simlevesque

Luca's announcement: https://bsky.app/profile/lcas.dev/post/3mhb4kavxd22y

Marvin's announcement: https://bsky.app/profile/marvinh.dev/post/3mhbtbl2ptk23

I use devcontainers in all the projects I use claude code on. [1] With it you can have claude running inside a container with just the project's code in write access and also mount a test folder with just read permissions, or do the opposite. You can even have both devcontainers and run them at the same time.

[1] https://code.claude.com/docs/en/devcontainer

If you want to try it just ask Claude to set it up for your project and review it after.

I think you'd like Arc Raiders.

I'm just wondering if there's a possible way to prevent this that wouldn't be intrusive or break existing features.

I love pentesting stories. Great blog post, I was smiling while reading most of it.

It reminded me of Deviant Ollam's stories such has his elevator security talk w/ Howard Payne: https://www.youtube.com/watch?v=oHf1vD5_b5I

What can Github do about this ?

Nah, the second you finish your build they release their version and then it's game over.

The first thing I recommend everyone using is devcontainers [1]. They're very simple to setup and make using LLMs a lot more secure.

[1] https://code.claude.com/docs/en/devcontainer

They want to be Claude so bad.

It's not much for a frontier AI but it can be a very useful specialized LLM.

No, Trump added "I will use the full power of the presidency to make them comply".

I'm pretty sure they can just prompt any convo in the background and ask "is this conversation sensitive ?" and the model can answer without this being added to the context of the convo.

Click on the link I provided and you'll know why I know. It's not markdown, it shows the code that was ran and the output.

It did run python code when I asked for a random number: https://gemini.google.com/share/dcd6658d7cc9

Then I said: "don't run code, just pick one" and it replied "I'll go with 7."

I recommend Eternal Terminal instead of mosh for this.

The paper: https://arxiv.org/abs/1909.10140

The first two I kind of understand what the author means. But "help me" and "severe pain" made me think that I'm just not the right public for this text.

Whitelist instead of blacklist seems like it would work.

It's been confirmed by @dang many times before. I'm not sure if that's what cut the title here but I've seen it many times in the last 10 years.

LLMs can't count. They need tool use to answer these questions accurately.

VSCode has vertical tabs for it's terminals like this, but on the right side.

> if all you need is time-related functions like stop watch, timer, multiple time zones etc.

But if you just need that, almost any watch will do. The Pebble is clearly not made for those people.

Why is the 9.2ms bar longer than the 104ms bar in the performance graph ?

If I've learned one thing in life: some people are totally shameless.

yeah but messenger.com was nice if you've blocked the facebook.com domain to try to use less social media.

I mean, some of the things on that homepage are truly the works of a genius IMO

For Go code I had almost no issue. PHP too. apparently for React it's not very good.

"/model claude-sonnet-4-6" works with Claude Code v2.1.44

"/model claude-sonnet-4-6" works with Claude Code v2.1.44

"/model claude-sonnet-4-6" works