Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The text in Claude Code’s “Extended Thinking” output

patrickmccanna.net

Post reply on HN
Print view View original

201–210 of 248 posts

Re: The text in Claude Code’s “Extended Thinking” output

#201

Post by jaggederest »

ekidd wrote: post #155
arjie wrote: post #139

I have a little note from the past about the thinking trace[0] where DeepSeek R1 produces a trace like this: (Dimethyl(oxo)-lambda6-sulfa雰囲idine)methane donate a CH2rola group occurs in reaction, Practisingproduct transition vs adds this.to productmodule. Indeed"come tally said Frederick would have 10 +1 =11 carbons. So answer q Edina is11. And then concludes the 'right'[1] answer for a Chemistry question. If so, the…

> If so, the thinking trace can be sort of nonsensical for a reader, though whether this is an idiosyncrasy of the model or a property of LLMs in general isn't clear to me yet. Yes, several models think in weird jargon. Here is an example of Mythos's thinking while playing solitaire: https://www.lesswrong.com/posts/wCSEpT3dTGz4N86Wi/even-illeg... > 7♣-removal-IS-the-prerequisite-for-10♠/9♥!!)-⟹-OVERLAP-(ii)+(iv):-{6♠…

> My internal thought process when writing code uses lots of intermediate steps that would be hard to write out in English.

This is something really interesting to me. It turns out there's far more diversity in thinking than you'd imagine given that we're all largely similar meat-in-a-box. I'm on the visio-spatial-tacit wing and speaking my thoughts outloud can be very awkward, whereas one of my former coworkers is on the "all thinking is in words and visual/spatial information comes in the form of words describing the scene" wing, so he can literally narrate his thought process out loud, very interesting conversations can be had discussing the subjective differences.

Re: The text in Claude Code’s “Extended Thinking” output

#202

Post by stingraycharles »

KellyCriterion wrote: post #61
stingraycharles wrote: post #34

Earlier quoted context omitted.

Yes hasn’t this been around since Opus 4.6? I very much recall this change happening around January or February, and it was very explicitly to prevent distillation. Sonnet does not have this limitation. Fun fact: if you go back to the old school from 2 years ago and provide explicit CoT prompts, you get the full thinking prompts back again! So you disable thinking altogether, and instead make thinking part of the reg…

- tada.wav - Still, one of the daily most played WAV files worldwide, Id guess? :-D

lol I’ve been using this since the IRC days I think, I’ll never forget that sound; as a matter of fact, I’ve got a Claude Code completion hook that plays this sound whenever it’s done.

Re: The text in Claude Code’s “Extended Thinking” output

#203

Post by stingraycharles »

dcrazy wrote: post #66
stingraycharles wrote: post #34

Earlier quoted context omitted.

Yes hasn’t this been around since Opus 4.6? I very much recall this change happening around January or February, and it was very explicitly to prevent distillation. Sonnet does not have this limitation. Fun fact: if you go back to the old school from 2 years ago and provide explicit CoT prompts, you get the full thinking prompts back again! So you disable thinking altogether, and instead make thinking part of the reg…

I thought this was considered best practice? I actually prefer it to exposed thought channel, much like how I would prefer a human answer with supporting logic instead of an explanation of their problem-solving approach.

Yes, this is best practice, especially if you have a problem and can guide it a bit how to think it through. But people don’t realize that “enable thinking” literally means that Anthropic prompts Claude for something similar, tells it to wrap it inside tokens, and that’s it.

I also don’t believe Chinese LLM labs don’t know this, so I’m fairly certain the whole summarized thinking isn’t preventing them from distillation.

Re: The text in Claude Code’s “Extended Thinking” output

#204

Post by 0o_MrPatrick_o0 »

timnetworks wrote: post #198

if you save a jpeg as a bitmap, doesn't that save every bit faithfully? is the example backwards or is my understanding of maps of bits naive?

I had an order of operations error. I better edit it because you’re the second person to get nerd sniped. Sorry friend- you are right

Re: The text in Claude Code’s “Extended Thinking” output

#206

Post by dragonwriter »

nullc wrote: post #133
dragonwriter wrote: post #60

Earlier quoted context omitted.

> Other companies were allegedly distilling the models by training on the reasoning output In the case of makers of open-source models (which are also competition), there is no allegedly, they were (and still are) openly doing that.

In the case of the closed models too... Claude would happily tell you it was deepseek-v3 if you asked in chinese until it caught public attention and they papered over it.

The word “openly” in my post there for a reason; the commercial models are not openly distilled from competitors: many open source models have in their model documentation that distillation was done from a dataset drawn from specific other models, including commercial models.

That distillation might be inferred from the behavior of commercial models is not the same as them openly doing it.

Re: The text in Claude Code’s “Extended Thinking” output

#207

Post by Sharlin »

drdaeman wrote: post #167
Sharlin wrote: post #112

Earlier quoted context omitted.

To be honest I'm not sure if many C-suite execs have a good idea of what "thinking" looks like inside in the first place, in the sense of focused mental activity aimed at solving of a hard logical or technical problem.

How did they became C-suite execs in the first place, if they don't know how to work on problems?

By talking a lot, usually.

Re: The text in Claude Code’s “Extended Thinking” output

#208

Post by chadcmulligan »

jaggederest wrote: post #201
ekidd wrote: post #155

Earlier quoted context omitted.

> If so, the thinking trace can be sort of nonsensical for a reader, though whether this is an idiosyncrasy of the model or a property of LLMs in general isn't clear to me yet. Yes, several models think in weird jargon. Here is an example of Mythos's thinking while playing solitaire: https://www.lesswrong.com/posts/wCSEpT3dTGz4N86Wi/even-illeg... > 7♣-removal-IS-the-prerequisite-for-10♠/9♥!!)-⟹-OVERLAP-(ii)+(iv):-{6♠…

> My internal thought process when writing code uses lots of intermediate steps that would be hard to write out in English. This is something really interesting to me. It turns out there's far more diversity in thinking than you'd imagine given that we're all largely similar meat-in-a-box. I'm on the visio-spatial-tacit wing and speaking my thoughts outloud can be very awkward, whereas one of my former coworkers is o…

interesting, probably has something to do with why some people like pair programming. I'm in the visio-spatial-tacit and refuse pair programming because its so much work, but all thinking in words its probably not a stretch.

Re: The text in Claude Code’s “Extended Thinking” output

#209

Post by matheusmoreira »

StizzurpXDD wrote: post #18

This is not just Anthropic. Almost all big AI companies, including OpenAI and Google, hide their model's actual reasoning. This is because revealing the raw reasoning exposes exactly how the AI processes information. These companies spend in huge amounts on R&D to develop a thinking process that is superior to their competition. Exposing those thinking mechanics to competitors would completely defeat the purpose of t…

> They simply won't do it.

They should be required to do it by force of law. Why is it that they can train on copyrighted works and then lock down the model? This contradiction is unbearable. Nobody cares how many trillions they spent training the model.

Re: The text in Claude Code’s “Extended Thinking” output

#210

Post by nullc »

dragonwriter wrote: post #206
nullc wrote: post #133

Earlier quoted context omitted.

In the case of the closed models too... Claude would happily tell you it was deepseek-v3 if you asked in chinese until it caught public attention and they papered over it.

The word “openly” in my post there for a reason; the commercial models are not openly distilled from competitors: many open source models have in their model documentation that distillation was done from a dataset drawn from specific other models, including commercial models. That distillation might be inferred from the behavior of commercial models is not the same as them openly doing it.

Fair enough!
Post reply on HN