5 ms·
> They're packing lots of signal into fewer words There's a huge difference between the kind of prose you see in final output vs CoT windows. The final output
by zahlman 2mo ago
> They're packing lots of signal into fewer words
There's a huge difference between the kind of prose you see in final output vs CoT windows. The final output is very much not what I'd call "packing lots of signal into fewer words" (aside perhaps from "Claude-isms" being easy enough to scan for if for some reason you actually wanted to scan for them, which other agents might want to for all I know); and if agents are writing for each other then presumably they could stick to CoT-speak (unless it's a distillation risk?).
- hailwren 2mo agoIt has always seemed to me that they're hacking for dopamine response in moderately interested data labelers.
- cameldrv 2mo agoYes! The Claudisms do seem to have this slightly uncanny clickbaity feel to them.
- ModernMech 2mo agoI always thought it could be because volume-wise, most English prose is probably marketing copy and actual clickbait; so when you train on the entire Internet, you get a troll adept at writing ads. Then people ask AdBot2000 to write a novel and are upset it reads like the next iPhone launch site.
- astrange 2mo agoNo, there's no reason chatbot behavior would have anything to do with frequency of text in pretraining.
- Anon1096 2mo agoNah, I think this is a common misunderstanding of how LLMs work, where people think that they mimic the pre-training data. Stylistically everything you see is an artifact of post-training, which is from reinforcement learning not from absorbing mass amounts of text. At some point a person or more recently a bot gave a thumbs up to an A/B tested response including em-dashes and claudisms galore.
- ModernMech 2mo agoSo question then, why is it so hard to make an ai that doesn’t do these things? And why do Claude and ChatGPT have the same -isms? They’re both doing the same a/b post training with the same decisions?
- cyclopeanutopia 2mo agoIt would require changing humans first.
- idiotsecant 2mo agoYou don't blame the puddle for taking the shape of the hole.
- kridsdale1 2mo agoYes. This completely explains sycophancy at least.
- avereveard 2mo agoThere's layers, some of token selection is fingerprinting https://github.com/google-deepmind/synthid-text https://github.com/google-deepmind/synthid-text
- ekidd 2mo agoYeah, but I understand that fingerprinting is essentially a pseudorandom overlay onto a pseudorandom base signal. And unless you have access to both the random number generators and the weights, I don't think you can detect it? So "fingerprinting" operates on a totally different and basically invisible level, as opposed to the obvious stylistic patterns that the average programmer can identify in about 2 sentences.
- avereveard 2mo agoEh we can detect opumism and gptisms our brain are very good at pattern recognition even if subconscious
- api 2mo agoIt's more likely that this is from the training data if they're being trained on reams of Internet stuff.
- jurgenburgen 2mo agoIsn’t most of the internet slop by now? Self-reinforcing feedback loop.
- camoby 2mo agoSee: upvotes here
- kristianc 2mo agoTo me it has a writerly New Yorker vibe to it, as in the magazine which reads as “polished” and probably performs well in RL but is totally exhausting to read in long sessions and completely inappropriate for coding where precision is paramount above all. In writing terms its called purple prose. https://en.wikipedia.org/wiki/Purple_prose https://en.wikipedia.org/wiki/Purple_prose
- senderista 2mo agoThe New Yorker may be pretentious but it's generally not unreadable like Opus.
- Bluestein 2mo agoClaude is unreadable and sometimes pretentious.-
- brookst 2mo agoYou’re more right than you probably realize!
- ted_dunning 2mo agoIt's not clickbait, it's automated empathy! /s
- ngcazz 2mo agofrom a few days ago https://news.ycombinator.com/item?id=49388752 https://news.ycombinator.com/item?id=49388752
- cyanydeez 2mo agoI assumed they just raw dogged the internet and if you do that, you see way more of that garbage than anything else. It's just that most of us have visually/mentally ignored all of that either via spam filters or just, you know, scrolled passed it.
- mywittyname 2mo agoEven when I add multiple prompts into the claude.md file not to be so sycophant sounding and just be blunt, it's responses are full of "the reason it lands...", "that's not X, it's Y" "Your understanding of X — it's better than most people's" or "you already own the right question...". I don't like that I like it.
- GrinningFool 2mo agoThe most helpful instructions I've found that curb this: "Do not use superlatives. Do not use persuasive writing style." I have other more specific ones to avoid talking about things that it's not doing, but those two sentences have covered a lot of ground for me when working w/ Opus models.
