Smarter Coding Agents for Rust with Symposium

4 Просмотры

Спасибо! Поделитесь с друзьями!

Вам не понравилось видео. Спасибо за то что поделились своим мнением!

Издатель Oct 3, 2026

AI coding agents promise faster development, but they often lack the context needed to write idiomatic and reliable Rust. At the same time, Rust's rich type system, compiler diagnostics and ecosystem tooling offer a powerful foundation for making agents smarter.

Join Orhun Parmaksız, Niko Matsakis and Jack Huey to explore how Rust and AI can make each other better. They'll discuss how agents can help developers understand compiler errors, navigate ownership and type-system challenges and move from diagnostic messages to informed solutions. You'll also learn about Symposium, an open source, vendor-neutral interoperability layer that equips coding agents with crate-specific skills, tools and workflows based on a project's dependencies.

Whether you already use coding agents or are curious about their role in Rust development, this livestream will offer a practical look at building more capable, efficient and reliable AI-assisted workflows.
#Rust #AI #LLM #Agents #OpenSource #developertools

00:00:00 Welcome and introductions
00:01:56 What sparked your interest in AI coding agents?
00:06:44 Can Rust and AI make each other better?
00:09:48 Why aren't compiler diagnostics alone enough?
00:20:02 What problem was Symposium originally trying to solve?
00:25:29 Which protocols does Symposium depend on?
00:29:37 What are crate-aware plugins?
00:34:22 Can't agents get the same context from docs or source code?
00:39:32 What information does Symposium pass to coding agents?
00:41:54 How does "Rust Token Killer" work with Symposium?
00:45:33 How does the recommendations repository work?
00:49:04 Who decides what gets added to the recommendations?
00:49:51 How would you describe Symposium in one sentence?
00:50:27 Live demo
01:06:36 What about prompt injection and supply-chain attacks?
01:10:18 Is that the whole setup, or do I need anything else?
01:13:45 Do I need to run sync, and what does it do?
01:15:38 What does "cargo agents use" do?
01:17:52 Can we see the skill working in real time?
01:19:36 How is the Rust best-practices skill updated?
01:23:50 How do you evaluate and benchmark skills?
01:26:34 How do you tell whether agent-written Rust is good?
01:29:12 Could this model work outside the Rust ecosystem?
01:34:43 Could better agents make Symposium obsolete?
01:41:59 How will AI change the role of open-source maintainers?
01:50:32 What's next for Symposium?
01:57:04 Wrapping up