It's very clear to me that however much you like Claude and Anthropic models, every professional user shoudl be actively cultivating a set of alternatives and fallbacks to use and seeking to ensure the overall ecosystem is as open, standardised and portable as possible.
Anthropic is highly opinionated, dogmatic and idealistic. They have made a pragmatic decision already to cut loose a whole swathe of users so they can pursue areas higher up the value chain. that was essentially the original decision they made when they went after software engineering instead of general users. The fact they are repeating this over and over by continuing to carve out sets of users and usages they don't want reinforces this as a persistent core principle rather than a single pragmatic business decision.
What this means is that at a higher level, their core principle is to sacrifice users they don't "need" rather than provide a broad, accessible service. Which means in turn, you have to accept that when your usage conflicts with their interests, you are at very high risk they will cut you out. If you already depend on Anthropic in any critical path of your daily life, this will be a significant issue.
Why, though? I have half a terabyte of VRAM in cards in a datacenter and it’s absolutely not clear that anything I’ve learned using these to program or inside an app has taught me anything significant.
I talked to my friend about all the things I do to ensure my personal assistant behaves reasonably. The prompt engineering, the fake site evals, everything. And then I have Opus 5.5 do it and it doesn’t need anything. It’s simply fully aligned to performance by default.
And what did I learn anyway. Going from one model to another changes what works. And you can simply learn everything from the Chinese model in days. The whole “building skill” thing is bogus.
I have this library of techniques, a scant few of which I’ve made available on my blog, and what is reality instead: the next model blows the need for the techniques into the dust. Back on level playing field.
It’s a false god, this cultivation. Waste of time. Use the best model you can. Once you hit the guardrails move out to the other models. Even Opus 4.8 is fine. If you legitimately want to dodge the guardrails use the -cyber models.
I haven't used an Anthropic model, or GPT heaviliy in a year. And I have access to both, I dont support either of those companies. I think they're both led by genuinely evil people.
Sam and Dario came from the EA-rationalist space. Their ultimate goal is to create an all powerful ASI before anyone else does and make sure no one can do it again. Of course it will be aligned, to their interests as a group and their moral philosophy.
In real life they "just" want to capture the US economy, replace most workers with AI, become trillionaires, then control the world by making ClaudeGPT influence peoples opinions and write backdoors into code. And they need to get open models banned in the process.
The Chinese just want to make money and beat America. Mistral and BFL want to prove Europe can still innovate. They don't have evil goals so they're not evil people.
"The Chinese just want to make money and beat America. They don't have evil goals so they're not evil people."
They straight up censored Kylian Mbappe a few days ago for cracking a joke about Mao Zedong. And very much all the models developed by China would have to conform with Chinese government's viewpoint, including censorship and such. Do you truly believe that they are not evil, or they will not ever be evil?
"However, in a narrow set of cases, we believe AI can undermine, rather than defend, democratic values. .. Mass domestic surveillance. Fully autonomous weapons."
How many other companies said no to Trump in any way?
Several companies (I believe at least Apple and Costco) still have DEI programs internally, if you count that as saying no to Trump. Though it's still crazy to me that most companies killed those programs voluntarily because Trump and Republicans don't like them, not because of any real law or contractual obligations.
But the main reason I wanted to reply was just to say -- saying no to Trump, by itself, does not mean that one is not evil. Not all evil in the world is Trump's evil.
Absolutely. I have been [cyber]'d so many times even Claude jokes about it now. Claude is great but it's operators are making it unreliable for me. I have moved one of my projects elsewhere and will move the remainder tomorrow as a stop gap until I get the right hardware set up to run abliterated local models for some percentage of my needs. Their message has been received loud and clear.
Issue with Opus is that you can’t do regular pr reviews where you say please try find security issues at best there is silent refusal and worst you get downgrade.
best part is, if you actually know what you're doing, you cna absolutely get it to find bugs/vulns/etc. so it just really about preventing long horizon agentic tasks or people who don't know what they're doing or those who are doing legit non-offensive security sfuff
What's ridiculous is this is all futile. Anthropic will go through all of this effort of attempting to block these things and it will be for nothing in the end.
For those who do not play with Local AI much. Qwen4 was announce to be released the first week of Nov. The preview for qwen4 is currently one of the best local models you can use, and is roughly opus 4.6 quality on local hardware that costs 3.5k-4k
Many people are expecting it to sit somewhere above Opus 4.8 to opus 5 (not above, but the unquanted version of their largest model is expected to perform around here)
Anyone who cares to keep their current capability should definitely be looking at localAI and using it for at least their more routine tasks , even if it is qwen4 27B quant 4 on their gaming rig.
If the weights aren't released yet it isn't a "local model".
That said, I hope there is a Qwen4 max. Because Qwen3.8-Max is the only thing that comes close to Kimi K3. We need competition in the open-weights space.
The preview for qwen4 == qwen3.8 flash next. And that is released currently.
I believe they have confirmed there will be Qwen4 Max, Plush, Flash, and 27B as the 4 open weights they plan to release. Starting next week, but they didnt specify if they plan to release all 4 at once, or if they plan to just release one of them, and trickle the others out over time. (They likely will do the latter to stay in the news longer)
They keep doing anti-competitive things over and over again, and the end users seems to refuse to see it. They do whatever they can to get vendor-lock in.
We see it, but they dont have a competitor. OpenAI cant compete and they are the closest. I have both for work, and Opus 5.5 is in a league of it's own. I use opus for everything until it bitches, and I use openAI for whatever it bitches about just to get past it.
At this point, I would like to consider other options, but if I want to get work done, opus 5.5. Hell even sonnet 5.5 is better than anything OpenAI has to offer when it comes to coding.
They do have competitors. They just dont have anyone with models equivalent to theirs. They win in the large context long thinking space by a lot.
I love local models, and I run 3.8 Flash Next on a dedicated machine. Alone with a number of image and video generation models. But if I'm doing REAL work, that isnt tinkering for fun on a side project, im using claude, because Qwen3.8 just cant handle large context project well without going off into left field.
Im SUPER jazzed for qwen4 with 1mil token context native support. I HOPE it a step up from 3.8 flash next. Ill be playing with all the qwen4 models the moment I can, but I dont see how someone working on large code bases doing REAL enterprise grade work is seriously using any local models in any capacity that comes remotely close to Opus 5.5+Claude
The size of a code base does not determine the size of the model you need. It's much more correlated with the complexity of the problem. I work in a monorepo with millions of lines, no issue; I have a side project with 20k lines, I had to step in and pull it out of non-progress - albeit by asking a single question
> But if I'm doing REAL work, that isnt tinkering for fun on a side project
Sorry, but this is either dilettante talk or an unhinged exaggeration - Qwen, DS, GLM and MM are perfectly fine for REAL work, unless you are an overly ambitious vibe coder, but then no model can save you. The Cl hand-holding may feel like something but it won't lead to the completion of real REAL work.
Honestly I feel like most people going "oh but I can't do this without Claude" are just suffering from FOMO. Especially since a lot of metrics are just subjective, you feel like you "have" to use the "best of the best glub shitto v5" model or else you won't be able to produce the best todo app.
Eh, we all say we don’t need the best model, but especially for more personalized request/Advice there is sort of a FOMO feeling, where you are afraid weaker models will miss something and directs you down a more inefficient path.
There is nothing comparable. I will take Opus5.5 (or even 5.0) over Astra for almost any workload. OpenAI's models are just too misaligned to work well outside their benchmarks, or a strict environment with very clearly defined goals and constraints.
And I've spent around $100-$120k in api calls this year (total, both providers). Terra is awesome (price-performance) on noncreative things, but I don't allow it to write anything besides one-off scripts. Astra is decent for adversarial review, but only when you watch closely for her to not derail project into directions that are away from goal but safe to build, and Sol I honestly don't know where I'd use.
would you rather treat your AIs like people, or would you rather find yourself down the line treating a person like an AI, but without really noticing it? because you are what you practice
What kind of work you do? I'm doing Linux kernel work in my free time, arguably one of the most complex coding task you can do, and the open source models are sufficient
You mean 3.1 Pro now, I assume? K3 is nice, but more expensive than an Claude subscription on a token-for-token basis (until the US frontier rugpull eventually comes), and not as good for my needs unless I'm trying to do "naughty" things, like reverse engineering.
And yes, I am taking full advantage of the scammy subscription scam, because it's the best value by far until the inevitable rugpull. Guess I'm not as smart as you and your "grownup" setup though.
Fortunately it is not a one horse race. GPT 6.1 Sol is an excellent model, sure it has some weaknesses (e.g. frontend design), but it also has strengths too (e.g. computer use / control; which is great for automated E2E QA).
On software engineering I find it's generally close enough to Opus 5.5 where it's more situational and even personal preference / skills / prompting style on which is better.
Most importantly, while neither company are saints, I don't curse at OpenAI models for classifiers anywhere as much as Claude models.
> but it also has strengths too (e.g. computer use / control; which is great for automated E2E QA).
Using 6.1 Terra for E2E workflow testing feels like cheating. I thought it would have all kinds of problems with the action space on certain steps, but it runs like a deterministic unit test.
I am convinced that OAI has the superior models for most practical scenarios. I've tried Claude recently and it makes me wonder if some of yall aren't actually massive trolls. It's genuinely not in the same room of performance if you place one foot outside junior programming or shiny blog posts about rewriting factorio to run on a ps2.
They are doing similar things, yes; but they are not the same especially when it comes to `cyber` classifiers, or the extent to which these models will assist you with ML and AI research.
