Skills — The Merge Was Right, The Word Wasn't
The default is model-invoked. For most of what we build, it should not be.
Yuya's Blog
This is Yuya's blog RSS feed
Latest posts
AI's Missing UI
The smartest AI agents fail without good UI for reviewing their output. That's why everything becomes a chat app.
LLM Application Components and Their Cloud Analogs
A deep dive into how modern AI applications mirror cloud infrastructure, breaking down each component from LLMs as compute engines to memory systems as databases, and exploring why a Heroku-like platform for LLMs could be transformative.
AI memory as frontier
Models are good enough. The next unlock is memory: persistent, structured, and personalized rather than a bigger context window.
Getting past LLM "day one" problem with MIRIX
A look at MIRIX, a modular memory architecture that helps AI assistants get past the "day one" problem
The Memory AI Needs Depends on What You're Using It For
The memory AI needs depends on what you are using it for.
Linear's sync engine architecture
My notes on Tuomas Artman's talk about Linear's sync engine: the object graph, the transaction queue, and how the read and write paths stay in sync.
Why I'm excited about local-first software
Local-first software is a new approach to building applications that prioritize the user experience and data ownership. It's exciting because it simplifies the development process and makes applications more reliable.
My note from "CRDTs for Mortals" by James Long at dotJS 2019
My note from "CRDTs for Mortals" by James Long at dotJS 2019
What "optimism" does to "apollo-cache-inmemory"
How Apollo Client uses "optimism" to improve the performance of the cache