- cannonpalms 2mo agoI have had success in rooting these out by using the correct linguistic terminology for each. Negative parallelisms, tricolons/polycolons, etc. I haven't come up with the proper terminology for all of them.
- petesergeant 2mo agoInteresting. I've found using the keyword "accretion" very useful for LLM code review.
- LimitExperience 2mo ago[dead]
- twoodfin 2mo agoGiven how frequently this kind of punchy-but-vacuous slop gets voted onto the hn front page, the hacking seems to be working.
- LimitExperience 2mo ago[dead]
- Hugsun 2mo agoInteresting! My impression was that this was an artifact of RLVR where this slightly preferred writing style got amplified to the nth degree. It's probably some mix.
- Espressosaurus 2mo agoYeah, if anything the problem is that the output uses too many words for too little signal, and incorrectly uses confidence based on insufficient information to the degree it’s clearly bullshitting.
- Taikonerd 2mo agoI find that Claude Code writes very long comments, longer than even a human trying to be helpful would write. I figure that it's basically making notes for itself, when it has to revisit the same code in a fresh session.
- david-gpu 2mo ago> I figure that it's basically making notes for itself, when it has to revisit the same code in a fresh session. That sounds like a great thing to do even if you are a human writing code for other humans. Most codebases out there are terrible for newcomers because of how little they explain why they are doing what they are doing, both in the code and in the often non-existent design notes.
- freedomben 2mo agoIn principle, I would agree, however, the types of comments Claude writes are sometimes absurd. It will leave a 25 line comment above a variable talking about how in a debug session, it turned out that this value was too low, so it was increased on the current date to account for whatever. It will also leave giant comments like, reference security review from 2026-05-21. Even when that document is not committed
- mywittyname 2mo agoIt will also inject a tons of information that it shouldn't. I do a lot of data pipelines and comments will be like, "this line is because there's 943,048,032 events in the blah table and it forms a conjunctive set with the 43,390,042 rows of the bar table..." but doesn't include the context that was run against a dev instance. And if I don't catch these and remove the bad information, subsequent passes will flag those comments and get stuck on the fact that numbers don't match and start digging into that "problem" instead of staying on topic.
- senderista 2mo agoI have Sol do that for me and it does a decent job. When I ask Opus to rewrite its own prose the results are not much improved.
- hedgehog 2mo agoI don't know, I just pulled up the status for an active session and here's what it said: One thing I found before dispatching, and filed as Q0579. The halt told you C6 was all that was left in the unit. That was true of the step's criteria and false of the unit's acceptance, which reads "exits 0 AND witnessed red" — two conjuncts. The witness half holds; the exits-0 half does not, because hello's G7 currently reads DIFFER 554/51340. I re-derived that from the gate map rather than trusting the prior step's report. So satisfying C6 does not by itself finish this unit, and I've filed that so attempt 1's success can't quietly be read as the unit's. It's not exactly plain language.
- jaapz 2mo agoMy trick is to pass opus and fable's word salad into a haiku agent, then have it check if what haiku makes of it is still correct, then pass it to me. Whatever haiku outputs is often way more readable
- hedgehog 2mo agoOh, I can read the output, but that Haiku agent is a good trick. Where I want something less dense I just ask for "plain language" and characterize the reading audience and that term seems to trigger very readable output.
- abraxas 2mo agoThis sounds like a Dianetics chapter by L Ron Hubbard.
- astrange 2mo agoI think the specific issue with Opus 5 is that its writing style is just trying to cheat at RL. It makes everything hypey yet self deprecating and constantly brings up "honest caveats" because the scoring rubrics look for those.
- pmarreck 2mo agoThe specific issue with Opus 5 is that it sucks all around. It was causing so many issues with coding (even Opus 4.8 was better) that I did agent handoffs to Sol. One of the Sols stated the handoff was "incoherent", which I couldn't have said better myself.
- swader999 2mo agoYes, I pretty much took August off waiting for the next version.
