Your compiler trusts you. It shouldn't. C# compiles a recursive loop that will slowly shove your database off a cliff. No warnings, no errors. Lean refuses to compile the exact same logic until you prove it terminates. 🧵

For every deviation between code and spec, three options: → Fix the code (scoped prompt, can't break other contracts) → Update the spec (if your intent was different) → Remove the spec (if it no longer applies)

This is the moment that matters 🤘 Syncs each requirement to the code that satisfies it, generates theorems, and proves them. Every theorem that doesn't hold comes back with the exact input that breaks it.

A local viewer to review every inferred contract before anything else runs. If a requirement looks wrong or too strict, you correct it here first. The formal check will then run against the contract you actually want.

Step 1: @PredictableCode reads the codebase and infers the behavioral contracts that the code itself implies it should satisfy. No specs written by hand. No annotations. Just the code and documentation, and what it formally promises to do.

The one question I keep coming back to with AI-generated code: does it actually do what I specified? Our first @PredictableCode case study answers that end-to-end on a real Java codebase, with a full recording of the workflow. predictablemachines.com/blog…

We'll be at @seattlefloworg Startup Day 2026 on May 15th, demoing @PredictableCode at the Startup Fair. Come see what it looks like to formally verify AI-generated code. Not "passes review," actually proven correct. 👌 Tickets 👉 seattleflow.org/startupday

AI writes the code. AI reviews the code. A developer clicks "merge" in under 10 minutes. In that single click, all the legal and regulatory liability concentrates in one human, for code they didn't write and couldn't fully read.

Every team has a doc that's quietly become a lie. The README wasn't updated when the API changed. The Notion page from six months ago. We used to shrug at it. Then we started pasting it into Claude as context for the next feature.

Code that runs, compiles, passes tests, and still isn't doing what you asked. Same requirement, three AI sessions, three incompatible implementations. The second piece of our pain points series is up. predictablemachines.com/blog…

Ten developers, ten AI sessions, ten slightly different ideas of what the codebase is supposed to do. First of a short series on the pain points we're solving with Predictable Code. predictablemachines.com/blog…

We just open-sourced lean4-json-schema: derive JSON Schema from Lean 4 types, with a compile-time proof that serialization always validates against it. No more trusting. The compiler checks it. 🧵 predictablemachines.com/blog…