Sure they will keep C/C++, and various low-level code: Swift is nice for developers but slow for execution (compared tose).

Try to get banned from any of these, or from the banking system, and find out

Not a hypothetical

https://lavialibera.it/en-schede-2447-francesca_albanese_und...

Everywhere I’ve ever worked, there was always some way to access a production system even if it required multiple approvals and short-lived credentials for something like AWS SSM. If the user has access, the agent has access, no matter how briefly.

Not if you require auth with a Yubikey, not if you run the LLM client inside a VM which doesn't have your private ssh key, ...

Supabase virtually encouraged it last year haha. I tried using it once and noped out after using it for an hour, when claude tried to do a bunch of migrations on prod instead of dev.

https://web.archive.org/web/20250622161053/https://supabase....

Now, there are some actual warnings. https://supabase.com/docs/guides/getting-started/mcp#securit...

I think LLMs are exposing how slapdash many people work when building software.

If the folder is versioned and commited regularly there is no problem. It also allows you to open the files in your IDE, do some other tasks or fixes for claude. It prevents claude from accessing any other folder, which is the idea of the post.

Until Claude nukes .git, assuming you're using git as the version/commit store. Solution use easy, just push to a remote on a reasonable cadence (that you can run reflog on, so a force push won't eat your data either). Git isn't backup though, it's a VCS, and those are two different things, even if they are somewhat alike.

I’ve seen Claude rm .git in rare occasions to “fix rebase hiccups”

Version control ain’t a match for a good backup

So? if it removes .git, just clone the project again and you are ok

In that case, you're just relying on your remote backup :-)

This breaks the non-interactive mode the post want to achieve. Claude will not be able to install some things and will require user action, which is not desired here.

Like what? It can already use npm/pip/etc. And if it needs a new APT package or config in /etc/ then you would want to know because you need to document it.

Claude Code on NixOS feels like it has super powers. Being able to spin up a nix-shell with needed dependencies on demand gives it access to all sorts of tools I don't have or want installed on my base system. My "book-recommendation" claude code uses sqlite to manage my reading history and to-read and maybe-read lists but I never installed tools for sqlite and they aren't present on my NixOS desktop. It just launches a nix-shell with sqlite anytime it needs to read/modify the database. As long as the database file is within the directory claude code was launched from, it doesn't need to prompt for permission. With the caching that NixOS does, it's fast enough to not even think about.

If you make claude work with c/c++, it may need apt for libraries or build tools.

Even with npm/pip, these may not be available on a base linux box.

Even then, some complex projects may need other tools that are not part of a base system (command line tools, redis, ...).

How would nuclear deterrence work for small entities like Denmark or Taiwan against huge entities like US or China? it only works at similar sizes

> How would nuclear deterrence work for small entities like Denmark or Taiwan against huge entities like US or China? it only works at similar sizes

It works as long as the harm that can be threatened is sufficient to outweigh any perceived gain of winning. Small states may not be able to sustain as large of an arsenal, but they also rarely offer as much value to a victor.

A nuclear deterrent is still a deterrent, no matter how small. No country (hopefully) wants to risk any kind of nuclear war. Ukraine would never have been invaded if it still had its nukes.

We cannot know. My best guess is that at some point in the future there will be a military conflict between two parties that have nukes. Pakistan vs India for instance. And although they have nukes they would fight conventionally unless one party is about to lose.

> And although they have nukes they would fight conventionally unless one party is about to lose.

In every war, eventually, one side will be about to lose.

It made North Korea essentially invulnerable.

If absolutely necessary you could take them out in an preemptive strike. They have no second strike capabilities such as SLBMs.

It is still risky of course and not advisable.

They could use it like Google Search, not as the first thing the user sees, but as a fallback

I get why you may not close <img> or <br> since they don't contain anything inside, but <p> and <li> should be closed to indicate the end of the content, otherwise it's shows you are mentally lazy and relying on some magic to do the work and guess what you wanted

To each their own. In simple lists for navigation menus I always omit closing li. There is no ambiguity on what I am intending even with those closing li omitted in such simple cases:

  <nav id=main-nav>
  <ul>
  <li><a href="/">Home</a>
  <li><a href="/hamburgers/">Hamburgers</a>
  <li><a href="/sausages/">Sausages</a>
  </ul>
  </nav>

<p> indicates a new paragraph. <li> indicates a new list item. Unless otherwise specified, the existing paragraph/list item continues. There's nothing magic about any of this, it's part of the HTML spec.

Laziness doesn't play a role. This isn't XML where you need to repeat yourself over and over again or abusing a bug in the rendering logic; it's following the definitions markdown language you're writing content in.

If you're not too familiar with the HTML language then it's always a safe bet to close your tags, of course.

The personal judgement isn't really helpful here?

If you don't close your <p> and <li> tags, you risk accidentally having content in the wrong place.

It's something to avoid because it can have bad consequences, not because it (somehow?) makes you a bad person.

Run it inside a VM, make snapshots of the VM if needed (or use vagrant/ansible to rebuild), commit regularly, ...

The VM still needs access to the network for the use cases they described though.

That seems incompatible with the parallel tasks of cleaning and cooking (at least for me, especially with kids around).

The VM is setup once, before you get to be "on the go": that's your development environment, you need one anyway

We don't need that to be pre-loaded on the device, do we? how often are they used? and when one is set to be used, the user is definitely online already, just grab the one then