Paul Kinlan

Homepage: https://paul.kinlan.me/

Paul Kinlan is a developer advocate at Google, where he focuses on the web platform and its future. He has been involved in various projects related to web development, mobile-first design, and progressive web apps. Paul is passionate about the web and its potential to connect people and ideas across the globe. He shares his thoughts on technology, development, and the future of the web through his blog and other platforms.

how might a modern lighthouse work with large language models?

please mind the model gap

I think i've got it: WebMCP is the new web intents

building a claw in the browser

shipping a prompt

how fast do ai models go stale?

are pwas cooked?

a business in a box

How might a browser be developed?

agent-do: my agent loop

webmcp is the new web intents ... maybe

damn claude, that's a lot of commits

the token salary

the llm whisperer

the prompt is the program

If NotebookLM was a web browser

headless cms was a stopgap

intercepting the web with an llm

calling apis from the browser shouldn't mean exposing keys

hypermedia was always the point

web components are packaging for ai-generated ui

what token slinging taught me about ai costs

on-device ai needs a trust model

embedding is the web's composability superpower

what falling ai model latency means for the web

A link is all you need

llms are becoming the new super-apps

this transition feels different