Found multiple real bugs in #AshFramework from this, and one of my monolithic applications went from 2 minutes full compile to 1 minute. This is a huge win for our CI pipelines, our agentic flows, and our developer productivity. Every time they release a new #ElixirLang version, we get:
- zero or effectively zero churn/breakages of import
- faster builds
- more correct programs
#ElixirLang ride or die.
Elixir v1.20 released! Now officially a gradually typed language: Elixir type checks every single line of code, finding bugs and dead code, without developer overhead (no typing signatures) and extremely low false positives rate. Plus a faster compiler! Links and reports below.
π LocalCan 3.0 beta 2 is here with a Free mode!
Local domains (.local), and traffic inspection now work with no license - free for local-only dev. Enjoy!
Release notes in comments π
90% of the types vs tests debate is typing advocates thinking all dynamic languages are like JavaScript and dynamic devs thinking all type systems are OO-based.
Launching our new paper on arXiv: we trained the largest multilingual food model ever built.
4.1M recipes. 7 languages. 1,790 ingredients. 300 dimensions.
All of human cooking compressed into 2 megabytes.
Always loved lean frameworks for quick work and as such I'm really happy to share that me and @nelsonmestevao built a "Sinatra" for @elixirlang based on Plug and Bandit π§΅(1/7)
https://t.co/WsaA8lpGwO
This is more or less what the majority of our large dev team at work has come to the conclusion on too. Iβm not sure if itβs the model or the harness but codex has far surpassed Claude code now.
Everyone is creating agentic interfaces for their applications (or agents are just calling your APIs).
#AshFramework for #ElixirLang has always made it easy to safely expose behavior over external interfaces, and MCP is no different. But agents can also do clever things! And we use them to do more than just "POST /some/thing" at this point.
We want agents to have powerful tools, but they need to also be safe and consistent.
AshLua (https://t.co/b2vgZqMaHT) gives you exactly that. You can provide the agent with a lua sandbox to call the actions defined in your application, but the actor is injected into every call ensuring they have no way of escaping their privileges.
The next release of AshLua takes this even one step further, and gives agents the ability to use transactions π. Why have an agent make 14 API calls in a sequence just to try and clean it up later, or leave things in a dangling state?
There is a better way!
Spotted in the NYC subway. βZero screen time.β An iPod Shuffle ad in 2026.
When we built the iPod, the goal was the technology disappeared and you could have your music wherever you were. 1,000 songs in your pocket.
Now weβre living through a moment where people are actively looking for ways to disconnect from the infinite feed, algos, and constant notifications. That doesnβt mean technology is bad. It means the best technology understands when to step back.
Not every problem needs another screen, another menu, or another layer of complexity. Constraints create freedom (read: @DavidEpstein new book Inside the Box). And often removing features creates a better product than adding them.
The future of technology shouldnβt just be more engagement. It should help us be more human.