@sandydoo Never looked at it this way 😃
The "smk" blues was a complain from an AV team. If you pay attention to the last years videos you will hear the clickers in the background.
I wish somebody would turn this clickers into something useful, like a proper one-key keyboard with mic.
We’re coming to NixCon 2026!
We’ll be in Kraków from September 25–28 to catch up with the #Nix community and share what we’ve been working on.
Come say hi at our booth, grab some swag, and catch the Flox team on stage:
🟣 @ronefroni Nix State of the Union 2026 and Nixing Artificial Intelligence
🟣 Yannik Sander: Adding to the Pile, or How to Extend nixpkgs
🟣 @garbas Flox lightning talk
See you in Kraków!
Full schedule: https://t.co/DKZw7oXlOr
@olvrgln Happy to show you what are we cooking at @floxdevelopment ... Would also like to learn what are you current problems to make sure we fix them for you.
We're going live with @temporalio on June 24.
James Bayer shows how a reproducible Temporal + AI dev environment goes from a pile of manual setup steps to one command: `flox activate`.
Plus more on sandboxing, security & observability.
📆 Jun 24 · 11am PT · https://t.co/Ycm9vuNeQ2
@semanticbeeng@floxdevelopment Also happy to dive deeper into your specific toolset. I'm also `garbas` in slack: https://t.co/64ro2CwPl3 ... feel free to ping me there and lets get you started
Yes, any combination of local llms and agents are possible. I'm going to assume few things but here is how I would set it up:
(create and go to a temporary folder for the flox environment)
$ cd $(mktemp -d)
(you can check the environment here: https://t.co/JdcjcPVGHk)
$ flox pull --copy flox/omlx
(optional: install some agent into the environment or use the onces you already have)
$ flox install flox/pi # or any other agent
(download the model from https://t.co/DWIgWl3AQG)
$ export HF_TOKEN="hf_xxxxxxxxxxx" # get it from here https://t.co/grcxsLC4pp
$ flox activate -- hf download "mlx-community/gemma-4-31b-it-8bit" --local-dir "$OMLX_MODEL_DIR/mlx-community/gemma-4-31b-it-8bit" # wait for model to download
(enter the environment with the omlx service running - you might need to wait few seconds)
$ flox activate --start-services
(start the agent from inside the environment)
(omlx) $ omlx launch pi
https://t.co/yRnXiq95n4 supports distributed teams across 100+ countries. But growth exposed familiar problems:
⚠️ Inconsistent dev environments
⚠️ “Works on my machine” failures
⚠️ Slower onboarding + rising release risk
With Flox, Fellow standardized reproducible environments across the SDLC without changing how engineers work.
Result: faster onboarding, higher velocity, fewer fire drills, and a more reliable path from dev → CI → prod.
“Flox gives us a dependable foundation so we can move fast with tight feedback loops.” — Samuel Cormier-Iijima, CTO & Co-Founder
Case study in comments.
I've spent most of my career building developer tools. And right now the whole DevEx market is changing in ways that are hard to ignore.
A senior engineer with Claude Code can prototype in a week what used to take a team a quarter. Companies that used to buy dev tools are building them in-house, and honestly it's becoming the rational choice. That's just one of the pressures.
But what I keep coming back to is something I didn't expect. The tools that lasted in my career were never the technically best. They were the ones where real trust existed between the people making the tool and the people using it. Not trust in code. Trust in humans.
AI can reproduce features. Someone can clone your tool with a test suite and an afternoon. But that relationship? Hard to prompt your way into. Maybe the moat was never the code.
The DevEx market is going to change. But we've been here before. This wave might be bigger, but nothing we can't handle. We're makers. Adapting is what we do.
https://t.co/c8HqsgujaJ
What do chefs and Nix engineers have in common?
We both obsess over a clean kitchen.
A good chef doesn't prevent burnt steaks by controlling every second. They keep the kitchen clean. Ingredients fresh, knives sharp, everything in its place. The creative parts figure themselves out.
I've been doing reproducible builds for 20 years. Locked deps, pinned environments, deterministic tooling. Got into it because "works on my machine" was driving me crazy.
Then AI showed up and something funny happened. All that plumbing turned out to be exactly what AI needed too. You can't make AI deterministic. But if everything around it is locked down, you actually know when something goes wrong whether it's the AI or your environment playing tricks.
20 years of clean kitchens and I was accidentally building the best AI strategy the whole time.
https://t.co/yrJlaXqyyk