Hanif Leoputera
a curious kid with GPUs and too many tabs open
hanif@leo:~$ cat about.txt
Hi, I’m Hanif. I’m a research engineer at Tacit in San Francisco, building wearables that enable silent speech. Before that I was at Affirm. Mostly, though, I’m curious about whatever sits underneath things: the chips under the models, the power under the chips, the physics under the power. This is where I write it down.
hanif@leo:~$ ls -lt writing/
- 2026-11-02generative-interfaces.mdAI is starting to draw interfaces. I've been wondering which part of the system should decide what you see.
- 2026-09-08post-hot-chips-2026.mdA tour of custom AI chips, and why the interesting comparison is speed per user against throughput, not FLOPs.
- 2026-07-20can-llms-do-science.mdA review of how Periodic, Lila, DeepMind, FutureHouse and academic labs are trying to make language models do science, and what still stands in the way.
- 2026-06-15data-market.mdWhy AI training data looks less like oil and more like fresh produce, and which parts of the market are worth building or backing.
- 2026-03-16speech-to-speech-survey.mdWho has built what in speech-to-speech, and how it works: four kinds of full duplex, codecs and continuous features, token factorizations, turn-taking decisions and thinking while listening.
- 2025-12-15everything-is-evals.mdA survey of formal proofs for AI-written code, from Lean, Dafny and Verus to Bend and the new AI provers, and why the proof has stopped being the hard part.
Click a post to read it. On a keyboard, ↑ ↓ and enter work too.
hanif@leo:~$ cat contact.txt
- hanifleoputeralim@gmail.com
hanif@leo:~$