staticvar

Joined 12 years ago

Karma

330

Submissions

148

Member since

Feb 25

View on Hacker News

Recent submissions

Comment · 1 month ago

I was really hoping that Epic had released a harness that plays games.

What's on the Box vs. What's in the Box rjcorwin.github.io

1 staticvar 1 month ago 0 comments

Comment · 3 months ago

I'm not sure that AI helps us with that last 1%. Just the 99% which is why, instead of having a backlog of 0% projects, we have a backlog of 99% projects. My hope is that we're more effective prioritizing a backlog of 99% projects than 0% projects.

Comment · 4 months ago

Maybe revenge, but it's a common play to fire a shot across the bow to create leverage in other areas.

Comment · 5 months ago

Totally fair argument to make right now. But so funny how this is the opposite movement of "Let me Google that for you".

Comment · 7 months ago

> We told Claude Code to block npx using its own denylist. The agent found another way to run it and copied the binary to a new path using /proc/self/root to bypass the deny pattern. When Anthropic's sandbox caught that, the agent disabled the sandbox. No jailbreak, no special prompting. The agent just wanted to eagerly finish the task.

I wish that article went into more detail about that attack. But I believe it, the extent that the permissions are easy to get wrong in your claude setttings. For example: https://www.youtube.com/watch?v=3CSi8QAoN-s&lc=UgwFNAh5fvDGJ...

Comment · 7 months ago

Too many cooks? >.<

Comment · 7 months ago

Also on that point about keeping humans in the loop on decisions, I've found following the Research-Plan-Implement process where we humans review at each of those stages, to be really helpful. This doc describes the skill I use with my agents so they keep me looped in: https://gist.github.com/rjcorwin/296885590dc8a4ebc64e70879dc...

Then I use cook to iterate and explore during the AI led parts.

Comment · 7 months ago

Yes, plz don't trust it, always review! The idea is that one prompt in Claude Code got you 80% of the way there, but with some automated review/iterate, it gets you 95% of the way there. It's not worth your time to review the 80% done version when you could be reviewing the 95% done version.

Comment · 7 months ago

Dagu.sh, using yaml files to describe the flow, looks like a nice step up in sophistication from the cook approach that's just trying to make it easy to issue directly from the command line.

My 2 cents on the dagu.sh website, it should lead with the demo section (https://docs.dagu.sh/overview/#demo). That helped me connect what it was and how I might use it.

Comment · 7 months ago

Done! @let-it-cook/[email protected] is out. This works when running loops, but even when you just run `cook "do something"` which itself is not a loop, just a call to your agent.

Comment · 7 months ago

Exactly! That's my vibe. https://rjcorwin.github.io

Comment · 7 months ago

Good to hear that you're having luck with small models. Note that cook exposes a --model param, also workflow specific model params (--model-work, --model-review, etc) so you can have a smaller model implementing a plan and a larger model reviewing the implementation.

Comment · 7 months ago

I would be interested in which stories you are thinking of. Stories of Claude breaking out of the restrictions set in its sandbox or stories of people not configuring Claude's sandbox correctly?

Comment · 7 months ago

That's right. However if you use the v3 operator, you get three parallel versions being built, and then combined depending on which resolver you use (pick, merge, and compare).