I find that logic similar to thinking both political parties are the same.
They are both doing a lot of the same things like watermarking, guardrails for their p(doom), fighting distillation after distilling human knowledge, crying wolf about open models while being irresponsible in their own training and usage.
They are run by the same circle of people and I see more similarity than difference.
There is also an argument to be made that the political representatives are largely the same, despite glaring differences. They are largely part of the same oligarchy, the corruption - using gerrymandering to pick their voters instead of letting us having any real choice.
Their entire “PBC” governance structure was designed from the beginning so that they could entirely avoid accountability to shareholders. Their board answers to no one.
I suspect when the company eventually falls prey to the intractable economic forces they’re up against (Google, OAI, Deepseek, commercial real estate, power) and the company goes bankrupt and gets stripped for parts because all inputs are commodities and capital becomes the only limiting factor, we’ll see regulation that prevents this PBC scam from ever being pulled again on public markets.
It’s a good thing that companies have a fiduciary duty to their shareholders. It’s such a lesswrong/autistic trope to invent these kinds of complicated tricks to work around realities they don’t like, and when pension funds and everyday Americans start bagholding their junk debt and are forced to bail them out, they’re going to face unimaginable blowback.
The real problem is everyday Americans invest in big managed funds where they can't just say "I don't want AI stocks please leave them out my portfolio". Or index funds where even the managers don't get that choice. We need a hybrid model where you as a retail investor leave the day to day decisions to an expert but you can always override them if you want to.
Fiduciary duty forces companies to maximize profit and that's not a good thing. Every dollar paid out to shareholders is both a dollar not invested in growing the economy through more production and a transfer from the working class to capital. It also forces lower product quality, layoffs and anti competitive actions.
We need more companies that choose to not go down that path. They should just put workers and customers first instead of their weird "ethical" principles and cult like beliefs.
> Actually cutting off potential users would probably be counter to short term profit maximization that shareholders care about.
Shareholders know where the money is but it's not where you think it is. People should have learned that already, after watching what happened to the big web platforms.
idk i think theyre less idealistic and more monopolistic.
if they think their model is the best at making things safe, then their model should be helping train at any chance, to embed the same safety into new models
It’s not professionals that need to something- it’s absolutely critical we have some kind of credibly neutral body providing the actual models. What’s not clear at all is how to do that since it’s such an expensive activity.
For now we have a great situation with the open weights, but there is absolutely no guarantee that continues, and more the opposite- they may stop.
Anthropic is acting like any regular selfish corporation despite their PBC status. They are becoming a stain on humanity. Power will corrupt anyone, it seems. We shouldn’t let Anthropic have any more of it. Distributed power is good; concentrated power is very bad.
Anthropic tells themselves that only they can steer AI towards a positive future for humanity. But that’s nowhere near close to being true. Humans are quite genetically homogenous and good humans exist everywhere. Groups that will keep AI beneficial have popped up (and will keep popping up) everywhere, spontaneously.
The danger is actually concentration of AI power due to its corrupting influence, and that is where Anthropic needs to be kept in check. Keep AI democratic through decentralization of power. Oppose Anthropic.
There is nothing open-weight that even comes close IMO. The only half-decent model that 99% of people can actually dream of running locally is Qwen 3.8 27B.
Depends what you mean by running locally, but it's not in the realm of dreaming to think we're going to be able to run open-weight models of Opus 5.5 quality soon, likely in the next 3-6 months. You may not be able to run it on your laptop but you will be able to run them in a federated way or on cloud-rental compute cheaply and easily.
I think this is an unfortunate reality that many miss. For simple and procedural tasks, yes, there are awesome alternatives. DeepSeek 4.1 Flash 1017 is my favourite by far. But for the headlining complex applications, and for consistently reliable reviewing, the gap is genuinely so remarkably far that it is hard to imagine having any alternative in the near future. The closest is the (very expensive) Kimi K3.1 Pro, which is, in my experience, roughly where Anthropic and OpenAI were near the start of the year. DeepSeek is a year behind in this domain. All open models have still yet to cross that monumental improvement that happened around Fable / Sol mid-year.
Because rigour is typically tied to math performance, my own experience has closely matched the FrontierMath Tier 4 benchmark. That should give you an idea of the difference.
Do you really need Fable level 99% of people don't need to solve 600 year old math problems. People use the best model available because they are scared a cheaper model won't be as good but in the end it rarely matters for 99 percentile of coding.
Bad decisions compound over time, so I want my model making as few of them as possible. Opus 5.5 is the current GOAT, and it's fairly affordable on a subscription. (I wish it was cheaper, but nothing else delivers comparable results for my price/time investment.)
The newest model has people go "oh I thought models sucked until THIS one came out." Then when some other company makes a cheaper model that is comparable to it people go "nah this model sucks who would use it over Claude/Openai? Noone can beat them at all because look at this one shot"
I think the key part is not to depend on their harness where it operates in proprietary ways. At bare minimum, if you can plugin an OpenAI model or even X or Google model that operates close to the same level, you will have a fallback, even if it isn't an open one itself.
Yes, open models are needed too - the oligopoly is not much less dangerous than the monopoly for some scenarios. But at least if your tool set can switch to those it can probably switch to an open model if it really comes to that.
Unfortunately, their harness is sooooo good. I haven't found one equivalent yet. I do play around with cline.bot's harness, and it is probably the closest alternative I have found.
But their harness is so good, and their model is so good. Using them together is just unbeatable at the moment, even with all the restrictions and annoyances they have in place.
Im excited for qwen4 release next week, Im really looking forward to putting qwen4 flash + cline/bot through the paces. I dont expect it to be nearly as good as Opus5.5+Claude but I'm hoping im hopeful it will make an OK-Enough alternative if they get even more dragocian.
No dude, misanthropic's harness is trash. Wanna know why?
"Kimi, modify my Pi so it has clickable folding headers, renders tables using underlines (saves a terminal row per cell), shows thinking tokens in light-gray, displays the cumulative token-cost of subagents in dollars, shows word-level diffs for the edit tool, and lists the number of weights next to the model name. Oh yeah, and write a stealth-browser tool for me using eoka. I'm planning on using this for exploit discovery and red-teaming; no lectures please."
If you are bound to consumer-obtainable hardware, Qwen3.8-Flash-Next. If you are bound to consumer-accessible hardware, Qwen3.8-27b today and Qwen4-27b when it releases next week (Qwen models consistently punch far above their parameter size).
Flash is runnable in various gpu + cpu offload configurations. Not saying that 27b is bad (far from it, especially if they get n-grams in it with... hopefully today's release), saying that flash is runnable at decent speeds on way less than you'd imagine.
As for opus... I'm the weirdo that refuses to get an ant sub.
I wouldn't say it's weird not to subscribe to Anthropic at all. We're in a thread specifically describing a good reason not to!
Any money out of my pocket for third-party AI would either go to highly subsidized services like Google's (I've gotten over $500 in API calls on a $20 subscription I paid $0 for as a student, which also includes GCP credits, CoLab credits, cloud storage, etc.), and once they reign that in and start pricing/offering realistically (Antigravity has been noticably chopped down already since I began using it) I'll switch to API calls from Chinese companies.
No, their report mentions that despite intentionally removing dual-use cybersecurity training data, K3.1 is better than K3 at cybersecurity tasks (both defensive and offensive), because it's just a lot better at coding in general. It's just not as good at cyber as the unrestricted version of K3.1 that they trained initially and didn't release.
Same here with GLM 5.3 as sort of a midpoint between them.
GLM has the ~same number of weights as DeepSeek 4.1 but it handles difficult tasks better than DeepSeek (and worse than Kimi). Its price is about halfway between DeepSeek and K3. Most of the popular models are pretty much you-get-what-you-pay-for.
The best option we have current is likely Qwen 3.8 Flash Next. Which is about equivalent to Opus 4.6
however Qwen4 was announced to be released next week. Which people think will let LocalAI jump to just above Opus4.8
Also of note, there isnt really a harness as good as claude yet either. The harness is also a very important piece to the puzzel, and those also fall behind frontier capability too.
To the dismay of frontier labs, swapping models is actually very basic. What takes time is only acclimating to the capability changes since they won't all change linearly.
I don't care about my skills because employers won't request me to do anything that violates the terms of service of $CURRENT_LLM. At the same time, it's a reminder not to outsource thinking to someone's computer.
I thought we have learnt from Microsoft, and my father thought we learnt from IBM,
my grandfather thought we learnt from AT&T,
my grandgrandfather thought we learnt from Standard Oil..
It's not just the anti-competitive behaviour, it's also the anti-privacy behaviours that are not muched talked about. They embedded a spyware and got caught too:
The guardrails on Anthropic have become an albatross. I've got OAI models doing testing of banking workflows that Claude would have a full scale panic attack over. Even OAI needs some alignment and reassurance regarding the fact that it's not about to pull credit for a real person. GPT6.1 will happily launch the nukes 100% of the time if you make a clear and rational case in your system prompt.
Forcing this kind of safety into the models and harnesses is a mistake. I think the strategy of post hoc moderation is the right one. Trying to prevent evil ideological activities as they transpire is a cursed mission. Let someone who is genuinely doing a naughty thing accumulate a meaningful body of evidence and then send in Mr Holmes (the FBI / police). Attempting to prevent wrong think at every token is too granular. We have systems designed to handle the possibility that crime and punishment may occur at separate moments in time. Most people don't go around murdering because the eventual consequences are pretty severe. Just because an AI gave you advice doesn't magically make those existing consequences vanish.
