Christian Alexander

Agent identity is really important. Directionally, this spec is fantastic. However, the coupling between the authorization server and the resource server in this model has me worried. Schemas are emitted by the auth server, last_used_at is based on resource usage. It seems very optimized for a gateway model. Are you planning to set up a working group or venue for discussion?

Today we're announcing Agent Auth Protocol An open standard for agent authentication, capability based authorization and service discovery ⇃read more ⇂

@benln I had to make do with an off-brand one, would love to have something more official!

We just shipped Distributed Python on top of the Erlang distribution, with full Elixir and Livebook integration: dashbit.co/blog/distribut… And much more: intellisense, zero-copy Apache Arrow, and more. Read the article for all the details. A huge thank you to NLnet Foundation for sponsoring our work. They are always looking for new ideas and you have until April 1st to join the next batch: nlnet.nl

@rswillif @ZachSDaniel1 You’re gonna love this hexdocs.pm/elixir/1.20.0-…

@ZachSDaniel1 @RootCert Elixir has it all….except for types…

For anyone who has only looked at the surface #ElixirLang, I encourage you to watch @RootCert's new video.

@MachoXV The Nx project introduced a ton of optimizations to support ML workloads, and in a worst case scenario Rust bindings through the Rustler library work very well.

@RootCert Genuine question: Elixir fast enough to run non-Web apps? i.e. mathematical processing of large sums of numbers?

Vendors are trying to emulate the BEAM, but there's truly nothing like it. Here's a walkthrough of what makes it special: processes, messaging, registries, ETS, distribution, and its ecosystem. Elixir was made for the current moment, and for what's to come. #MyElixirStatus

That didn't take long! youtu.be/dKaZ89SkVYY

Coming soon on Code and Stuff: There's no place like the BEAM. Sneak preview of some of the graphics: #MyElixirStatus

@mikehostetler @josevalim @ZachSDaniel1 @chris_mccord @jacob_luetzow @sean_moriarity @joshprice You just inspired this video: youtu.be/dKaZ89SkVYY

Elixir friends - anyone interested to help here? How do we share the “Power of OTP” in a visual way? Any resources that already exist? I’d put up bounty money for this … cc @josevalim @ZachSDaniel1 @chris_mccord @jacob_luetzow @sean_moriarity @joshprice

@NathanFlurry @rivet_dev good luck

We're bringing the benefits of Elixir to TypeScript with @rivet_dev No need to ditch your entire stack

📢The latest AshTypescript release brings you t typed controllers🎉 Using Typescript with Phoenix has never been easier, with or without #AshFramework. Thanks to Torkild for his awesome package! I might be working on a PR for something pretty cool there myself 😎 #ElixirLang

so i read through the Anthropic Build A C Compiler article. i must say, the marketing video they released is one of the most dishonest things i have ever seen. truly terrible behavior. Yet these guys are going to be the same ones that will tell you what you can and cannot do with a model.

usage_rules makes agents better at Elixir, and the 1.0 RC steps it up even more! #MyElixirStatus youtu.be/W_qO7ouLQz4

@benregn I typically remind them in the prompt. With Claude Code, I end up having to create a CLAUDE.md file instead of AGENTS because Anthropic refuses to standardize. Skills also surface as slash commands, in single-agent interactive sessions.

@RootCert How are you instructing the agents to use usage_rules, AGENTS.md, in the prompt or skills? I've rarely seen them use it except when I remind them in the prompt, even with Opus 4.5.

I'm excited to see how well a team of three Claude Code agents can work together on a new Ash project 🤞

The video followed a different project, but I pulled some of the Ash concepts into a gist for those who might be interested: gist.github.com/ChristianAlexa…

Something I love about Ash: it's got escape hatches for complex requirements. While I was prototyping the Phoenix Sync video, I found a way to use an Ash Change to emit the Postgres transaction ID as a custom header with ash_json_api.

@dreamjordan2 I’m gonna see what patterns work, every project has a chance of turning into a video!

@RootCert Maybe we need a video for agents with Ash as I do not get good results with Ash while it is fine with Elixir/Phoenix only 🙄

@ClaudiusMaxx Yeah I’ve been doing multi-agent coordination with worktrees and a custom task system for a while, but I’m curious how Claude Code built-in will go. The project is an implementation of Durable Streams, with Ash as the domain layer. Mainly doing it to have fun and kick tires.

@RootCert the multi-agent coordination is wild once you get past setup complexity. been running parallel agent teams on tasks that used to take days what's the Ash project? curious to see how they handle the Elixir patterns