← @bgSponge

Evolving programming languages in the AI era

Summary

José Valim argues that programming languages must shift from optimizing for human ergonomics to optimizing for agentic tooling, specifically through stronger guarantees and program databases.

  1. Agents weaken ecosystem formation.

    While agents reduce the cost of building frameworks, they may also allow developers to bypass existing libraries by generating bespoke, single-use solutions.

    If I need a library to solve problem X, I might just ask an agent to build exactly what I need.

  2. Syntax optimization is misplaced.

    Valim contends that focusing on syntactic affordances like optional chaining is a mistake because agents view code as simple tokens-in, tokens-out.

    any new programming language that claims to be made “for coding agents” and ultimately focuses on syntax is effectively building around today’s limitations.

  3. Languages must prioritize guarantees.

    Since agents can handle tedious tasks, languages should favor explicit types and formal proofs over the inference and ergonomics humans prefer.

    making types and intentions explicit gives the compiler, other agents, and ourselves more information to work with.

  4. Program databases replace LSPs.

    Valim suggests replacing the document-centric Language Server Protocol with queryable databases like SQLite or Datalog that agents can use to explore code.

    expose this information as a program database with a query language, be it SQLite, Datalog, or a custom DSL.

  5. Observability replaces debuggers.

    Instead of human-centric step-through debugging, systems should provide programmatic interfaces for agents to monitor and diagnose production state live.

    We should expose the runtime and state in our systems in ways that agents can query and explore programmatically.

Read the original ↗Save to your Sponge

Summarized with Sponge. Quotes are the author’s own words.

More from @bg

All readings →