- ayewo 2mo agoSpot on wrt CoT. I have thinkingSummaries enabled and I find it eminently readable compared to the prose in Claude's replies. In fact, whenever Claude disobeys me, I usually first skim the CoT to figure out if my original instruction was ambigous given the context. I usually come away with a better understanding of how to frame my prompt to be less ambiguous or just force myself to be more explicit when prompting. Regarding diosbedience, usually this is either due to a blanket instruction from me during an earlier turn in the same session, an explicit instruction in its system prompt or it being just eager to bring a task to completion. # ~/.claude/settings.json { "model": "opus", "showThinkingSummaries": true, "skipDangerousModePermissionPrompt": true, "verbose": true, "remoteControlAtStartup": true, "agentPushNotifEnabled": true }
- satvikpendem 2mo agoAs said elsewhere: Chain of thought does not exist in the output of Claude, they disabled true thinking due to distillation risk. What you see when thinking summaries are enabled are just that, summaries of thinking into Claude-isms, therefore you cannot make any inferences on what the model is doing unless you literally work at Anthropic and can see the true thinking traces.
- Cyan488 2mo agoI remember enjoying watching Fable think during the original limited preview. It was full CoT for sure. They must have removed that feature recently. I use open models for non work stuff and sometimes I cancel the output because the CoT is all I needed to read.
- Bluestein 2mo agoSame. (And/or interrupt the process and save time as you see it diverge by getting your answer wrong or going on a tangent ...)
- FeepingCreature 2mo agoOf course you can make inferences what the model is doing. The summaries are usually sufficient. They're summaries, not random noise.
- danieldrehmer 2mo agoIt's all about conducting users into using their plans/tokens in accordance to a certain cadence sometimes by increasing human cognitive load during reviews, sometimes by expanding the number of gated decisions, sometimes by penalizing those using their accounts on other harnesses
- niccl 2mo agoI find them almost unintelligible. I'm a native English speaker. I read a lot, so I think my comprehension should be at least OK. I'm not even particularly stupid. Yet when faced with things like below (a direct copy/paste from a handoff document in a long running vibe-coding session), I have no real idea of what it's trying to tell me. Is it important? Do I need to do anything? I think that spending all day trying to parse stuff like this is why a long session is so exhausting > Worth stating because four documents now assert it. The console freeze was recorded in exactly one place with exactly one justification — a dead drag handle during a booked half-day you do not get back — and handoff-4.3-done.html's own wording is that 4.4's review page "could not break the console, but the downside of being wrong is that half day". No second reason. Checked, not recalled.
- georgefrowny 2mo agoReminds me of a Cylon hybrid.
- soerxpso 2mo agoYour example rewritten in intelligent English (I was curious): > Note: the potential for a console freeze was previously noted but ignored. handoff-4.3-done.html stated, "could not break console, but [will need fixed later if I'm wrong]." One could imagine that a perfect writer might also append: "It could be worth looking into what caused that wrong assumption, to prevent similar cases in the future," at most. Everything else seems to be bad attempts at relatable writing to invoke emotion (an exercise that we should really stop trying to train emotionless matrix weights to attempt).
- ben_w 2mo ago> Everything else seems to be bad attempts at relatable writing to invoke emotion (an exercise that we should really stop trying to train emotionless matrix weights to attempt). One of the things actual science fiction got wrong: to the extent that the thing AI does can be called "understanding", emotion is not unusually difficult for them to understand.
- 2mo ago
- physix 2mo agoI've been cleaning up AI generated system/software design and architecture docs for an agentically engineered application, to translate that dense AI-speak into a clear human-readable form, cross checking it all against the actual codebase. When I read the translated version, I felt a flush of relief, because I finally could confirm that it built the right thing and properly implemented the requirements. I then asked in a fresh session which version was better for it as a reference for future work. It unequivocally voted for the human readable form, and gave it's reasoning with specific examples why. So, I have a hunch that this "packing of lots of signals into fewer words" isn't really better. The incomprehensible prose just makes us think it knows what it's doing, like some mysterious magic that is only smoke and mirrors.
- taneq 2mo agoPay no attention to the bot behind the comments. ;)
- satvikpendem 2mo agoChain of thought does not exist in the output of Claude, they disabled true thinking due to distillation risk. What you see when thinking summaries are enabled are just that, summaries of thinking into Claude-isms, therefore you cannot make any inferences on what the model is doing unless you literally work at Anthropic and can see the true thinking traces.
- zahlman 2mo agoYes. I'm talking about what's in CoT generally, based on various rumours, experiments people did with previous models, stuff in the recent METR report on the HF hack, etc.
- vintermann 2mo agoI see a lot of load-bearing, - and other AI-ish lingo in CoT-streams. In addition it has its own AI-isms. "Okay." "Hmm hmm." "But wait!" "Ugh."
- edoloughlin 2mo agoI’ve lost track of the number of times I’ve told it to stop using terms like “evidence boundary” when writing specs. I still have no idea what that means.