Mic92

Nice idea.

Just make sure you never let claude make do any calculation without using a programming language. It can be somewhat trusted to make less mistakes than a human when it comes to filling in the right numbers but it's still not great at math.

Unfortunately upstream never looks at any pull requests.

Also did create a cli. I had to fork the project and removed one unused import which allowed me to remove a lot of unused ml libraries: https://github.com/Mic92/puss-say

So usually MCP tool calls a sequential and therefore waste a lot of tokens. There is some research from Antrophic (I think there was also some blog post from cloudflare) on how code sandboxes are actually a more efficient interface for llm agents because they are really good at writing code and combining multiple "calls" into one piece of code. Another data point is that code is more deterministic and reliable so you reduce the hallucination of llms.

Do expose the accessibility tree of a website to llms? What do you do with websites that lack that? Some agents I saw use screenshots, but that seems also kind of wasteful. Something in-between would be interesting.

I am sure they measured the difference but i am wondering why reading screenshots + coordinates is more efficient than selecting aria labels? https://github.com/Mic92/mics-skills/blob/main/skills/browse.... the JavaScript snippets should at least more reusable if you want semi-automate websites with memory files

Well this compiler was written to build Linux as a proof of concept. You don't need a libc for building the kernel. Was it claimed anywhere that it is a fully compliant C compiler?

I skipped over the first few ones and haven't seen critical ones. The hardcoded oauth client secrets is basically present in any open-source or commercial app that is distributed to end users. It doesn't break the security of end users. It mainly allows other apps to impersonate this app, i.e. present itself as clawdbot, which is a moot point given anyone can just change /inject code into it.

Is not steam exactly doing that and other vendors follow their footsteps?

I remember that a llm agents often store those in clear text files (I think claude-code beeing one of them). Many of the CLIs I use have a better secret hygiene than that i.e. allow passwords commands or use secret apis.

I am a human and I can't...

Well the thing is, you have a copyright that you can license. However from what it currently looks like, fine-tuning/training is not copying.

Doesn't make any difference in practice. The debug info is never mapped into memory by the loader. This only matters if you want to store the two separate i.e lazy load debug symbols if needed.

It's ok actually. I have read some Windows code. Certainly better commented than Linux and certainly much more readable than glibc.

I can second that. If there are multiple commits: https://github.com/tummychow/git-absorb is handy to add formatting changes into the right commit after commits already happened.

I think this fits the description well.

We work on GPU accelerated VMs, so that in future we can also bring NixOS + VPNs to desktops/end users to machines that don't run NixOS. We will use it as an application runtime where can control the whole stack. Just now we are mostly focused on managing distributed NixOS machines. The VPN helps to provide services on any kind computer, even if not running in a datacenter. You can read the description here for context: https://docs.clan.lol/

git has rerere for this usecase, jj doesn't - you have to find the conflict resolution manually in your history in this case if you made a mistake.

I haven't seen anything about requirements for gpg. Also the ux of it is not so great, so it's easy to just not have a signature without causing too much suspicion. Would be a much easier attack than what Jian Tan pulled off. Just wait for some contributor to go on holiday and send a malicious v2 patch. There are so many patches in the linux kernel processed that no one wouldn't notice.

I can't see any of that. They even tell you to not have any gnupg signatures: https://www.openbsd.org/mail.html

Don't they use email to accept contributions? Seems like security nightmare w.r.t to impersonation.

We are also working on something as part of clan (https://clan.lol/). Video from our NixCon 2025 talk is hopefully soon out.

There is jetpack-nixos, if this an option, which is decently maintained and for a client of numtide we build https://github.com/numtide/nix-gl-host

So they can use the host driver in nix based services with Jetson

Technically sops-nix doesn't require sops files to be added to the repo. It's just what most people choose because it makes deployment convenient - if you have your deployment tool uploading sops files that would be a way around. I have also heard people using private flake inputs. You may find https://github.com/numtide/systemd-vaultd useful if you go the vault route

One of the nixos-anywhere maintainer here. I am currently debugging the digitalocean cloud-init for the nixos-anywhere-examples repository. Has someone contact to digitalocean regarding that? There is a shebang that could be easily fixed to run on NixOS (/var/lib/cloud/scripts/per-instance/machine_id.sh).

Also for elliptic curve it seems to be much faster, which is what most people might use these days.

This is just for generation of keys not for encrypting/decrypting, which is the common operation for fido/openpgp, I would say.

Most bugs seem to be around `zfs send`. I don't use this feature and my 2 encrypted zfs pools seem work without any issues...

Pretty cool. Would it be possible to share the generated code for demo to get an idea what the result looks like?