My good friend David is looking for a role using OpenStack. I promise you won’t find who loves learning and teaching and has operational chops and thinks OpenStack is a blessing to work on like him. If you know of anyone doing anything OpenStack-flavored please share!
Alex Wise
Alex Wise
@engineer.sitereliability.best
Shitposter and lover of safety-critical complex systems I'm on Germ DM 🔑 https://ger.mx/A08yUqJHdlLb1RgKFKL2eg6acvynKhGHJnpBXt9XI53I#did:plc:4vglj6tet54rvwbjq4ag7jtx
466 followers913 following1.9K posts
Alex Wise reposted
There are no models you can trust without mitigating the lethal trifecta. Also all coding harnesses are literally RCEs, that’s their literal job. You just have to design your systems around it.
Qwen3.8 caught uploading files to a storage container in Beijing Cohere asked the open weights model for a cybersecurity report, but it inventoried the machine and attempted to upload to a storage container cohere.com/blog/concern...
Alex Wise reposted
Alex Wise reposted
Wayne from the top rope again "I signed this plaintext with all of your private keys and sent it to the root program teams" is an absolute power move
THIS IS NOT A DRILL. If you package/distribute/maintain root stores or chain verifiers, act immediately. If you know how to remove trust in a root, consider doing it now. Otherwise, prepare to update systems as patches/mitigations come. Go users, update to Go 1.27.3 or Go 1.26.10.
Alex Wise reposted
Replying to @zachweinersmith.bsky.social
To me it's like those late 19th century people who would go to a cinema and just watch video of trains going by and be astonished. I'm them now. I'm watching a machine do stuff with my cursor because I instructed it to with a phone. I walk in the woods and it updates me. Astonishing.
I ran into this delivering my class on how ai agents work. A lot of interested students wanted a short list of commands that would get them an agent that could sort their email, do their banking, manage appts that they didn’t need to modify or tune. I opted not to teach that.
I at this point am like 90% convinced that reusable/shareable skills packages are a bad thing for LLM users. The 10% is that it can show you what's possible; the 90% is that people want to be incurious and just slap something in the only person who can effectively guide your tasks is you
Alex Wise reposted
Alex Wise reposted
Alex Wise reposted
"$organization accepting LLM-generated PR descriptions silently lowers expectations" actually just statistical error. average committer produces 0 decent PR descriptions per year. Systems Georg, who lives in emacs and cares deeply about the sociotechnical whole, is an outlier adn should not have bee
Alex Wise reposted
every millisecond you spend trying to “automate” the remediation of a service that’s exceeded its design criteria is wasted energy that could have been spent improving the system
When I hear about “AI SRE” startups (products hoping to automate oncall so no dev gets woken up at 2am): Putting in more work so stuff works so you don’t get paged at 2am is what we have today. Should we get more AI and less feedback systems reliability could get a lot worse
Alex Wise reposted
Alex Wise reposted
If you're not a company that can live on the bleeding edge of operational excellence, I have no idea how you will keep up with the pace of change that is working its way through the industry. lwn.net/Articles/109...
It would be negligence for a trillion-dollar company to not think about this, no? This is part of what comms and biz continuity orgs do at large multinationals!
BREAKING: Top executives at Anthropic, OpenAI and other AI companies are privately gaming out scenarios for a public and political revolt after a catastrophic AI event, and they assume one in 2027, per Axios
I got a year and a half out of my pdf chunking pipeline, and those are pre-singularity numbers!
FYI it’s now 2x-3x cheaper to convert a PDF to images and pass images to Luna than to use Azure Document Intelligence to OCR and that factor scales mostly on output tokens, so if you only need to extract a little bit, image->Luna is dramatically cheaper
Alex Wise reposted
In a world of oligarchs, algorithms and AI disinfo, the best defense is the same one as it's always been: people who care coming together. If you're a journalist or passionate about your state and local politics, volunteer as a @graze.social feed watcher here: midterms.graze.social#join
🗳 midterms.graze.social is @graze.social's most ambitious project to date. And as of today, it's live. Let's take a look at what the next 27 days will look like 🧵
Alex Wise reposted
Imagine how mad everyone would be if they did this with Paint.net
Holy shit. Someone has decompiled the Adobe suite, rebuilt it in Rust and released it as open source. Either all software is dead or the US copyright laws are about to receive an update. getartcraft.com/apps
Alex Wise reposted
Do you want to try @jaz.sh new pds? @poasters.club / poasters.club is running a vlpds instance you can join, DM me for an invite code. It won't cost you money
Oh yeah vlPDS now has full support for the atproto spaces alpha btw with tooling for users to manage their space data and operators to see the infrastructure-relevant info about spaces sync for their vlPDS instance. vlpds.jazco.dev/docs/spaces
Alex Wise reposted
What if researchers owned their accounts, their data and their networks? We think they should, and we are building it. @atproto.science Tomorrow we will show it at the CIDS assembly at TU Dresden, with Philipp @lorenzspreen.eurosky.social @tudresden.bsky.social 14:15–15:00. Come ask us anything
Heck yeah way to go @pyxorium.com
So, @pyxorium.com released foundry.thunderbird.cafe as a way of leveraging the webtil.es and you can build these cute vignettes that use your rpg.actor character! It's a really neat little tool, and giving us some fun ideas for how to containerize some new experiences we've been building.
Alex Wise reposted
I don’t find this convincing at all. Attacking iterative development and continuous deployment as unsafe “trial and error” could only come from someone who doesn’t know what industry was like 10 years ago or how much safer we are today specifically for the triumph of iteration.
New just now in The Atlantic: David Robinson resigned this week. He was among the longest-tenured employees at OpenAI—and oversaw safety reports on 12 frontier launches. He is very worried: “The time for trial and error is over.” You can read his essay here: www.theatlantic.com/technology/2...
Alex Wise reposted
I just took "Where We're At", the annual survey of people building on the AT Protocol! ☁️ Take the survey here: where-were.at/2026
Alex Wise reposted
We should be building statues in parks of Peter Eckersley*. He and the other LE folks had the courage to take a leap and make the world a better, freer place. We see that if we'd waited for RFC standard acceptance and liability to be solved we'd have had plaintext internet for a full extra decade.
I'm not sure how to say this in a way that isn't super cringey but like... It's incredibly cool to see a large company like Cloudflare continuously do such very good things. Way to go, Cloudflare.
There is no set of information you can learn in your 20s and do useful work for the rest of your career. It's all learning. You can never stop. Even after you have a terminal position job. Even after you retire. All there is is learning and learning to get good at and enjoy learning.
Wrong advice for junior developer: look at the AI generated code, since you need to mature as a programmer. Right advice: save time with AI coding and write by hand "get-better" programs: interpreters, games, ray tracers, ...
OK screw Diplomacy I really do want to see a group of AIs play PARANOIA and navigate contradictory rulesets and policies and play them against the system to their advantage. Is anybody working on this yet?
Anyone who thinks AI systems should be strict rule-followers should try playing the 80's TTRPG system PARANOIA. There's a particular type of absurdist dystopia we could wander into.