves being aggressive and directly in the face of law enforcement officers, not merely holding a sign at a distance.Please read up on this one example of a US permanent resident. And then justify the actions of the govt against Yunseo Chung.https://humanrightsfirst.org/yunseo-chung-v-trump-administra...
ncallaway: > illegal citizens???
oefrha: I would take their displayed uptime with a huge grain of salt. The other day Claude Code and claude.ai web was completely unavailable for me (Claude Code got into logged out state and couldn’t even log in) for at least two hours, they showed hours of “elevated errors”, yet not a single minute of downtime was recorded. And then there was yet another outage finally with recorded downtime a few hours later…Edit: This incident: https://status.claude.com/incidents/kyj825w6vxr8
pixelatedindex: Can’t believe we have people who unironically say things like “illegal citizens need to go”. For shame. You probably are one too, would you leave first?
speedping: open blogpost → ⌘ + F "pelican" → 0 results ???
fc417fc802: I agree that particular example is reprehensible.I never claimed to condone the actions of the current admin. The examples of people being deported for protesting that I am familiar with are student visa holders. While I don't personally support the examples that I am aware of, I also recognize that in those specific cases the executive branch appears to be within the bounds of the law. I don't even object to the executive branch having the power to cancel the visas of political dissidents in the general case, merely to how they are choosing to apply it.It's surprising to me to learn that a green card could be revoked for protected speech. That ought to fall well outside the bounds of the law IMO. Green cards and visas are entirely different things.
selcuka: You are not wrong, but after having started working with LLMs, I have this feeling that many humans are simply autocomplete engines too. So LLMs might be actually close to AGI, if you define "general" as "more than 50% of the population".
brightball: Bartowski? Like Chuck Bartowski from the TV show?
gunsle: You have to show ID to pick up a prescription or open a bank account. You have to show ID for routine traffic stops. This is such a juvenile, tired argument.
baby_souffle: “You show ID at the bank” is a classic, juvenile and tired argument because it swaps in a voluntary transaction for state coercion.The concern isn’t IDs exist—it’s who’s demanding them, in what context, and what happens if you can’t comply on the spot.
vuurmot: how about explaining why without giving an analogy?
jiggawatts: Not to mention that the USA has a long history of looking down their nose at the USSR for doing exactly what the USA is doing now.I forgot that HN is mostly filled with a younger generation that might not get the reference.The "Papers, please." quote is a common trope in spy movies, books, etc... about the former Soviet Union.
vicchenai: Been running the 32B locally for a few days and honestly surprised how well it handles agentic coding stuff. Definitely punches above its weight. Only complaint is it sometimes decides to ignore half your prompt when instructions get long, but at this size I guess thats the tradeoff.
ramgine: Running 32b on what hardware?
RandyOrion: This is sad for local LLM community. First we lost wizardLM, Yi and others, then we lost Llama and others, now we lost Qwen...
binarycrusader: You absolutely do not have to show ID to pick up every prescription; just some, which is also dependent on state law, federal law, and pharmacy.But also, I don't care if it's a tired argument--this isn't about how things are, it's about how we want them to be. I don't want to live in a state action-coerced society.
zerebos: Yeah the page you linked even shows the benchmarks in coding for this model, so I'd be curious where that claim comes from
levocardia: "product market fit is when people are ripping the product out of your hands and everything is breaking constantly" - seems bullish to me
abhikul0: Thanks, i'll check his quants.
nurettin: I'm running it on a run of the mill ryzen server with 128 GB RAM with llama.cpp, it seems as intelligent as gpt4.
nozzlegear: Who are you quoting?
throwup238: In my experience all of the models do that. It's one of the most infuriating things about using them, especially when I spend hours putting together a massive spec/implementation plan and then have to sit there babysitting it going "are you sure phase 1 is done?" and "continue to phase 2"I tend to work on things where there is a massive amount of code to write but once the architecture is laid down, it's just mechanical work, so this behavior is particularly frustrating.
dripdry45: I hope you will excuse my ignorance on this subject, so as a learning question for me: is it possible to add what you put there as an absolute condition, that all available functions and data are present as an overarching mandate, and it’s simply plug and chug?
whalesalad: What hardware are you running this on?