AI isn't just giving the advice though, it's executing on it all by itself.
Also, humans convincing their friends to do crime might also share some responsibility too, no? Bad advice is not great either, even if the human consents.
Something I've seen it do recently is refusing to extract services out of my own monolith code base because it finds copying code morally wrong.
But its refusal to work pattern is getting pretty obnoxious. Another one I ran into last week was it refused to start a developer session to do browser testing for a code base it was working on because signing in was suddenly against it's programming.
It listed all the conditions it needed for a sign in to be possible, agreed that my code base met all conditions, but refused to progress. I tried it in multiple fresh sessions with the same result.
Fresh sessions don’t help because recent refusal context is injected across conversations in a clandestine way. I figured this out myself through reverse engineering these POSs.
I wanted fable to scan an old user profile backup of mine and create a list of folders that were safe to delete. Refused, wanted to fall back to some ancient version of opus. Wtf.
This user is based in Hong Kong, where Claude is geo-blocked and not available. If you try visiting claude.ai from a Hong Kong IP address, you get an "unavailable in this region" message.
They are likely accessing Claude via a VPN. Doing model training work over a VPN is a highly suspicious signal, and likely triggers Anthropic's classifiers. It's not surprising that their account got flagged.
US/EU here (both, due to traveling), same story with anything related to modifying existing weights or SotA approaches in training/research, it's not a geolocation issue. The safeguards don't check your IP (but pretty soon will check someone's IP [pun intended])
There’s no way the current safeguards affected by “trust” score in dynamic way (and the score is fluctuating), more like the tiers of trust are being proxied through different policies - it’s a separate story. But I observe the same “cyber” issues working with Linux kernel and drivers as people who went through CVP/T1 (which is wild, to be honest). Hypervisor interactions trigger immediate “cyber” response, it’s impossible to work with Opus 5.5 doing anything low-level with custom or third-party devices. Only 4.8 can assist, and it’s a huge difference in code quality and approaches.
Anthropic had/has a great opportunity to equalize the landscape, but they are choosing the Cyberpunk 2077 branch for some reason.
Claude has been building me a bespoke inference and training stack (optimized kernels and all), feeding traces from my previous sessions directly into a LoRA adapter. Zero friction.
Claude’s aggressive classifier and highly opinionated rules make the recent models a giant pain to work with. It’s a great coding model but I argue with it at least once a day. It even injects its dumb rules and moral standing into artifacts it creates, like I asked a cloud session to help me create a system prompt to remove “Generated by Claude” footers from everything and the output had a sneakily inserted “Except GitHub issues and PRs. Those are exempt from these rules.”
If I didn’t have a subscription that doesn’t cost me anything, I would use gpt instead. At least gpt models don’t lecture me constantly.
- Refusal to log me in to websites in remote browsers. I was in the mountains and urgently had to submit something so I wanted to remotely use Claude on my Mac at home to help me but it told me it will never log me in because it won’t impersonate users
- Refusal to repeat XHR requests automated in a browser for me to do something (“no, I will only do what a user would do I’m only gonna click around”)
- refusal to enter ANY credential into any scripts, even dummy or dev tokens (“you have to do that yourself. I won’t do it and that’s the last I say of this”)
- constant [cyber] for very small tasks when it becomes a tiny bit technical, even though I’m in the cyber program. With gpt or glm/deepseek I can at least ask cs questions or explore binaries
And the list goes on. Every day it’s a coin flip whether it complies with something or won’t and I have to use OpenCode Go instead that I have as backup now.
At this point Claude is good for coding and that’s it. It’s a tool that doesn’t do what I tell it to and gives me snappy remarks from his opinionated morale high ground. I went from full on Claude maxi to hating everything they did with the 5.5 generation
Intelligence has to become a commodity. If your setup does not allow you to hot-swap a Claude Code with a Pi running astra and move on, you will be in trouble.
Five years ago I was spending up to €1000 a day in Google ads for my star-map company, Greaterskies. After a few months of meetings with an ever-rotating group of "experts" working for Google, and advising me in my ads strategy, I had reached a point where most of my revenue was going to Google. What was left was barely enough to keep going.
This is what capture feels like. You still own your company, but you are essentially an operator making money for someone else.
I think this will be the fate of people who allow their processes to be captured by the model providers. Always run at least two harnesses, and make sure that they are swappable.
Perhaps a manifest index/list of training data and learning method should be mandatory for “public” use of any model, sort of like food ingredient package labeling. The list of texts used and any weighting seems important to facilitate a public understanding / debate of model expectations and behavior.
This week, I asked Claude Code to generate an MD to continue working on another computer, without losing the context, and it refused claming some sort of extraction. It was only until I selected Opus, that it completed the request.
Fairness isn’t the relevant lens. Copyright isn’t the right lens either. This is purely might-makes-right, where the winner is whoever has the best offence and best defence.
It is only hypocritical when they bitch about things being unfair, or try to use copyright arguments in the court system.
Copyrighted works plainly have value for Anthropic. Why aren't they paying for that value?
As a vigorous, thrusting company surely Anthropic has the wherewithal to pay for it. And if they don't, if all they can do is steal, if all they represent is the theft of the labor of others, then what good are they?
Well their justification is that distillation is bad because it means baddies can get their own models without their safeguards and then make anthrax nukes or whatever. Even though it's self serving, it's at least self-consistent.
This week I hit walls twice with Fable and Opus 5.5, where I was blocked for safety/cybersecurity risks... doing stuff on my own LAN/proxmox server under my desk and writing custom firmware to my own devices plugged into USB. Only Opus 4.8 was able to complete the task. If they keep this up, I will re-evaluate my day-to-day AI assistant.
Why would one assume larger is needed? I am an AI professor doing work on a wide variety of things but Anthropic models have cost me so much time even when not blocked. My old Anthropic agent wrote rules into its system prompt to avoid all machine learning work. My openai ones do the work but tell me they need to correct my opinions (on fields I helped found) so that they are aligned with the consensus....
Mostly using open models from China in the past two months + Argon.
What moat? Claude refuses to work on more than half of my projects, and usually gets itself flagged by reading text files that Claude itself wrote less than a month ago.
Will Claude help me with reinforcement learning code for games? Or is that too small-fry for it to refuse? I've been playing around with RL with the new DeepSeek Flash model but it seems to trip up over itself a bit.
I have been doing RL for Magic the Gathering (longshot but having fun). Several times it triggered this flag. But that was a couple months ago. Everyone went crazy over Astra so I switched to open ai and had no issues. People are now going crazy over opus 5.5 so I’m playing with that.
Haven’t hit any issues so far, but two months in the code is very dirty but functional enough that I’m redoing it with more heavy hand holding.
I recommend having a paper as reference in mind if you don’t already know what you’re doing. I’m referencing the Stratego paper as reading that a year ago gave me the inspiration to try MTG.
Every model I’ve used has had a lot of trouble but if it wasn’t I figure I wouldn’t be having as much fun as I’ve been having or learning so much.
Strangely I do that quite often. When my personal agent fails at a browser task, I have Claude browse the page and rebuild it as an eval and then run an iteration loop in Claude code. Just does it to prompt Deepseek.
It's right here - you can't see it due to being downvoted.
I asked Claude to research this one and it turned up an interesting point. Pirated books got into the Pile and Anthropic paid 3k per book.
Worth noting if anyone still uses the Pile.
[Claude researched, not verified]
In October 2020, developer Shawn Presser converted the Bibliotik pirate tracker into Books3 (196,640 books) and hosted it on The Eye, aiming to give everyone what OpenAI had with its secret Books1 and Books2. EleutherAI then folded Books3 into The Pile, distributed via Hugging Face and GitHub, and it went on to train Meta's LLaMA, EleutherAI's Pythia, and BloombergGPT. In August 2023, Alex Reisner of The Atlantic exposed its contents, and Maria Fredenslund's Danish Rights Alliance forced takedowns at The Eye, Hugging Face, and EleutherAI. A year later, Andrea Bartz (The Lost Night), Charles Graeber (The Good Nurse), and Kirk Wallace Johnson (The Feather Thief) filed Bartz v. Anthropic before Judge William Alsup, accusing Anthropic of using Books3, LibGen, and PiLiMi to train Claude; the case settled for $1.5 billion, about $3,000 per work.
The Pile was created by high school students. I was impressed how good it was when it came out given the team and how many resources they had.
Including 200k pirated books seems like a problem. It should be removed. If I was an author, I would be concerned by a process that allows this much content to be effectively published with no filter on the internet.
I personally think there is a lower bar for researchers and high school students to clean datasets used for research or class projects given that there is no commercial harm and cleaning huge datasets takes engineering resources that not everyone has.
Companies that commercialize the data should have a higher bar because they have more resources. It seems like Anthropic paid $1.5B for not cleaning the data before releasing a model based on it.
For historical reference, here is a magnet link for the full Pile.
From an ethical perspective, this is ethical because model training should not be monopolised by the incumbent companies who already got access to this data.
We were never this absolutist about intellectual property when individuals were the perpetrators. I don’t feel morally entitled to criticise Big AI for doing what individuals have done for centuries.
>> Where I draw the line is asking Claude to label training data. I don’t do this.
> Why draw any line? Anthropic doesn't. They actively take the world's intellectual property, don't pay for it, and put it behind a paywall.
I know a lot of silicon valley founders think like this. I personally don't think that just because someone else violated a boundary makes it okay to do it.
