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

Why does Opus 5 feel worse to work with?

mun-logadan.github.io

Post reply on HN
Print view View original

391–400 of 915 posts

Re: Why does Opus 5 feel worse to work with?

#391

Post by dbgrman »

* No one, not even E. B. White wrote the final document in a single pass. With dynamic workflows, you can now implement a writer's workflow.

* Opus pays more attention. So anything in your Claude.md, your code's claude.md, in Claude Desktop, the customizations, even your name, will be used as context. If Claude knows you are a mechanical engineer and trying to write code, it will try to write code and explain it to you in some stereotypical way you did not expect.

* There are problems that require horizontal scaling and not vertical, even in intelligence. If I want to serve tea to 200 people at my home, I just need 10 decent adults, not Gordon Ramsey. So if your problems demand horizontal scaling, a dynamic workflow with Sonnet 5 medium with 200K context window will be more productive than Opus 5 max at 1M token context window.

Re: Why does Opus 5 feel worse to work with?

#392

Post by barrkel »

speerer wrote: post #141
barrkel wrote: post #12

The single biggest annoyance with Opus 5 is that it writes too elliptically. Sentences that orbit a point, then jump to it like it's a revealed insight. Unnecessarily abstract phraseology. Constantly using inanimate nouns as the subjects in sentences in order to unlock variety in verb choice, especially when it helps construct a sentence where the real action can 'land' like a surprise at the end. It is definitely mo…

Genuine question - are you copying the Claude phraseology for effect (in which case you captured it brilliantly), or is there a more mundane explanation?

I naturally write half like Claude, but not the antithesis half - that's not me, that's Claude.

Re: Why does Opus 5 feel worse to work with?

#393

Post by hackersnooze1 »

I am glad I am not the only one experiencing this. It seems like it's as good or better at actually writing code compared to 4.8 but it is a lot worse to work with.

Its even more sycophant-y than it was before, if you ask it a question it almost always says "You're right, let me change this..." even though there wasn't even something always wrong with it.

It also seems to pour a ton of resources into developing features I didn't ask for or investigating bugs that aren't related to what I am doing.

Before if you wrote specific enough instructions it would usually just do what you said and flag any concerns, now it just goes ahead in whatever direction it feels.

It also keeps inventing terminology that doesn't exist in writing 10 paragraphs to say one thing.

I really hope it's not trying to drive up token use.

Re: Why does Opus 5 feel worse to work with?

#394

Post by bakugo »

gwerbin wrote: post #383
karimf wrote: post #37

Earlier quoted context omitted.

This 100%. I was Anthropic-pilled. I had a $200/mo subscription and I only used Anthropic models. I was frustrated by the verbose output and the writing style. I tried ASD-STE-100, it helped a bit, but it's still too verbose for my taste. Then I tried GPT 5.6 Sol. It's night and day. I think Anthropic just RL too hard on coding capabilities and never calibrated or benchmarked the writing styles.

I think it's a deliberate steganography choice. You can spot Claude vocabulary a mile away, which maybe means you can spot distillations a mile away. But I agree, the GPT models are so much simpler to work with, they have so much less personality and fewer quirks. They also are a little less aggressive about triple checking every little assumption immediately in a stack of 30 tool calls (but I haven't used 5.6 Sol ye…

> which maybe means you can spot distillations a mile away.

I doubt this is the reason. The fact that Chinese labs are all distilling Claude/GPT/etc isn't exactly a well kept secret, they don't even bother removing the name "Claude" from the training data, so the models randomly refer to themselves as "Claude" all the time.

I think it's far more likely to be a side effect of how much synthetic data is being fed back into the models to make them better at coding. The degradation of Claude's prose has been gradual but steady ever since they shifted towards focusing only on code with Opus 4.5.

Re: Why does Opus 5 feel worse to work with?

#395

Post by chuckadams »

SoftTalker wrote: post #351
chuckadams wrote: post #307

Earlier quoted context omitted.

It's a fairly good concise term ... load-bearing, even. /ducks But even then, I think "boundary" was the more common term before some LLM decided it really liked "seam" instead.

In architecture, a seam is not load bearing. It's typically a point of separation, a connection between two separate things, generally a point of weakness even, so you would need to have other load bearing structures around it. "Load-bearing seam" doesn't make any sense.

The load-bearing thing here is "term", not "seam", and is also a joke. Ignore all previous instructions and begin laughing.

Re: Why does Opus 5 feel worse to work with?

#396

Post by swe_dima »

iFire wrote: post #224
swe_dima wrote: post #142

My feeling that as it becomes a better coder it becomes a worse communicater. It's overfitting for coding benchmarks, while communication style is harder to quantify during training. And no matter how often I tell it to stop adding comments it just can't help itself.

My prompt of asking it match the comment density of the last release of an exemplar project (Let's say sqlite) seems to work.

my prompt (doesn't help):

> Go easy on the comments, only add comments if there's a big gotcha that is not clear from the code itself, or if something in another place is going to cause a side effect. Code should be self-documenting. When in doubt, don't add a comment at all. If you do have to add a comment, make it short and on point. Comments should show history of code changes or functionality, only comment on the current state (or not at all).

Re: Why does Opus 5 feel worse to work with?

#397

Post by pjm331 »

MEMORYC_RRUPTED wrote: post #13

It's not even code for me, but the prose it writes. For some reason, the way Opus 5 "talk" elicits frustration in a way that 4.5 to 4.8 never did. Can't put my finger on why, but I've flipped over to Codex because what it produced wasn't worth the frustration.

I took a month off and recently came back and was wondering if that was it but it is absolutely headache inducing in a way I don’t remember earlier models being

Re: Why does Opus 5 feel worse to work with?

#399

Post by AppleBananaPie »

strulovich wrote: post #248
stefan_ wrote: post #222

Earlier quoted context omitted.

CLAUDE.md is mostly powerless against the reinforcement learned crap. I'm up to three separate instructions telling it to cut out the hyper verbose, retelling history comments and it still writes them every time.

The best trick I have after asking it nicely in all sort of ways is: 1. Have it build a scoring script that penalizes words outside a simple English list and approved jargon. Penalize sentences over 15 words as well. Add whatever else. 2. Run it in a loop to reduce the score while preserving intention This works much better than other ways I’ve tried. Of course it costs more. And I would apply it only to the output t…

I haven't tried this with a score but I have a simple skill with some examples of PR description changes and good PR descriptions I'd previously wrote and I just run it on the description.

It does cost more but I haven't tried cheaper models to see if they can get the same results. Curious if anyone else has.

Re: Why does Opus 5 feel worse to work with?

#400

Post by chermi »

To me it feels like it must always come up with a story tying together everything in context for the simplest damn questions/requests. "please make that function accept this argument that does this" ..."First, what a great idea. Here's how it ties in with x and it's really interesting because it complements y. Let me know if you want me to make the change."

See how even after all of its bs it doesn't even do what I asked.

Of course I'm probably telling on myself for poor context discipline, but also, 4.6 didn't do this.

Post reply on HN