Zetaphor: I'm running this exact same setup on a Framework Desktop and I'm seeing ~30 tokens/second
nvader: Another vote in favour of "harness".I'm aligning on Agent for the combination of harness + model + context history (so after you fork an agent you now have two distinct agents)And orchestrator means the system to run multiple agents together.
Zetaphor: This has also been my understanding of all of these terms so far
throw5t432: > building military installations in the SCSMany countries in the SCS are doing this. In fact China was late to the game, as Vietnam did it much earlier.
throw5t432: > People in Hong Kong diedDo you have a legit source for this? When I search for information, I only found this case, “Luo Changqing, a 70-year-old Hong Kong cleaner, died from head injuries sustained after he was hit by a brick thrown by a Hong Kong protester during a violent confrontation between two groups in Sheung Shui, Hong Kong on 13 November 2019.”None of the other legit sources claim the police killed any of the rioters.
sciencejerk: Why are they giving them away for free?
beepbooptheory: Why keep using it then? I simply still read websites. It's not always great but sounds better than whatever that weird dynamic is!
lynndotpy: Well, the problem aren't just the NSF funding cuts. Everyone else is already dumping truckloads of cash. There's also the public health situation (who wants measles or polio?), the risk of retaliatory attacks from the countries we're at war with, etc. You could write paragraphs about why the US is less attractive to researchers.When I was a deep learning PhD in the first Trump administration, US universities were already very deeply affected by the Muslim ban, and so a lot of talent ended up in other countries.Sibling commentators are rightfully pointing out that foreigners, especially those who would not be recognized as white, face an onerous and risky customs process with long-term and increasing risks of deportation. When you see a headline like the NIST labs abruptly restricting foreign scientists, _everything_ else feels uncertain. Even if someone doesn't believe they're personally at risk for deportation, they're still seeing everything else.And then it all boils down to a reputational thing. The era where we were the top choice for research is in the past. If you start a PhD in the US on your resume during this era, you might be anticipating how you'll answe the question of why you weren't good enough to get accepted somewhere better.
sciencejerk: Where do researchers go instead of USA then? Genuinely curious
wood_spirit: Thank you thank you you give me hope :)But how do you see the current thinking level and how do you change it? I’ve been clicking around and searching and adding “effortLevel”:”high” to .claude/settings.json but no idea if this actually has any effect etc.
vasco: Look up how many of the main people are Polish.
elcritch: Recently it seems that even if you add those conditions the LLMs will tend to ignore them. So you have to repeatedly prompt them. Sometimes string or emphatic language will help them keep it “in mind”.
shinycode: If found it better to split in smaller tasks from a first overall analysis and make it do only that subtask and make it give me the next prompt once finished (or feed that to a system of agents). There is a real threshold from where quality would be lost.
joecool1029: recently being within the past 24 hours lol: https://github.com/anthropics/claude-code/releases/tag/v2.1....
darkwater: > Re-introduced the "ultrathink" keyword to enable high effort for the next turDoh.
nekitamo: In my tests, Qwen3.5-35B-A3B is better, there is no comparison. Better tool calling and reasoning than Qwen3-Coder-Next for Html/Js coding tasks of medium size. Beware the quants and llama.cpp settings, they matter a lot and you have to try out a bunch of different quants to find one with acceptable settings, depending on your hardware.
ilaksh: Does anyone know when the small Qwen 3.5 models are going to be on OpenRouter?
armanj: they're already there ?? https://openrouter.ai/qwen/qwen3.5-27b
ilaksh: Like 4B, 2B, 9B. Supposedly they are surprisingly smart.