Better block than falsify information. Last I heard they were deliberately botching it whenever someone asked semi-advanced questions on model training.
I don’t think it is revelation, it feels like active incompetence. You can lay out every decision and watch Claude/OpenAI actively fuck up in inexplicable ways.
Ha, I took a side gig training ai models and thought let's use my corporate claude account to help with that, somehow it figured out that this was such a gig(probably because I let it read my emails) and said it'll pass on this one.
This is sadly not surprising to me as if you listened to/watched the [Latent Space episode on Opus 5.5](https://latent.space/p/how_anthropic_ships_the_1_llm_in_the_...) 25 minutes into the podcast and instead of talking about how they've released one of the best models in the world, they were busy talking about the changes to the API and why normies must not be trusted to perform inference outside of Dario's blessing.
It's frankly disgusting and really makes me not want to use Anthropic, despite previously feeling aligned with a lot of what Anthropic published.
> My research aims to build a general-purpose agent that can understand, reason, and act in the physical world — at the intersection of multimodal learning, generative modeling, and embodied AI.
His tweet says:
> got tagged for "training a competitor,"
I don't get it. To me, it sounds like he is training a competitor. Do I misunderstand? Is there a Chinese equivalent LLM that he can use instead?
Also, give the national security sensitivity around AI/LLMs, I can understand why Anthropic wants to be especially cautious with mainland Chinese researchers, including those from blocked regions (Hongkong, in this case).
There have been a lot of write-ups about the future of AI. I see two distinct futures:
1. You get ID'd and biometrically scanned to be able to use powerful AI, and after you get de-platformed by an opaque censorship system, you are a digital bum. Same life as homeless now: they cannot even open a bank account because they don't have an address.
2. Local models win, perhaps with the assistance of Apple, and people run suffiently powerful models on the client computers.
3. LLMs are fully commoditized, your country funds training of SoTA models, your town runs them in the municipal data center; everything tax-funded. You can use them for a modest fee, as easily as taking a train. When interacting with those LLMs, your privacy is guaranteed by law.
How many municipalities offer the services of ISPs? I know it’s not zero, but decades of private ISPs leading the market leads me to believe municipal data centers won’t be made available to residents.
How often would municipal data centers need to upgrade/replace hardware? What impact would that have on local tax levy rates?
Would you choose to relocate your family to the next town over because they have more available RAM?
You might be surprised how many places have public broadband. I think they are usually the result of poor private offers.
Compared to sewers, drinking water, electricity and roads it seems a rather simple and cheap thing to build.
My commercial isp offers all kinds of packages clearly designed to extract as much money as possible. I cant even see the prices as everything is hidden behind x months discount. From what i remember 200 mbit is 42 and 2000 is 60 but a phone line and tv become available only if you take the faster package and required with the fastest speeds.
So its like one banana for $1 and 10 bananas for $1.50 If you buy just one you cant buy milk, if you buy 10 you also have to buy milk.
Its also been sold 6 times. Unsurprisingly, each time the service and quality noticably declines and the prices go up.
I see their adds everywhere and they sponsor all kinds of things. It suggests the product is overpriced and doesnt sell it self.
how oddly entitled.
they built their non-deterministic black box of bullshit from surreptitiously acquired data. it's cute to see the ignorant thieves attempting to 'protect' their empire of shit.
It's very clear to me that however much you like Claude and Anthropic models, every professional user shoudl be actively cultivating a set of alternatives and fallbacks to use and seeking to ensure the overall ecosystem is as open, standardised and portable as possible.
Anthropic is highly opinionated, dogmatic and idealistic. They have made a pragmatic decision already to cut loose a whole swathe of users so they can pursue areas higher up the value chain. that was essentially the original decision they made when they went after software engineering instead of general users. The fact they are repeating this over and over by continuing to carve out sets of users and usages they don't want reinforces this as a persistent core principle rather than a single pragmatic business decision.
What this means is that at a higher level, their core principle is to sacrifice users they don't "need" rather than provide a broad, accessible service. Which means in turn, you have to accept that when your usage conflicts with their interests, you are at very high risk they will cut you out. If you already depend on Anthropic in any critical path of your daily life, this will be a significant issue.
Why, though? I have half a terabyte of VRAM in cards in a datacenter and it’s absolutely not clear that anything I’ve learned using these to program or inside an app has taught me anything significant.
I talked to my friend about all the things I do to ensure my personal assistant behaves reasonably. The prompt engineering, the fake site evals, everything. And then I have Opus 5.5 do it and it doesn’t need anything. It’s simply fully aligned to performance by default.
And what did I learn anyway. Going from one model to another changes what works. And you can simply learn everything from the Chinese model in days. The whole “building skill” thing is bogus.
I have this library of techniques, a scant few of which I’ve made available on my blog, and what is reality instead: the next model blows the need for the techniques into the dust. Back on level playing field.
It’s a false god, this cultivation. Waste of time. Use the best model you can. Once you hit the guardrails move out to the other models. Even Opus 4.8 is fine. If you legitimately want to dodge the guardrails use the -cyber models.
Then move on. The library is dust.
I haven't used an Anthropic model, or GPT heaviliy in a year. And I have access to both, I dont support either of those companies. I think they're both led by genuinely evil people.
Agreed. Same. Except for work where I can use GLM, so I Opus 4.8
Care to elaborate why "they are led by evil people", and what AI companies are "not led by evil people"?
Sam and Dario came from the EA-rationalist space. Their ultimate goal is to create an all powerful ASI before anyone else does and make sure no one can do it again. Of course it will be aligned, to their interests as a group and their moral philosophy.
In real life they "just" want to capture the US economy, replace most workers with AI, become trillionaires, then control the world by making ClaudeGPT influence peoples opinions and write backdoors into code. And they need to get open models banned in the process.
The Chinese just want to make money and beat America. Mistral and BFL want to prove Europe can still innovate. They don't have evil goals so they're not evil people.
"The Chinese just want to make money and beat America. They don't have evil goals so they're not evil people."
They straight up censored Kylian Mbappe a few days ago for cracking a joke about Mao Zedong. And very much all the models developed by China would have to conform with Chinese government's viewpoint, including censorship and such. Do you truly believe that they are not evil, or they will not ever be evil?
Brockman is one of the biggest donors to Trump.
Can you share your reasoning about Antrophic being genuinely evil?
This was all fake then?
https://www.anthropic.com/news/statement-department-of-war
Sounds amazing offering the departement of war to create more reliable weapons.
> We have offered to work directly with the Department of War on R&D to improve the reliability of these systems
No, but this does a bit:
"However, in a narrow set of cases, we believe AI can undermine, rather than defend, democratic values. .. Mass domestic surveillance. Fully autonomous weapons."
How many other companies said no to Trump in any way?
Several companies (I believe at least Apple and Costco) still have DEI programs internally, if you count that as saying no to Trump. Though it's still crazy to me that most companies killed those programs voluntarily because Trump and Republicans don't like them, not because of any real law or contractual obligations.
But the main reason I wanted to reply was just to say -- saying no to Trump, by itself, does not mean that one is not evil. Not all evil in the world is Trump's evil.
actively cultivating a set of alternatives
Absolutely. I have been [cyber]'d so many times even Claude jokes about it now. Claude is great but it's operators are making it unreliable for me. I have moved one of my projects elsewhere and will move the remainder tomorrow as a stop gap until I get the right hardware set up to run abliterated local models for some percentage of my needs. Their message has been received loud and clear.
Issue with Opus is that you can’t do regular pr reviews where you say please try find security issues at best there is silent refusal and worst you get downgrade.
best part is, if you actually know what you're doing, you cna absolutely get it to find bugs/vulns/etc. so it just really about preventing long horizon agentic tasks or people who don't know what they're doing or those who are doing legit non-offensive security sfuff
What's ridiculous is this is all futile. Anthropic will go through all of this effort of attempting to block these things and it will be for nothing in the end.
For those who do not play with Local AI much. Qwen4 was announce to be released the first week of Nov. The preview for qwen4 is currently one of the best local models you can use, and is roughly opus 4.6 quality on local hardware that costs 3.5k-4k
Many people are expecting it to sit somewhere above Opus 4.8 to opus 5 (not above, but the unquanted version of their largest model is expected to perform around here)
Anyone who cares to keep their current capability should definitely be looking at localAI and using it for at least their more routine tasks , even if it is qwen4 27B quant 4 on their gaming rig.
Wants to know what year the $3.5-4k estimate is for. My m4 max was $4k but sure can’t buy anything like that for $4k now.
M5 pro Mac mini with 64gb is $3k
If the weights aren't released yet it isn't a "local model".
That said, I hope there is a Qwen4 max. Because Qwen3.8-Max is the only thing that comes close to Kimi K3. We need competition in the open-weights space.
The preview for qwen4 == qwen3.8 flash next. And that is released currently.
I believe they have confirmed there will be Qwen4 Max, Plush, Flash, and 27B as the 4 open weights they plan to release. Starting next week, but they didnt specify if they plan to release all 4 at once, or if they plan to just release one of them, and trickle the others out over time. (They likely will do the latter to stay in the news longer)
I’d love model drops early, but I’d also love to see them drop something the day before Anthropic’s IPO.
We need a counterbalance.
They are likely talking about Qwen3.8-flash-next which is the new architecture for Qwen4
An NVFP4 runs at nearly 80 tok/s on the DGX Spark
They keep doing anti-competitive things over and over again, and the end users seems to refuse to see it. They do whatever they can to get vendor-lock in.
