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

Claude Opus 4.7

anthropic.com

Post reply on HN
Print view View original

831–840 of 1001 posts

Re: Claude Opus 4.7

#831

Post by gawa »

wsmhj wrote: post #827

Tried 4.7 on a few of my regular workloads. The quality ceiling is definitely higher than 4.6 when it actually engages — but that's the problem. "Adaptive thinking" seems to actively avoid thinking on tasks where I'd expect it to reason carefully, and I end up getting flat, fast answers where I wanted depth. Turning off adaptive thinking and bumping effort to high gets me closer to what I want, but at that point the…

How did you disable adaptive thinking for your experiment? In the documentation of claude code [0] it says:

> Opus 4.7 always uses adaptive reasoning. The fixed thinking budget mode and CLAUDE_CODE_DISABLE_ADAPTIVE_THINKING do not apply to it.

[0] https://code.claude.com/docs/en/model-config#adaptive-reason...

Re: Claude Opus 4.7

#832

Post by captainkrtek »

I use Claude Opus 4.6 as an enterprise user, and have also noticed a lobotomization. In recent weeks it's been much more self-correcting even within singular responses ("This is the problem - no wait, we already proved it can't be this - but actually ...") I'm wary of 4.7 being a change in this pattern, it's frustrating to have such a substantial change in experience every few months.

Re: Claude Opus 4.7

#833

Post by spaceman_2020 »

EmanuelB wrote: post #808

I can't notice any difference to 4.6 from 3 weeks ago, except that this model burns way more tokens, and produces much longer plans. To me it seem like this model is just the same as 4.6 but with a bigger token budget on all effort levels. I guess this is one way how Anthropic plans to make their business profitable. During the past weeks of lobotomized opus, I tried a few different open weight models side by side wi…

Yeah my rate limits are getting exhausted way faster now. Its also way slower and overplans unless you steer it closely.

I can’t rely on this anymore.

Re: Claude Opus 4.7

#834

Post by SnowLprd »

gghootch wrote: post #270
cupofjoakim wrote: post #20

> Opus 4.7 uses an updated tokenizer that improves how the model processes text. The tradeoff is that the same input can map to more tokens—roughly 1.0–1.35× depending on the content type. caveman[0] is becoming more relevant by the day. I already enjoy reading its output more than vanilla so suits me well. [0] https://github.com/JuliusBrussee/caveman/tree/main

Caveman is fun, but the real tool you want to reduce token usage is headroom https://github.com/gglucass/headroom-desktop (mac app) https://github.com/chopratejas/headroom (cli)

This smells heavily of astroturfing. Particularly because Headroom is a paid product, and that fact is not mentioned here or in the GitHub README.

Here was my experience…

I download and run the Mac application, which starts installing a bunch of things. Then the following happens without advance notice:

- Adds background item(s) from "Idiosyncratocracy BV"

- Downloads over 2 GB of files

- Pollutes home with ~/.headroom directory

- Adds hook(s) to ~/.claude/hooks/

- Modifies your ~/.claude/settings.json to add above hook(s)

… and then I see something in the settings that talks about creating an account. That's when I realized that this is a paid product, after all of the above has happened.

Headroom seems to use https://github.com/rtk-ai/rtk under the hood. What does Headroom offer over the actually-free RTK? Who knows.

At this point I have had it with this subterfuge — I immediately trash the app and every related file and folder I can find, of which there are many. Hopefully I got them all, but who knows. There should have been an easy way to uninstall this mess, but of course there isn't.

The lack of transparency here is really concerning.

Re: Claude Opus 4.7

#835

Post by elAhmo »

EmanuelB wrote: post #808

I can't notice any difference to 4.6 from 3 weeks ago, except that this model burns way more tokens, and produces much longer plans. To me it seem like this model is just the same as 4.6 but with a bigger token budget on all effort levels. I guess this is one way how Anthropic plans to make their business profitable. During the past weeks of lobotomized opus, I tried a few different open weight models side by side wi…

Its funny to think that with a model release Anthropic can slide in some instructions ("be a bit more detailed" or something similar) that affect the token output by a few percent, 5-10%, which will not be noticeable by most users but over the course of the year would bring solid growth (once the VC craze is over, if ever) and increase income.

"Regular companies" would love to have a growth like that without effectively doing anything.

Re: Claude Opus 4.7

#836

Post by paulluuk »

EmanuelB wrote: post #808

I can't notice any difference to 4.6 from 3 weeks ago, except that this model burns way more tokens, and produces much longer plans. To me it seem like this model is just the same as 4.6 but with a bigger token budget on all effort levels. I guess this is one way how Anthropic plans to make their business profitable. During the past weeks of lobotomized opus, I tried a few different open weight models side by side wi…

If open weight models are sufficient for your engineering problems, then you should absolutely use them. But I haven't seen a single open weight model that can get even close to the complexity in my projects. They sometimes work for small toy examples or leetcode puzzles, but not very any real project. Really curious what models you've found that could replace current state of the art.

Re: Claude Opus 4.7

#837

Post by keyle »

captainkrtek wrote: post #832

I use Claude Opus 4.6 as an enterprise user, and have also noticed a lobotomization. In recent weeks it's been much more self-correcting even within singular responses ("This is the problem - no wait, we already proved it can't be this - but actually ...") I'm wary of 4.7 being a change in this pattern, it's frustrating to have such a substantial change in experience every few months.

Frustrating that the experience changes, and then they retire the better older model because it costs more, although it was better for everyone. The new ones are just geared better towards beating the benchmarks at a cheaper cost!

Re: Claude Opus 4.7

#838

Post by mattmanser »

EmanuelB wrote: post #808

I can't notice any difference to 4.6 from 3 weeks ago, except that this model burns way more tokens, and produces much longer plans. To me it seem like this model is just the same as 4.6 but with a bigger token budget on all effort levels. I guess this is one way how Anthropic plans to make their business profitable. During the past weeks of lobotomized opus, I tried a few different open weight models side by side wi…

I just don't believe you.

The vast gulf between open weights and frontier models that existed 6 months ago has suddenly disappeared?

It's far more likely you're just bad at assessing model output.

Re: Claude Opus 4.7

#839

Post by OrangeMusic »

trueno wrote: post #455

noticing sharp uptick in "i switched to codex" replies lately. a "codex for everything" post flocking the front page on the day of the opus 4.7 release me and coworker just gave codex a 3 day pilot and it was not even close to the accuracy and ability to complete & problem solve through what we've been using claude for. are we being spammed? great. annoying. i clicked into this to read the differences and initial exp…

HN is literally full of contradictory stories when it comes to which model is better. It's impossible to know if one is objectively better than another, I suspect they're more or less equivalent. People who just recently had a particularly good experience will post that the model they used is better than the other ones. People who just recently had a particularly bad experience will say the model "got worse"...

It's all based on vibes!

Re: Claude Opus 4.7

#840

Post by rl3 »

captainkrtek wrote: post #832

I use Claude Opus 4.6 as an enterprise user, and have also noticed a lobotomization. In recent weeks it's been much more self-correcting even within singular responses ("This is the problem - no wait, we already proved it can't be this - but actually ...") I'm wary of 4.7 being a change in this pattern, it's frustrating to have such a substantial change in experience every few months.

>..."This is the problem - no wait, we already proved it can't be this - but actually ..."

Ditto. Has me wondering why there isn't a reconciliation pass somewhere on the final output.

At least it's a decent signal for when model confidence is low.

Post reply on HN