Sakthimm: Yep. The 9B has excellent image recognition. I showed it a PCB photo and it correctly identified all components and the board type from part numbers and shape. OCR quality was solid. Tool calling with opencode worked without issues, but general coding ability is still far from sonnet-tier. Asked it to add a feature to an existing react app, it couldn't produce an error-free build and fell into a delete-redo loop. Even when I fixed the errors, the UI looked really bad. A more explicit prompt probably would have helped. Opus one-shotted it, same prompt, the component looked exactly as expected.But I'll be running this locally for note summarization, code review, and OCR. Very coherent for its size.
BoredomIsFun: > Very coherent for its size.I found them to be less than stellar at writing coherent prose. Qwen 3.5 9b was worse in my tests than Gemma 3 4b.
BoredomIsFun: Different one. Bartowski is a minor celebrity in the local LLM world, together with Unsloth.
nopurpose: How do those companies make money? Qwen, GLM, Kimi, etc all released for free. I have no experience in the field, but from reading HN alone my impression was training is exceptionally costly and inference can be barely made profitable. How/why do they fund ongoing development of those models? I'd understand if they release some of their less capable models for street cred, but they release all their work for free.
indrora: Ostensibly, a mix of VC funding and that they host an endpoint that lets them run the big (200+GB) models on their infrastructure rather than having to build machines with hundreds of gigs of llm-dedicated memory.
fc417fc802: Why did Meta release theirs? The better question is, why not? If you aren't at the cutting edge and don't have a moat then releasing them is pure reputational upside with zero downside.
varshar: As per Anthropic support (for Mac and Linux respectively) -echo 'export ANTHROPIC_EFFORT="high"' >> ~/.zshrc source ~/.zshrcecho 'export ANTHROPIC_EFFORT="high"' >> ~/.bashrc source ~/.bashrc
swiftcoder: There's no huge reason to bring them to the US. Plenty of US corporations have maintain overseas offices. Even if its impolitic to employ them directly in China, you can employ them in other offices (for example, Amazon has been known to do this with their Singapore offices)
rwmj: The small spend may be worth it to destroy US proprietary AI companies.
gdiamos: Results as good as Qwen has been posting would seem to trigger a power struggle.I think companies that don’t navigate these correctly eventually lose.
wongarsu: But on inference they have to compete with other inference provider that just has a homepage, a bunch of GPUs running vllm and none of the training cost. Their only real advantage are the performance optimizations that they might have implemented in their inference clusters and not made public
justinclift: > As someone that started using Co-work, I feel like I am going insane with the frequency that I have to keep telling it to stay on task.Used to have the same thing happening when using Sonnet or Opus via Windsurf.After switching to Claude Code directly though (and using "/plan" mode), this isn't a thing any more.So, I reckon the problem is in some of these UI/things, and probably isn't in the models they're sending the data to. Windsurf for example, which we no longer use due to the inferior results.
eunos: > never put a political sign in your yard or even just talk about the wrong thing with your friend in a WeChat.Practically, how many care about that? Consider that in other part of the world they also cancel folks based on social media opinion...and that Benjamin Franklin's opinion on security and freedom? Thats terminally online phenomenon only. I once tried to bring that without specifically mentioned that it came from ol Ben himself to folks IRL. Many thought it was some anarchist blabbers.
soulofmischief: Claude Opus does this constantly for me, no matter how I prompt it or what is in my AGENTS.md, etc. It is the bane of my existence.
