So the trick is to always set to max, and then begin every task with “this is an extremely complex task, do not complete it without extensive deep thinking and research” or whatever.

You’re basically fighting a battle to make the model think more, against the defaults getting more and more nerfed to save costs.

Time to make a deal with the kids - i’ll verify you for instagram if you verify me for ChatGPT

https://youtu.be/FBSam25u8O4

I expected this to become less necessary over time as models got faster, but the opposite has happened. It feels like Claude has actually gotten slower (but in fairness does more per prompt), meaning worktrees are even more essential now.

What is the network effect of Google Search?

Claude will typically resolve these in a surface level way when in practice they often require deeper changes (to prompts, routing to different model, more general cleaning code, etc) and it’s hard/impossible to have Claude do these without some input.

Other noise arises from Claude thinking related issues are unrelated and so solving them separately, and also just intermittent infrastructure type issues that are clearly transient being “solved” with some weird code change.

By what metric?

The exception is when a fix is a) trivial or b) affecting a real user and therefore needs to be fixed quickly, in which case the current workflow is useful. But yeah, the real step-change was having Claude hitting the Sentry APIs directly and getting the info it needs, whether async or not.

I'd also imagine that people's experiences with this vary a lot depending on the size and stage of the company - our focus is developing new features quickly rather than maintaining a 100% available critical production service, for example.

the timing is not a huge issue though because the type of bugs being caught at this stage are rarely so critical they need to fixed in less time than that - and the bandwidth is limited by someone reviewing the PR anyway.

the other issue is crazy token wastage, which gets expensive. my gut instinct re triaging is that i want to do it myself in the prompt - but if it prevents noise before reaching claude it may be useful for some folks just for the token savings.

no, I don’t receive alerts because i’m looking at the PR/issues list all day anyway, it would just be noise.

I have a github action that runs hourly. It pulls new issues from sentry, grabs as much json as it can from the API, and pipes it into claude. Claude is instructed to either make a PR, an issue, or add more logging data if it’s insufficient to diagnose.

I would say 30% of the PRs i can merge, the remainder the LLM has applied a bandaid fix without digging deep enough into the root cause.

Also the volume of sentry alerts is high, and the issues being fixed are often unimportant, so it tends to create a lot of “busy work”.

Curious what plans you’re using? running 24/7 x 5 agents would eat up several $200 subscriptions pretty fast

Comments URL: https://news.ycombinator.com/item?id=47012749

Points: 1

# Comments: 1

OpenClaw runs with an Anthropic/OpenAI API key though?