Alt text:

Slrpnk.net is currently offline due to an unforseen hardware failure in combination with the main system-administrators having no physical access to the server location until mid July due to work or summer-holiday related travels.

We are very sorry for this unforseen down-time, but slrpnk.net will return for sure and we already have some plans for a nice relaunch, so stay tuned!

Their HA infrastructure is all built on open source projects. The thing they have that we don’t is teams of SREs on-call 24/7.

So they claim.

Looking through the commit history there are numerous “Manually fixed…” commits, where the LLM doesn’t do what the programmer wants after repeated prompting, so they fix it themself.

And here is the problem. It required expert supervision for the prompts to be repeatedly refined, and the code manually fixed, until the code was correct. This doesn’t save any labour, it just changes the nature of programming into code review.

If this programmer wasn’t already an expert in this problem domain then I have no doubt that this component would be full of bugs and security issues.

Meaning it’s really just running in their closet,

You think they’re lying about having their hardware in a colo?

maybe it’s finished

The fundamental difference is that hash functions are designed to be irreversible (one-way functions), whereas encryption is designed to be reversible (where the inverse operation is called “decryption”).

You should treat security questions like passwords and use strong alphanumeric passwords as the answers. Just make sure to store them in your password manager in such a way that you can remember which one goes with which question!

You don’t want your password to be encrypted, you want it to be hashed.

I think you’ve misunderstood what AGI means.

If AGI is made of components, you could argue that it isn’t “general”.

Can you elaborate? I don’t get what you mean by this.

It depends on the provenance of the code and who (if anyone) is downstream.

A project that’s packaged in multiple distros is more likely to be reliable than a project that only exists on github and provides its own binary builds.

I remember having that 8.10 Intrepid Ibex wallpaper for quite a while before I went looking for some custom wallpapers.

Slightly off-topic, but I hate that this website only loads the images after I scroll to them. Is that the website doing that, or my browser?

I spent ages looking at one near the top and in the meantime, my browser could have been loading all the images further down the page. But instead it didn’t, and so when I scrolled down, now it has to load each image as I reach it. I estimate about 16MB of images in total, since each thumbnail is actually the full size image.

Interesting, thanks!

Thanks for the reply.

Currently, most packages are built from git HEAD on alpine:edge or debian-unstable build containers. So if the fix for this affected libwebp is shipped to the images that the build containers are based on (likely because we use edge/unstable images), then any affected packages would also automatically receive this fix.

How often do packages get rebuilt? Is it only when there’s a new version? The problem in that case would be that a package that is no longer developed (or has very long release cycles) would not receive the fix.

I think we can make an exception for soup and ice-cream, no?

Does it store a complete dependency graph for each of your statically built (or containerized) applications?

For example, if there’s an exploit for libwebp and you need to update all the binaries that link it, can it find which binaries need updating from that information?

💩

Built on IPFS

true censorship resistance

Last time I checked, IPFS was not censorship resistant.