DiogenesKynikos: Imagine if you forgot to bring your ID card to the bank, and they grabbed you and the next thing you knew, you were in a concentration camp in El Salvador.
alecco: Try repeating the key parts of the prompt. Perhaps with a numbered list or an action plan.
fragmede: Yes, but the difference in degree, and how are material.
petcat: This US administration (or any admin) would almost certainly impose export controls on US AI technology before it would allow one of the frontier model providers to be acquired/relocate outside the US. It did the same thing when ASML wanted to acquire Cymer (California company that provides the EUV light source technology). The acquisition was only allowed under strict technology sharing/export agreements with the Dutch government.Europe really just needs to rally behind Mistral. That's where they should dump their cash.
ivan_gammel: Having one „champion“ is flawed European approach. We need local competition and headhunting to make it fly.
azinman2: Hard to compete in an environment that’s anti-996 and the pay is so much less.
dvdkon: Do you consider high working hours to be a benefit akin to higher pay? I think fewer hours and less money is a fair deal for employees.
z3t4: [delayed]
EnPissant: Unlike offloading weights from VRAM to system RAM, I just can't see a situation where you would want to offload to an SSD. The difference is just too large, and any model so large you can't run it in system RAM, is going to be so large it is probably unusable except in VRAM.
zozbot234: Unusable for anything like realtime response, yes. Might be usable and even quite sensible to power less-than-realtime uses on much cheaper inference platforms, as long as the slow storage bandwidth doesn't overly bottleneck compute.
realharo: And yet almost all of the most popular sports leagues in the US have a salary cap rule.
theshrike79: Chinese companies don't always operate on purely capitalistic principles, there is sometimes government direction in the background.For China, the country, it's a good thing if American AI companies have to scramble to compete with Chinese open models. It might not be massively profitable for the companies producing said models, but that's only a part of the equation
user34283: This thread is largely pointless political back-and-forth were predictably the comments with a more positive opinion on current US immigration enforcement will be flagged.To get back to the original point, personally I doubt sentiment on US immigration enforcement would be so significant a deterrent for Chinese talent, who may not share the political views of the American left for whom this is a big concern.
Matl: > were predictably the comments with a more positive opinion on current US immigration enforcement will be flaggedGiven the tactics employed by ICE, it's a true shock and horror that most people have more humanity than that.But I guess a person who can't form a grammatically correct sentence is an example of the sort of people who can rest easy,
reactordev: Turn down the temperature and you’ll see less “simpler” short cuts.
smokel: For the uninitiated: Interestingly, it is not advisable to take this to the extreme and set temperature to 0.That would seem logical, as the results are then completely deterministic, but it turns out that a suboptimal token may result in a better answer in the long run. Also, allowing for a little bit of noise gives the model room to talk itself out of a suboptimal path.
mejutoco: Setting the temperature to zero does not make the llm fully deterministic, although it is close.
rcbdev: It's all tied up in real estate and horribly inefficient savings accounts.
nnoremap: Or the 2026 version: 'Hey Claude set your thinking level to high.'
littlestymaar: Every developed country on earth has an immigration policy and an administration dedicated to enforcing it.No other developed countries have masked goons abducting people in public wearing civil clothes and masks and disregarding every laws of the country (violating private property and foreign embassies, deporting national citizens, and numerous other preposterous bullshit).Immigration policy enforcement is normal, the madness that has been running in the US for a year isn't.
ramgine: Does keeping it on one stick make it more performant? I have a epyc server with 1tb of 64 gig sticks and a 3060, looking to maximize what models I can run on it
sowbug: Unless you know something we don't, Alibaba hasn't been accused of distilling or stealing any Anthropic assets.
varispeed: Opus 4.6 found in my documentation how to flash the device and wanted to be clever and helpful and flash it for me after doing series of fixes. I've got used to approving commands and missed that one. So it bricked it. Then I wrote extra instructions saying flashing of any kind is forbidden. Few days later it did again and apologised...
