Case Studies

Cognition shipped Code Scans on September 16: a plan/shard/map/reduce pipeline that turns a vague engineering goal like ‘cut our build time’ into a batch of pull requests. Cognition’s own numbers claim a 96% PR merge rate at Philips and a 64% Rust build-time cut — but the pipeline runs on SWE-2, a model post-trained on someone else’s base, the same open question this blog has been asking about Cognition since its Series E.

Cognition engineer Eric Lu, working with multiple Devin agents, factored RSA-260 — a 260-digit number unsolved since the 1991 RSA Factoring Challenge — using a GPU-accelerated rewrite of the CADO-NFS sieving toolkit that Devin designed and operated with only periodic human steering. We independently re-multiplied and primality-tested the published factors ourselves: the claim checks out. RSA-2048 is unaffected.

Five production engineering teams — from a 23,000-person Latin American e-commerce giant to a 50-person legal AI startup — show what spec-driven development looks like at real scale. The numbers: 90% autonomous code generation, 6x task completion rates, 500K engineer-hours reclaimed. The patterns behind the numbers are more interesting than the numbers.

You gave Claude Code to 40 engineers. Now everyone’s writing their own prompts, their own workflows, their own shortcuts. Here’s how one team turned that chaos into a shared skill marketplace — and what they learned building it.