Andri Óskarsson

I've been struggling with managing an ever increasing amount of balls in the air and GitHub's UI hasn't been very helpful. So, after spending a week with Paseo, I decided to scratch my own itch and build a "did, doing, waiting, next" board for myself as a Paseo plugin.

I've been struggling with managing an ever increasing amount of balls in the air and GitHub's UI hasn't been very helpful. So, after spending a week with Paseo, I decided to scratch my own itch and build a "did, doing, waiting, next" board for myself as a Paseo plugin.

These series of dangerous vulnerabilities is scary for sure. But, the really scary part is that state sponsored intelligence services have been playing with these models for a long time now. With less restrictions.

Humans with AI are lazy. So, if humans no longer create requirements, wireframes, user-stories or tasks. Then how the hell are we going to build products or projects and feel somewhat in control? I think I have a solution. github.com/andrioid/req...

I got the idea early this year and have been testing and iterating on it for a while and I think that others should be using it too. - Iterating on a small requirements.yaml file is much faster than iterating on a codebase - Brainstorming potential products - Discover requirements w. AI

Why is the TLS leak critical? I've been asked. Our ability to run banking, payment, etc online is highly dependent on our ability to encrypt traffic from the user and the server without network operators (or hackers) peeking inside.

With a leaked private-key for a trusted certificate-authority (CA) anyone can create a trusted certificate for any domain. That means unknown networks like airports, cafés, etc now technically have the ability to read all your encrypted traffic without you knowing it.

Sitting here, after hours, watching grown people comment spam the PRs for the Docker PHP image. As a result. They locked comments - which failed the release CI script. Thanks. Thanks so much.

If you're running anything PHP based, consider taking it offline until PHP patches become available today at 14:00 UTC cert.europa.eu/publications...

Exploits are out. Patches are not. Unconfirmed mitigation is disabling session.upload_progress.enabled in your php.ini x.com/k_firsov/sta...

Anyone else noticing considerable more delay in GitHub's UI recently in Firefox? I'm observing seconds long delays in both rendering and typing. It wasn't great - but now it's bordering on unusable.

Experimenting with Firecracker microVMs at work for developer environments. So far I have: - Developers login with GitHub and get a box - Box is reachable via SSH and will auto-start - Box can be stopped/recreated from UI Each box - 32G ram - Mise with tools/agents pre-installed - Paseo relay

Non-goals: - Multiple host-machines or a cluster scheduler. I'm not building Kubernetes! - Anything that unnecessarily increases the exposure of the VM manager API

Experimenting with Firecracker microVMs at work for developer environments. So far I have: - Developers login with GitHub and get a box - Box is reachable via SSH and will auto-start - Box can be stopped/recreated from UI Each box - 32G ram - Mise with tools/agents pre-installed - Paseo relay

Host machine and architecture - Nobody needs to access it - VM API is a non-root user w. KVM access - Management CLI for setting up systemd/firewall - Services run via SystemD - Web interface is separate from VM management Stack is Go and a little SQLite

I've tried migrating my blog to a CMS multiple times, but always end up going back to Markdown files. Then it occurred to me that my PDS is the perfect place to store this stuff. I don't need to host anything, but I can if I want to. So, I'm working on "atcms" in my spare time. #atproto

I never built a CMS in the 2000's like everyone else. So here goes 🙃 - Browser only (no backend to run) - Syncs through service worker - Built-in layouts for CMS related collections - Supports all lexicons (fallback if no layout) - Asset previews/browsing

Several sources are reporting that the password used to leak 8 million Danish social-security numbers was "123456". Most people who have worked in IT for a long time have at one point or another been sloppy with passwords. But, I'd hate to be THAT person right now.

I have mixed feelings about using LLMs for programming. Dislike one-shot slop (except for prototyping); but digging into legacy code with agents is just wonderful. Asking it to find patterns, anti-patterns or seams. Discuss idiomatic code and present it in HTML slide-decks is incredibly valuable.