Why & how we rewrote Bun from Zig to Rust

I’ve always liked to think of the A in AI as augmented instead of artificial. Humans plus AI instead of versus. This is hardly controversial if you look at it in terms of any technology invention: cars made horse buggies obsolete. EVs and solar panels are making petrol engines go the way of the dodo. And in every case, the buggy drivers suffered and had to reinvent themselves, but the human quality of life improved.

The problem I've been interested in is this: if you have content or principles that you'd like to learn from a book, is it possible to build an interactive agent that can help you grok that content easier? In Part 1 I built a Socratic programming tutor as a CLI agent on Flue. It ended on a dare. The tutor was a terminal app — an 

YouTube video by Apple Developer

In 1984, Benjamin Bloom published a finding that has haunted education ever since: students tutored one-on-one performed two standard deviations better than students in a conventional classroom. The average tutored kid beat 98% of the untutored ones. The catch—the reason it's called the 2 Sigma Problem—is that one-on-one tutoring never scaled. We've known the best way to teach for forty years and could never afford it.

TLDR: I've rebound the space key so that when held it behaves like a combo of Ctrl+Option+Cmd+Shift (the Hyper key) using kanata, and then bound Space+letter combinations using Raycast to launch and switch between apps instantly. Why? Once in a while I make a change to my workflow that feels like it's worth blogging about. This time, I was inspired by keyboard workflows that primarily seem to come from the Linux world…

Last weeks: trip to Japan, Chronicle news, and links.

Last week: switching back to Cursor AI, Alapuzha Gymkhana, Path to Transcendence & the Japan visa.

Chronicle combines AI with taste to help you go from raw thoughts to polished presentations. Designed for serious work: Start with templates, create with AI workflows, and collaborate with your team t...

Last week: electric vehicles, configuring Zed, a CD player, visa application for Japan & movies.

Last week: migrating vishnugopal.com, book quotes, reading glasses & AI and React stuff.

Last week: Thunderbolts, visiting temples, discussing the war with friends, and working on features while managing health and diet.

Last week: a houseboat trip, impending launch at Chronicle, a blog transition plan & personal fitness progress.

Last week: two movies, editing Indian skin tones, Nordic design, and an old game made new.

Last week: Talekat, visit to Jatayu Park, Pravinkoodu shapu & side projects.

Las week: family from north India, XState musings, small world adventures, and getting fish.

Last week: reflections on language learning, Achan's birthday, a Beatles' visit, Rikuoh, and the kutty thevangu.

Last week: new job at Chronicle, workouts and diet, photography experiments, and a couple of movies.

TLDR: I find current AI-enabled editing interactions to be a bit of before-after magic. You click a button & wham: your entire text changes. I built an AI editor called Draftflow that works more like "another person in the room" in a collaborative environment. There is a video below in case you don't want to read much. The Problem: Gen AI-enabled Corrections…

I built an AI cowriter and director called Coweave to help anybody write interactive fiction, a task that until now required some knowhow of programming.

TLDR: I learnt about Apple’s new Live Caller ID lookup feature last month, and I built out a simple implementation of that in TypeScript called sharded-seal-lookup. This post explains how the…