⚡️Introducing WebGPT⚡️
Just this month, Chrome announced WebGPU's release. What does this mean? Near-native GPU speeds, from the web!
I took the opportunity to build WebGPT: a package to run GPT models entirely on the browser.
Here's why this is a big deal: twitter.com/i/web/status/1…
WebGPT is just four Javascript files. No installations, no libraries, just ~1000 lines of vanilla JS!
Soon, every device in the world can run models... just by clicking a link. No friction, amazing interoperability.
But why does this matter?
👉 Preserve privacy.
👉 Instant latency.
👉 Costs less (it's free).
💡Add next-level autocomplete into your website.
💡Easily fine-tune GPT on your private data.
💡Crowd-sourced distributed model training across thousands of devices.
Interested? I've open-sourced the code on Github:
github.com/0hq/WebGPT
If you have Chrome Canary (or v113), try it here:
kmeans.org
Keep in mind: I started this knowing nothing about transformers or attention, GPUs or matmuls, so the code is still rough/unoptimized. Contributions welcome!
Shoutout to @asciidiego for pitching me this project at a party two weeks ago (+ @karpathy for the lectures, as always).
Share this Scrolly Tale with your friends.
A Scrolly Tale is a new way to read Twitter threads with a more visually immersive experience.
Discover more beautiful Scrolly Tales like this.