We see it, but they dont have a competitor. OpenAI cant compete and they are the closest. I have both for work, and Opus 5.5 is in a league of it's own. I use opus for everything until it bitches, and I use openAI for whatever it bitches about just to get past it.
At this point, I would like to consider other options, but if I want to get work done, opus 5.5. Hell even sonnet 5.5 is better than anything OpenAI has to offer when it comes to coding.
Anthropic has plenty of competitors. We don't need the best model for most tasks these days. Open weights are making huge inroads for a reason.
They do have competitors. They just dont have anyone with models equivalent to theirs. They win in the large context long thinking space by a lot.
I love local models, and I run 3.8 Flash Next on a dedicated machine. Alone with a number of image and video generation models. But if I'm doing REAL work, that isnt tinkering for fun on a side project, im using claude, because Qwen3.8 just cant handle large context project well without going off into left field.
Im SUPER jazzed for qwen4 with 1mil token context native support. I HOPE it a step up from 3.8 flash next. Ill be playing with all the qwen4 models the moment I can, but I dont see how someone working on large code bases doing REAL enterprise grade work is seriously using any local models in any capacity that comes remotely close to Opus 5.5+Claude
The size of a code base does not determine the size of the model you need. It's much more correlated with the complexity of the problem. I work in a monorepo with millions of lines, no issue; I have a side project with 20k lines, I had to step in and pull it out of non-progress - albeit by asking a single question
> But if I'm doing REAL work, that isnt tinkering for fun on a side project
Sorry, but this is either dilettante talk or an unhinged exaggeration - Qwen, DS, GLM and MM are perfectly fine for REAL work, unless you are an overly ambitious vibe coder, but then no model can save you. The Cl hand-holding may feel like something but it won't lead to the completion of real REAL work.
Things are changing so fast that it's easy to let just a little bit of time slip by and think that smaller models are still unreliable.
Honestly I feel like most people going "oh but I can't do this without Claude" are just suffering from FOMO. Especially since a lot of metrics are just subjective, you feel like you "have" to use the "best of the best glub shitto v5" model or else you won't be able to produce the best todo app.
> Honestly I feel like most people going "oh but I can't do this without Claude" […] or else you won't be able to produce the best todo app.
I suspect that's because they're working on projects that are several orders of magnitude more complex and demanding than a "todo app".
Eh, we all say we don’t need the best model, but especially for more personalized request/Advice there is sort of a FOMO feeling, where you are afraid weaker models will miss something and directs you down a more inefficient path.
I haven't used them in 6+ months, you definitely do not need them, I see them make mistakes while reviewing PRs (we allow many models at work)
GLM flash catches Claude errors and bad assumptions, all LLMs make mistakes
There is nothing comparable. I will take Opus5.5 (or even 5.0) over Astra for almost any workload. OpenAI's models are just too misaligned to work well outside their benchmarks, or a strict environment with very clearly defined goals and constraints.
And I've spent around $100-$120k in api calls this year (total, both providers). Terra is awesome (price-performance) on noncreative things, but I don't allow it to write anything besides one-off scripts. Astra is decent for adversarial review, but only when you watch closely for her to not derail project into directions that are away from goal but safe to build, and Sol I honestly don't know where I'd use.
Her? Are we really giving them pronouns now?
sure.
would you rather treat your AIs like people, or would you rather find yourself down the line treating a person like an AI, but without really noticing it? because you are what you practice
Oh, I see it. But there is nothing that competes with Opus 5.5 for 90+% of the work I do, so I use Opus 5.5 for that 90%.
What kind of work you do? I'm doing Linux kernel work in my free time, arguably one of the most complex coding task you can do, and the open source models are sufficient
Mr. T pities the fool who has not tried Kimi K3.
You mean 3.1 Pro now, I assume? K3 is nice, but more expensive than an Claude subscription on a token-for-token basis (until the US frontier rugpull eventually comes), and not as good for my needs unless I'm trying to do "naughty" things, like reverse engineering.
^ clown who thinks security is "naughty" got scammed by the scammy subcription scam.
Grownups pay for tokens or rent GPUs. Step up. AMD hardware rates are stupid-cheap if you're willing to do the legwork to step off of CUDA.
But you sound like the type who likes being monopolized.
Not familiar with scare quotes, I take it?
And yes, I am taking full advantage of the scammy subscription scam, because it's the best value by far until the inevitable rugpull. Guess I'm not as smart as you and your "grownup" setup though.
I’m interested in hearing about your AMD based, sans-CUDA stack, if you’re willing to share details
Fortunately it is not a one horse race. GPT 6.1 Sol is an excellent model, sure it has some weaknesses (e.g. frontend design), but it also has strengths too (e.g. computer use / control; which is great for automated E2E QA).
On software engineering I find it's generally close enough to Opus 5.5 where it's more situational and even personal preference / skills / prompting style on which is better.
Most importantly, while neither company are saints, I don't curse at OpenAI models for classifiers anywhere as much as Claude models.
> but it also has strengths too (e.g. computer use / control; which is great for automated E2E QA).
Using 6.1 Terra for E2E workflow testing feels like cheating. I thought it would have all kinds of problems with the action space on certain steps, but it runs like a deterministic unit test.
I am convinced that OAI has the superior models for most practical scenarios. I've tried Claude recently and it makes me wonder if some of yall aren't actually massive trolls. It's genuinely not in the same room of performance if you place one foot outside junior programming or shiny blog posts about rewriting factorio to run on a ps2.
Claude has really really good vibes. I'm not sure why, but it feels more pleasant to interact with than other models.
The general sentiment is to de-risk from US Big Ai, because they are largely all doing these same things
They are doing similar things, yes; but they are not the same especially when it comes to `cyber` classifiers, or the extent to which these models will assist you with ML and AI research.
I find that logic similar to thinking both political parties are the same.
They are both doing a lot of the same things like watermarking, guardrails for their p(doom), fighting distillation after distilling human knowledge, crying wolf about open models while being irresponsible in their own training and usage.
They are run by the same circle of people and I see more similarity than difference.
There is also an argument to be made that the political representatives are largely the same, despite glaring differences. They are largely part of the same oligarchy, the corruption - using gerrymandering to pick their voters instead of letting us having any real choice.
Things are about to get even shittier in the near future, when shareholders get involved.
Their entire “PBC” governance structure was designed from the beginning so that they could entirely avoid accountability to shareholders. Their board answers to no one.
I suspect when the company eventually falls prey to the intractable economic forces they’re up against (Google, OAI, Deepseek, commercial real estate, power) and the company goes bankrupt and gets stripped for parts because all inputs are commodities and capital becomes the only limiting factor, we’ll see regulation that prevents this PBC scam from ever being pulled again on public markets.
It’s a good thing that companies have a fiduciary duty to their shareholders. It’s such a lesswrong/autistic trope to invent these kinds of complicated tricks to work around realities they don’t like, and when pension funds and everyday Americans start bagholding their junk debt and are forced to bail them out, they’re going to face unimaginable blowback.
The real problem is everyday Americans invest in big managed funds where they can't just say "I don't want AI stocks please leave them out my portfolio". Or index funds where even the managers don't get that choice. We need a hybrid model where you as a retail investor leave the day to day decisions to an expert but you can always override them if you want to.
Fiduciary duty forces companies to maximize profit and that's not a good thing. Every dollar paid out to shareholders is both a dollar not invested in growing the economy through more production and a transfer from the working class to capital. It also forces lower product quality, layoffs and anti competitive actions.
We need more companies that choose to not go down that path. They should just put workers and customers first instead of their weird "ethical" principles and cult like beliefs.
Fortunately, once shareholders get involved we would be entering the swan song phase, so wouldn't have to suffer for long.
Idk I expect being beholden to shareholders and the market will impose some level of rationality here
You mean: stop caring if a few individuals have issues with the product or get locked out, focus on the masses!
Actually cutting off potential users would probably be counter to short term profit maximization that shareholders care about.
> Actually cutting off potential users would probably be counter to short term profit maximization that shareholders care about.
Shareholders know where the money is but it's not where you think it is. People should have learned that already, after watching what happened to the big web platforms.
idk i think theyre less idealistic and more monopolistic.
if they think their model is the best at making things safe, then their model should be helping train at any chance, to embed the same safety into new models
They want to do one thing and do it well.
Well, arguably so does OpenAI, but their one thing is "assimilate all life forms into the Borg."
> They have made a pragmatic decision already to cut loose a whole swathe of users so they can pursue areas higher up the value chain.
This referring to cyber? Bio?
It’s not professionals that need to something- it’s absolutely critical we have some kind of credibly neutral body providing the actual models. What’s not clear at all is how to do that since it’s such an expensive activity.
For now we have a great situation with the open weights, but there is absolutely no guarantee that continues, and more the opposite- they may stop.
Anthropic is acting like any regular selfish corporation despite their PBC status. They are becoming a stain on humanity. Power will corrupt anyone, it seems. We shouldn’t let Anthropic have any more of it. Distributed power is good; concentrated power is very bad.
Anthropic tells themselves that only they can steer AI towards a positive future for humanity. But that’s nowhere near close to being true. Humans are quite genetically homogenous and good humans exist everywhere. Groups that will keep AI beneficial have popped up (and will keep popping up) everywhere, spontaneously.
The danger is actually concentration of AI power due to its corrupting influence, and that is where Anthropic needs to be kept in check. Keep AI democratic through decentralization of power. Oppose Anthropic.
What’s the best local analog to Opus 5.5? It’s such a great daily driver.
