Amplifypartners

An introduction to reinforcement learning in imagination, AKA generating synthetic environment data via world models.

Why every software company is now a dev tools company.

For all its quirks, the industry has (mostly) converged on Git for version control. But what happens when agents are writing most of our code?

Types and theorem provers: why formal verification matters in the age of AI code.

The story of Gradium, the anatomy of audio AI models, and why smaller labs continue to edge out larger ones when it comes to voice.

How CedarDB reimagined every piece of the relational database like code generation, a SOTA query optimizer, and design for modern hardware.

A story of real context engineering: how Hightouch's engineers built a long running agent that smoothly handles sparse and open ended marketing tasks.

A primer on Deterministic Simulation Testing, the testing framework that's taking the systems world by storm, plus how Antithesis is bringing it to the masses.

Why we should care about using autocurricula to train digital agents and how unsupervised environment design and evolutionary algorithms might help us get there.

Behind the scene's of Datology's incredible data infrastructure, from clever implementation of deduplication to their custom versions of Spark and Flyte.