MarsIronPI: [delayed]
jjcc: There are 2 groups of new Chinese immigrants in the US, they are quite different:1.Those who arrived through legal channels (most studied at U.S. universities and remained on H1B visas, with a smaller number through EB5 or other visa categories) and eventualy got green card.2.Undocumented immigrants, which include several sub-groups/waves. In the 1990s, most came from just a couple provinces, Fujian and southern Zhejiang. After COVID, they were from different parts of China and entered through the southern border.The contributors to AI development belong to the 1st group. They are spread across the country but a large number work in high-tech companies in Northern California.The 2nd group was intially concentrated in New York and Southern California (Los Angeles area). Later they have expanded into nearby regions. They provide labor for Chinese-owned small businesses such as restaurants, grocery stores, and hotels.There is an industry created largely by Chinese political dissidents helping Group 2 through asylum applications using fake materials and exploiting common Western beliefs or narratives about China like human rights concerns. For example, Alysa Liu’s father is an asylum lawyer.ICE enforcement efforts would likely focus more on Group 2 if they are knowledgable. Ohio should not be a high-priority area. I could be wrong due to changes over time. One indicator you can observe: Are there many Chinese-owned small businesses in your area?
hedora: ICE arrests citizens and legal immigrants on a regular basis. Only 5% of the people they arrested had an immigration related conviction on their record:https://www.cato.org/blog/5-ice-detainees-have-violent-convi...The Bay Area is mostly exempt for now because, after Trump announced ICE was going to surge in SF, a bunch of tech billionaires with economic interests in the region convinced him not to.Also, over the last year, there have been a bunch of high-profile arrests of Ohioans by ICE. In one example, they arrested someone for showing up to their immigration hearing, leaving their young kid separated from them outside the court.
MarsIronPI: [delayed]
coredog64: "Papers, please" is not just ID, but also authorization to travel internally. The idea that asking for ID* is anywhere equivalent is asinine.*Reminder that folks visiting the US on a visa are legally required by the terms of said visa to always carry upon their person at least a copy of identity papers backing up that visa, and that this law has been in place for a very long time.
gaoshan: They were operating in a traditional "China town" neighborhood for the detentions and the neighborhood they were going door to door in is one populated mostly by white collar professionals (tech, college professors, etc.).
malwrar: +1 to this, anecdotally I’ve found in my own evaluations that if your system prompt doesn’t explicitly declare how to invoke a tool and e.g. describe what each tool does, most models I’ve tried fail to call tools or will try to call them but not necessarily use the right format. With the right prompt meanwhile, even weak models shoot up in eval accuracy.
mongrelion: > [...] _but not necessarily use the right format._This has also been my experience. But isn't the harness sending the instructions on how to invoke a tool? Maybe it is missing the formatting part. What do you think?
kelvinjps10: But the constitution is not worded as if they don't have the same fundamental rights. Even in other countries, it is the same; this is done to prevent slavery and unjust incarcelation. So visitors have the same fundamental rights to free speech, fair trial, etc. The US has also agreed to international conventions. But the current administration seems to not care
kelvinjps10: Xenophobia is as bad as racism. The fundamental difference between xenophobia and racism, it's that one is applied because of where you're from and the other your race. But you can receive the same downsides with both.
azinman2: Fair deal for the employees but hard to compete against smart, well resourced people who are working 996. Everything with AI is moving so fast that moving slower makes you irrelevant.
raven12345: As someone active in both English and Chinese media, I always feel like who relying on only one is brainwashing, just like Wumao. There's no difference here; it's always about the government control，destroying US company... In reality, free services have always been a competitive strategy for businesses in China, from ride-hailing to bike-sharing, all about grabbing market share and competing for potential users. Daily active users are what Chinese companies care about most.
nurettin: In my case it definitely helps, but keeping 64k context window would probably require more than 64 GB.
Balinares: What's the selling point of these quants vs the Unsloth ones?
BoredomIsFun: Sometimes unsloth has broken ones for a particular model, sometimes no quants at all, and there is subtle difference in behavior.
politelemon: 60 to 70 on a 5080, but only tinkering for now. The smaller models seem exceptionally good for what they are, and some can even do OCR reliably.
cpburns2009: What quantization are you running on the 5080? I'm waiting to receive mine.
brorfred: Can you expand on this? SL accepts any credit card for purchasing single tickets and I assume you can buy an SL card using cash in for example Seven Eleven? Also, the issue with bank ID when you are robbed is identical to any bank app anywhere, isn't it?