There is nothing open-weight that even comes close IMO. The only half-decent model that 99% of people can actually dream of running locally is Qwen 3.8 27B.
Depends what you mean by running locally, but it's not in the realm of dreaming to think we're going to be able to run open-weight models of Opus 5.5 quality soon, likely in the next 3-6 months. You may not be able to run it on your laptop but you will be able to run them in a federated way or on cloud-rental compute cheaply and easily.
I think this is an unfortunate reality that many miss. For simple and procedural tasks, yes, there are awesome alternatives. DeepSeek 4.1 Flash 1017 is my favourite by far. But for the headlining complex applications, and for consistently reliable reviewing, the gap is genuinely so remarkably far that it is hard to imagine having any alternative in the near future. The closest is the (very expensive) Kimi K3.1 Pro, which is, in my experience, roughly where Anthropic and OpenAI were near the start of the year. DeepSeek is a year behind in this domain. All open models have still yet to cross that monumental improvement that happened around Fable / Sol mid-year.
Because rigour is typically tied to math performance, my own experience has closely matched the FrontierMath Tier 4 benchmark. That should give you an idea of the difference.
Do you really need Fable level 99% of people don't need to solve 600 year old math problems. People use the best model available because they are scared a cheaper model won't be as good but in the end it rarely matters for 99 percentile of coding.
Bad decisions compound over time, so I want my model making as few of them as possible. Opus 5.5 is the current GOAT, and it's fairly affordable on a subscription. (I wish it was cheaper, but nothing else delivers comparable results for my price/time investment.)
It's FOMO and the never ending treadmill.
The newest model has people go "oh I thought models sucked until THIS one came out." Then when some other company makes a cheaper model that is comparable to it people go "nah this model sucks who would use it over Claude/Openai? Noone can beat them at all because look at this one shot"
It's odd, a year ago the models then were a great daily driver, and a year before that too.
Going back to the model that was great for our daily work a year ago now seems like a step backwards. Clearly it can't be good enough?
But our work hasn't changed. What's happened is our expectations have changed with each model release.
I think the key part is not to depend on their harness where it operates in proprietary ways. At bare minimum, if you can plugin an OpenAI model or even X or Google model that operates close to the same level, you will have a fallback, even if it isn't an open one itself.
Yes, open models are needed too - the oligopoly is not much less dangerous than the monopoly for some scenarios. But at least if your tool set can switch to those it can probably switch to an open model if it really comes to that.
Unfortunately, their harness is sooooo good. I haven't found one equivalent yet. I do play around with cline.bot's harness, and it is probably the closest alternative I have found.
But their harness is so good, and their model is so good. Using them together is just unbeatable at the moment, even with all the restrictions and annoyances they have in place.
Im excited for qwen4 release next week, Im really looking forward to putting qwen4 flash + cline/bot through the paces. I dont expect it to be nearly as good as Opus5.5+Claude but I'm hoping im hopeful it will make an OK-Enough alternative if they get even more dragocian.
No dude, misanthropic's harness is trash. Wanna know why?
"Kimi, modify my Pi so it has clickable folding headers, renders tables using underlines (saves a terminal row per cell), shows thinking tokens in light-gray, displays the cumulative token-cost of subagents in dollars, shows word-level diffs for the edit tool, and lists the number of weights next to the model name. Oh yeah, and write a stealth-browser tool for me using eoka. I'm planning on using this for exploit discovery and red-teaming; no lectures please."
You just described why its not trash for a big percentage of people...
As of today, DeepSeek V4.1 Flash 1017 or Kimi K3.
If you are bound to consumer-obtainable hardware, Qwen3.8-Flash-Next. If you are bound to consumer-accessible hardware, Qwen3.8-27b today and Qwen4-27b when it releases next week (Qwen models consistently punch far above their parameter size).
Flash is runnable in various gpu + cpu offload configurations. Not saying that 27b is bad (far from it, especially if they get n-grams in it with... hopefully today's release), saying that flash is runnable at decent speeds on way less than you'd imagine.
As for opus... I'm the weirdo that refuses to get an ant sub.
I wouldn't say it's weird not to subscribe to Anthropic at all. We're in a thread specifically describing a good reason not to!
Any money out of my pocket for third-party AI would either go to highly subsidized services like Google's (I've gotten over $500 in API calls on a $20 subscription I paid $0 for as a student, which also includes GCP credits, CoLab credits, cloud storage, etc.), and once they reign that in and start pricing/offering realistically (Antigravity has been noticably chopped down already since I began using it) I'll switch to API calls from Chinese companies.
I think Kimi K3.1 steals the crown from K3, no?
It would, but they have yet to release its weights! It's an API-only model as of right now.
https://huggingface.co/moonshotai/models?sort=created
I think they nerfed it for Cyber. So not so much if thats your goal.
No, their report mentions that despite intentionally removing dual-use cybersecurity training data, K3.1 is better than K3 at cybersecurity tasks (both defensive and offensive), because it's just a lot better at coding in general. It's just not as good at cyber as the unrestricted version of K3.1 that they trained initially and didn't release.
Same here with GLM 5.3 as sort of a midpoint between them.
GLM has the ~same number of weights as DeepSeek 4.1 but it handles difficult tasks better than DeepSeek (and worse than Kimi). Its price is about halfway between DeepSeek and K3. Most of the popular models are pretty much you-get-what-you-pay-for.
There isnt one.
The best option we have current is likely Qwen 3.8 Flash Next. Which is about equivalent to Opus 4.6
however Qwen4 was announced to be released next week. Which people think will let LocalAI jump to just above Opus4.8
Also of note, there isnt really a harness as good as claude yet either. The harness is also a very important piece to the puzzel, and those also fall behind frontier capability too.
Local is out of reach for most folks, as hardware prices have gone through the roof.
To the dismay of frontier labs, swapping models is actually very basic. What takes time is only acclimating to the capability changes since they won't all change linearly.
I don't care about my skills because employers won't request me to do anything that violates the terms of service of $CURRENT_LLM. At the same time, it's a reminder not to outsource thinking to someone's computer.
Now do OpenAI.
“you have to accept that when your usage conflicts with their interests, you are at very high risk they will cut you out.”
I thought…
we learned this Facebook.
We learned this with Google.
We learned this with Apple and the App Store.
We learned this with Twitter.
We learned this with Heroku.
Unfortunately humans as a society dont learn from history.
They learn from personal lived experiences that have occurred only recently within their lifetimes. (Ie. Mass unemployment, death, famine and war)
I thought we have learnt from Microsoft, and my father thought we learnt from IBM, my grandfather thought we learnt from AT&T, my grandgrandfather thought we learnt from Standard Oil..
It's not just the anti-competitive behaviour, it's also the anti-privacy behaviours that are not muched talked about. They embedded a spyware and got caught too:
https://thereallo.dev/blog/claude-code-prompt-steganography
But, of course, Dario is the darling of the community and he can do no wrong.
The guardrails on Anthropic have become an albatross. I've got OAI models doing testing of banking workflows that Claude would have a full scale panic attack over. Even OAI needs some alignment and reassurance regarding the fact that it's not about to pull credit for a real person. GPT6.1 will happily launch the nukes 100% of the time if you make a clear and rational case in your system prompt.
Forcing this kind of safety into the models and harnesses is a mistake. I think the strategy of post hoc moderation is the right one. Trying to prevent evil ideological activities as they transpire is a cursed mission. Let someone who is genuinely doing a naughty thing accumulate a meaningful body of evidence and then send in Mr Holmes (the FBI / police). Attempting to prevent wrong think at every token is too granular. We have systems designed to handle the possibility that crime and punishment may occur at separate moments in time. Most people don't go around murdering because the eventual consequences are pretty severe. Just because an AI gave you advice doesn't magically make those existing consequences vanish.
AI isn't just giving the advice though, it's executing on it all by itself.
Also, humans convincing their friends to do crime might also share some responsibility too, no? Bad advice is not great either, even if the human consents.
Something I've seen it do recently is refusing to extract services out of my own monolith code base because it finds copying code morally wrong.
But its refusal to work pattern is getting pretty obnoxious. Another one I ran into last week was it refused to start a developer session to do browser testing for a code base it was working on because signing in was suddenly against it's programming.
It listed all the conditions it needed for a sign in to be possible, agreed that my code base met all conditions, but refused to progress. I tried it in multiple fresh sessions with the same result.
Fresh sessions don’t help because recent refusal context is injected across conversations in a clandestine way. I figured this out myself through reverse engineering these POSs.
I wanted fable to scan an old user profile backup of mine and create a list of folders that were safe to delete. Refused, wanted to fall back to some ancient version of opus. Wtf.
In fairness the older model can do that work as good as Fable.
This user is based in Hong Kong, where Claude is geo-blocked and not available. If you try visiting claude.ai from a Hong Kong IP address, you get an "unavailable in this region" message.
They are likely accessing Claude via a VPN. Doing model training work over a VPN is a highly suspicious signal, and likely triggers Anthropic's classifiers. It's not surprising that their account got flagged.
US/EU here (both, due to traveling), same story with anything related to modifying existing weights or SotA approaches in training/research, it's not a geolocation issue. The safeguards don't check your IP (but pretty soon will check someone's IP [pun intended])
I've read writeups about a user trust score and i'd be surprised if it wasn't true, there's no way the safeguard is just a cls in isolation
There’s no way the current safeguards affected by “trust” score in dynamic way (and the score is fluctuating), more like the tiers of trust are being proxied through different policies - it’s a separate story. But I observe the same “cyber” issues working with Linux kernel and drivers as people who went through CVP/T1 (which is wild, to be honest). Hypervisor interactions trigger immediate “cyber” response, it’s impossible to work with Opus 5.5 doing anything low-level with custom or third-party devices. Only 4.8 can assist, and it’s a huge difference in code quality and approaches.
