911–920 of 1001 posts
Too late, personally after how bad 4.6 was the past week I was pushed to codex, which seems to mostly work at the same level from day to day. Just last night I was trying to get 4.6 to lookup how to do some simple tensor parallel work, and the agent used 0 web fetches and just hallucinated 17K very wrong tokens. Then the main agent decided to pretend to implement tp, and just copied the entire model to each node...
They've increased their cybersecurity usage filters to the point that Opus 4.7 refuses to work on any valid work, even after web fetching the program guidelines itself and acknowledging "This is authorized research under the [Redacted] Bounty program, so the findings here are defensive research outputs, not malware. I'll analyze and draft, not weaponize anything beyond what's needed to prove the bug to [Redacted]. I…
I can understand the wishes to make LLMs even more self driven. After all that's the idea of a lose prompt. No matter how short, LLM figures out what most users are expecting. Thanks to RLHL it accomplishes wonders. My desire though is to be able to steer the model exactly where I want. Assuming token cost isn't an issue, it doesn't remove the need for costly review. I would rather think first and polish up my abilit…
This is a CC harness thing than a model thing but the "new" thinking messages ('hmm...', 'this one needs a moment...') are extraordinarily irritating. They're both entirely uninformative and strictly worse than a spinner. On my workflows CC often spends up to an hour thinking (which is fine if the result is good) and seeing these messages does not build confidence.
There’s one that’s like “Considering 17 theories” that had me wondering what those 17 things would be, I wanted to see them! Turns out it’s just a static message. Very confusing.
Assuming /effort max still gets the best performance out of the model (meaning "ULTRATHINK" is still a step below /effort max, and equivalent to /effort high), here is what I landed on when trying to get Opus 4.7 to be at peak performance all the time in ~/.claude/settings.json: { "env": { "CLAUDE_CODE_EFFORT_LEVEL": "max", "CLAUDE_CODE_DISABLE_BACKGROUND_TASKS": "1" } } The env field in settings.json persists across…
Subagents are very useful. But sometimes it uses sonnet or haiku. You can try something like "always use opus for subagents" if you want better subagents.
I'm not sure how much I trust Anthropic recently. This coming right after a noticeable downgrade just makes me think Opus 4.7 is going to be the same Opus i was experiencing a few months ago rather than actual performance boost. Anthropic need to build back some trust and communicate throtelling/reasoning caps more clearly.
What I want to know is why my bedrock-backed Claude gets dumber along with commercial users. Surely they're not touching the bedrock model itself. Only thing I can think of is that updates to the harness are the main cause of performance degradation.
Earlier quoted context omitted.
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.
Fundamentally they're the same technology with the same exact algorithms under the hood; only the post-training alignment differs. That is, the difference you see is either placebo effect or you being lucky and better aligning with model post-training bias.