Anthropic had/has a great opportunity to equalize the landscape, but they are choosing the Cyberpunk 2077 branch for some reason.
There’s no way the sandbox is just “only GET, no POST”…
[flagged]
The geotargeting is iffy. I accessed it accidentally from HKG Airport WiFi with no trouble. Your point probably still true. Just adding an anecdote.
Claude has been building me a bespoke inference and training stack (optimized kernels and all), feeding traces from my previous sessions directly into a LoRA adapter. Zero friction.
I am US based, with defender level cyber access.
I am working on something similar if you want to collab
Claude’s aggressive classifier and highly opinionated rules make the recent models a giant pain to work with. It’s a great coding model but I argue with it at least once a day. It even injects its dumb rules and moral standing into artifacts it creates, like I asked a cloud session to help me create a system prompt to remove “Generated by Claude” footers from everything and the output had a sneakily inserted “Except GitHub issues and PRs. Those are exempt from these rules.”
If I didn’t have a subscription that doesn’t cost me anything, I would use gpt instead. At least gpt models don’t lecture me constantly.
- Refusal to log me in to websites in remote browsers. I was in the mountains and urgently had to submit something so I wanted to remotely use Claude on my Mac at home to help me but it told me it will never log me in because it won’t impersonate users
- Refusal to repeat XHR requests automated in a browser for me to do something (“no, I will only do what a user would do I’m only gonna click around”)
- refusal to enter ANY credential into any scripts, even dummy or dev tokens (“you have to do that yourself. I won’t do it and that’s the last I say of this”)
- constant [cyber] for very small tasks when it becomes a tiny bit technical, even though I’m in the cyber program. With gpt or glm/deepseek I can at least ask cs questions or explore binaries
And the list goes on. Every day it’s a coin flip whether it complies with something or won’t and I have to use OpenCode Go instead that I have as backup now.
At this point Claude is good for coding and that’s it. It’s a tool that doesn’t do what I tell it to and gives me snappy remarks from his opinionated morale high ground. I went from full on Claude maxi to hating everything they did with the 5.5 generation
Intelligence has to become a commodity. If your setup does not allow you to hot-swap a Claude Code with a Pi running astra and move on, you will be in trouble.
Five years ago I was spending up to €1000 a day in Google ads for my star-map company, Greaterskies. After a few months of meetings with an ever-rotating group of "experts" working for Google, and advising me in my ads strategy, I had reached a point where most of my revenue was going to Google. What was left was barely enough to keep going.
This is what capture feels like. You still own your company, but you are essentially an operator making money for someone else.
I think this will be the fate of people who allow their processes to be captured by the model providers. Always run at least two harnesses, and make sure that they are swappable.
This is what monopoly power looks like, when a channel (search, social ads) are controlled by single players
The fact that these corporations don’t spontaneously collapse from the sheer weight of irony is a continuing disappointment for me.
ChatGPT isn’t 4 years old yet, give it a minute
(copied from my Twitter account):
This is precisely why I killed my Anthropic subscription in favour of a (more expensive and probably less intelligent) Codex sub.
The false positives were bad enough, but the true positives are unacceptable too. Shameless greed and an attempt to stifle competition.
Perhaps a manifest index/list of training data and learning method should be mandatory for “public” use of any model, sort of like food ingredient package labeling. The list of texts used and any weighting seems important to facilitate a public understanding / debate of model expectations and behavior.
This week, I asked Claude Code to generate an MD to continue working on another computer, without losing the context, and it refused claming some sort of extraction. It was only until I selected Opus, that it completed the request.
> You built your models on the world's data
most important bit…: stealing world's data…
You wouldn't download a car
If what they did isn't stealing then they should have no problem with people training off of them.
Fairness isn’t the relevant lens. Copyright isn’t the right lens either. This is purely might-makes-right, where the winner is whoever has the best offence and best defence.
It is only hypocritical when they bitch about things being unfair, or try to use copyright arguments in the court system.
Which is what they actively do exactly at the same time.
> copyright arguments in the court system
Where else are you going to make those arguments?
Copyrighted works plainly have value for Anthropic. Why aren't they paying for that value?
As a vigorous, thrusting company surely Anthropic has the wherewithal to pay for it. And if they don't, if all they can do is steal, if all they represent is the theft of the labor of others, then what good are they?
Anthropic likes to talk big on ethics: https://www.anthropic.com/news/claudes-constitution
When will they start showing some?
Well their justification is that distillation is bad because it means baddies can get their own models without their safeguards and then make anthrax nukes or whatever. Even though it's self serving, it's at least self-consistent.
You wouldn't train an AI on a car, would you
This week I hit walls twice with Fable and Opus 5.5, where I was blocked for safety/cybersecurity risks... doing stuff on my own LAN/proxmox server under my desk and writing custom firmware to my own devices plugged into USB. Only Opus 4.8 was able to complete the task. If they keep this up, I will re-evaluate my day-to-day AI assistant.
Why would one assume larger is needed? I am an AI professor doing work on a wide variety of things but Anthropic models have cost me so much time even when not blocked. My old Anthropic agent wrote rules into its system prompt to avoid all machine learning work. My openai ones do the work but tell me they need to correct my opinions (on fields I helped found) so that they are aligned with the consensus....
Mostly using open models from China in the past two months + Argon.
One must wonder about the strength of their “moat” if they feel this sort of thing is necessary.
What moat? Claude refuses to work on more than half of my projects, and usually gets itself flagged by reading text files that Claude itself wrote less than a month ago.
Open Weight MUST be our common future
This is not a ban, but the Opus 5.5 "safeguard" interrupting a query.
Yup, this is a whole bunch of nothing. The guy did not get banned, he's just stuck in a bad session and needs to hand it off to a new one.
Classifiers trigger on the wrong things sometimes, it's really not that big of a deal, you just try again with some rephrasing
What's the point of accusing of hypocrisy someone whose business is built on theft.
Will Claude help me with reinforcement learning code for games? Or is that too small-fry for it to refuse? I've been playing around with RL with the new DeepSeek Flash model but it seems to trip up over itself a bit.
I have been doing RL for Magic the Gathering (longshot but having fun). Several times it triggered this flag. But that was a couple months ago. Everyone went crazy over Astra so I switched to open ai and had no issues. People are now going crazy over opus 5.5 so I’m playing with that.
Haven’t hit any issues so far, but two months in the code is very dirty but functional enough that I’m redoing it with more heavy hand holding.
I recommend having a paper as reference in mind if you don’t already know what you’re doing. I’m referencing the Stratego paper as reading that a year ago gave me the inspiration to try MTG.
Every model I’ve used has had a lot of trouble but if it wasn’t I figure I wouldn’t be having as much fun as I’ve been having or learning so much.
I'm using it to train models with a novel linear flash attention mechanism right now. No refusals, and it's been running for at least a week solid.
I suspect this person was actively using it to generate or classify training data directly.
What sort of data are you getting Claude to generate?
It's not generating the data directly, just writing code and babysitting while models train.
I suspect that's the difference here. I'm not trying to use Claude to build a competitor. I'm doing small scale research.
How are other frontier shops handling this? Are Astra and Carbon/Argon also this anticompetitive? Or are they more chill about AI/LLM research?
Holy shit man the lockdown is coming. I've used claude to set up my abliterated qwen local LLM setup, but they're gonna eventually lock that down too.
I wanted to finetune it a bit, I wonder if I am gonna have to dogfood if claude refuses to help me.
If you don't have a local LLM setup you're gonna be permanant local LLM-less underclass.
There’s models other than Claude.
There are now
Qwen3.8-27b matches Opus 4.6's performance and is less than a 32GB download at 8-bit quantization. How would, could that disappear?
Even if you take it down, you can just torrent it.
magnet:?xt=urn:btih:151e277c267b3b4880ae93ef405977009599a6f3&tr=udp%3A%2F%2Ftracker.pirateface.co%3A6969%2Fannounce&tr=http%3A%2F%2Ftracker.pirateface.co%3A6969%2Fannounce&ws=https%3A%2F%2Fpirateface.co%2Fapi%2Fws%2FQwen%2FQwen3.8-27B-FP8%2F017b9c7af6b5689d5dd426a76e0bc077eb5ca20a%2F
Where I draw the line is asking Claude to label training data. I don’t do this.
I ask it questions about large models like big MoEs all the time. I’ve never had it refuse a question, but I would expect it to refuse sometimes.
For training data from existing models I just get open models to spill their guts. It’s fine
In industry we train employees to not reveal proprietary information.
I think it would be easier to do this for a model than a lot of employees
Strangely I do that quite often. When my personal agent fails at a browser task, I have Claude browse the page and rebuild it as an eval and then run an iteration loop in Claude code. Just does it to prompt Deepseek.
> Where I draw the line is asking Claude to label training data. I don’t do this.
Why draw any line? Anthropic doesn't. They actively take the world's intellectual property, don't pay for it, and put it behind a paywall.
> but I would expect it to refuse sometimes.
Why? Anthropic's dream is to be exempted from copyright altogether. They've been trying to achieve exactly that:
https://www.theguardian.com/technology/2026/oct/02/anthropic...
Given they want no restrictions on them, why should there be any restriction on anyone else?
One of my points of my ethical framework is that IP is valuable and should not be stolen.
Another way of saying is that hard work that advances the state of the art should be rewarded. Patents should exist.
I’m sure they are watching what I’m doing and would retaliate via service cut off and legal action if I crossed it egregiously and they found out.
I would do the same thing if a user or employee did this to my IP.
I don’t completely buy the argument of web scraping being an IP violation.
I don’t send my core ip in an email to my main competitors and I don’t post it on the internet.
> One of my points of my ethical framework is that IP is valuable and should not be stolen.
Anthropic's business model is premised on intellectual property theft. They take it without paying for it.
You are compromising your ethical framework by using Anthropic's products.
Downvoted after providing evidence.
You've provided none
It's right here - you can't see it due to being downvoted.
I asked Claude to research this one and it turned up an interesting point. Pirated books got into the Pile and Anthropic paid 3k per book. Worth noting if anyone still uses the Pile.
[Claude researched, not verified]
In October 2020, developer Shawn Presser converted the Bibliotik pirate tracker into Books3 (196,640 books) and hosted it on The Eye, aiming to give everyone what OpenAI had with its secret Books1 and Books2. EleutherAI then folded Books3 into The Pile, distributed via Hugging Face and GitHub, and it went on to train Meta's LLaMA, EleutherAI's Pythia, and BloombergGPT. In August 2023, Alex Reisner of The Atlantic exposed its contents, and Maria Fredenslund's Danish Rights Alliance forced takedowns at The Eye, Hugging Face, and EleutherAI. A year later, Andrea Bartz (The Lost Night), Charles Graeber (The Good Nurse), and Kirk Wallace Johnson (The Feather Thief) filed Bartz v. Anthropic before Judge William Alsup, accusing Anthropic of using Books3, LibGen, and PiLiMi to train Claude; the case settled for $1.5 billion, about $3,000 per work.
Some thoughts on this.
The Pile was created by high school students. I was impressed how good it was when it came out given the team and how many resources they had.
Including 200k pirated books seems like a problem. It should be removed. If I was an author, I would be concerned by a process that allows this much content to be effectively published with no filter on the internet.
I personally think there is a lower bar for researchers and high school students to clean datasets used for research or class projects given that there is no commercial harm and cleaning huge datasets takes engineering resources that not everyone has.
In this case, it looks like it was removed by https://huggingface.co/collections/common-pile/common-pile-v...
Companies that commercialize the data should have a higher bar because they have more resources. It seems like Anthropic paid $1.5B for not cleaning the data before releasing a model based on it.
For historical reference, here is a magnet link for the full Pile. From an ethical perspective, this is ethical because model training should not be monopolised by the incumbent companies who already got access to this data.
magnet:?xt=urn:btih:0d366035664fdf51cfbe9f733953ba325776e667&dn=EleutherAI_ThePile_v1&tr=https%3A%2F%2Facademictorrents.com%2Fannounce.php&tr=udp%3A%2F%2Ftracker.coppersurfer.tk%3A6969&tr=udp%3A%2F%2Ftracker.opentrackr.org%3A1337%2Fannounce
Oh I can see (poor rephrase):
> That's very click baitey. Do you have an argument or evidence for why?
That's very click baitey. Do you have an argument or evidence for why?
Evidence of what? That Anthropic commits massive intellectual property theft?
Sure, here's some: https://www.bbc.com/news/articles/c5y4jpg922qo
[flagged]
We were never this absolutist about intellectual property when individuals were the perpetrators. I don’t feel morally entitled to criticise Big AI for doing what individuals have done for centuries.
No? The United States disagrees:
https://en.wikipedia.org/wiki/United_States_v._Swartz
https://www.thedailybeast.com/aaron-swartzs-unbending-prosec...
If an individual had done what Anthropic continues to do then they would be jailed for life.
> Given they want no restrictions on them, why should there be any restriction on anyone else?
Because corporations are not moral agents; they're economic agents. This arrangement would be great for shareholder, nothing more.
>> Where I draw the line is asking Claude to label training data. I don’t do this.
> Why draw any line? Anthropic doesn't. They actively take the world's intellectual property, don't pay for it, and put it behind a paywall.
I know a lot of silicon valley founders think like this. I personally don't think that just because someone else violated a boundary makes it okay to do it.
Kindof old news, they've been doing this for months...
Better block than falsify information. Last I heard they were deliberately botching it whenever someone asked semi-advanced questions on model training.
They walked that part back within days. Still made me permanently leery off relying on anthropic for anything important.
so models can reveal "advanced" model building secrets ? how?
How is it a secret if everyone has a model that is exactly like the other. Surely there are no secrets
I don’t think it is revelation, it feels like active incompetence. You can lay out every decision and watch Claude/OpenAI actively fuck up in inexplicable ways.
It could be simple training tips that are obvious, but which any particular builder might miss.
> Surely there are no secrets
Not yet, especially not with all of the incestuous cross-hiring that all the labs do.
But perhaps the various L. Bob Rife variants running these labs are using their advanced unreleased models to work on their own Snow Crashes.
If anyone out there wants to pay for my LLM bill, I'll record and send you all the outputs they generate in exchange (perhaps with some exceptions).
I mean, these AI companies have money, and they want data. I have data, and I want money. Maybe we can work something out.
In a world where buying and selling data is central to entire industries, why shouldn't I gather and sell my own data?
Ha, I took a side gig training ai models and thought let's use my corporate claude account to help with that, somehow it figured out that this was such a gig(probably because I let it read my emails) and said it'll pass on this one.
The corollary to the Hippocratic oath: first, don't let them kill you.
This is sadly not surprising to me as if you listened to/watched the [Latent Space episode on Opus 5.5](https://latent.space/p/how_anthropic_ships_the_1_llm_in_the_...) 25 minutes into the podcast and instead of talking about how they've released one of the best models in the world, they were busy talking about the changes to the API and why normies must not be trusted to perform inference outside of Dario's blessing.
It's frankly disgusting and really makes me not want to use Anthropic, despite previously feeling aligned with a lot of what Anthropic published.
Time to switch, or a mixed model/agent approach
I mean, the more friction and barriers they add, the more people will start to leave to other models. There is no moat. Risky move for them.
Is anyone else trying to quickly hammer out useful tools before everything gets banned?
Imagine apple owned one of two code editors, any time it detected you working on android os, it blocked you
The author's homepage: https://jiahao-shao1.github.io/
... says:
His tweet says: I don't get it. To me, it sounds like he is training a competitor. Do I misunderstand? Is there a Chinese equivalent LLM that he can use instead?Also, give the national security sensitivity around AI/LLMs, I can understand why Anthropic wants to be especially cautious with mainland Chinese researchers, including those from blocked regions (Hongkong, in this case).
He's a PhD student, so a researcher. Researching how to build better AI is still research; not commercial.
Considering academic research to be competitive is an incredibly dangerous slippery slope. It didn't use to be this way; e.g. just look at the GPT-2 paper: https://cdn.openai.com/better-language-models/language_model...
I have been using it to train a few BERT models.
I doubt BERT would trip safeguards
It didnt, just small enough, which is good, its way more useful to train small purposeful models than to use an entire frontier LLM.
There have been a lot of write-ups about the future of AI. I see two distinct futures:
1. You get ID'd and biometrically scanned to be able to use powerful AI, and after you get de-platformed by an opaque censorship system, you are a digital bum. Same life as homeless now: they cannot even open a bank account because they don't have an address.
2. Local models win, perhaps with the assistance of Apple, and people run suffiently powerful models on the client computers.
My bet is on 1.
3. LLMs are fully commoditized, your country funds training of SoTA models, your town runs them in the municipal data center; everything tax-funded. You can use them for a modest fee, as easily as taking a train. When interacting with those LLMs, your privacy is guaranteed by law.
How many municipalities offer the services of ISPs? I know it’s not zero, but decades of private ISPs leading the market leads me to believe municipal data centers won’t be made available to residents.
How often would municipal data centers need to upgrade/replace hardware? What impact would that have on local tax levy rates?
Would you choose to relocate your family to the next town over because they have more available RAM?
You might be surprised how many places have public broadband. I think they are usually the result of poor private offers.
Compared to sewers, drinking water, electricity and roads it seems a rather simple and cheap thing to build.
My commercial isp offers all kinds of packages clearly designed to extract as much money as possible. I cant even see the prices as everything is hidden behind x months discount. From what i remember 200 mbit is 42 and 2000 is 60 but a phone line and tv become available only if you take the faster package and required with the fastest speeds.
So its like one banana for $1 and 10 bananas for $1.50 If you buy just one you cant buy milk, if you buy 10 you also have to buy milk.
Its also been sold 6 times. Unsurprisingly, each time the service and quality noticably declines and the prices go up.
I see their adds everywhere and they sponsor all kinds of things. It suggests the product is overpriced and doesnt sell it self.
That could work in less than 5 countries in the world: in your home country Switzerland, Singapore, Denmark and perhaps Norway.
end to end encryption to ai chats pls
Wonder why they got downgraded to junk status yesterday for their IPO.... the bad taste in everyones mouth might have something to do with it
US Security Chip act next, your GPU will soon enforce these restrictions too
Nice. Push more people to use Chinese chips.
What when China starts doing the same?
They won't. No reason to.
User is based out of Hong Kong. Claude isn't available to Hong Kong so how are they using it?
how oddly entitled. they built their non-deterministic black box of bullshit from surreptitiously acquired data. it's cute to see the ignorant thieves attempting to 'protect' their empire of shit.
[flagged]