🚀 With Meta's recent paper replacing tokenization in LLMs with patches 🩹, I figured that it's a great time to revisit how tokenization has evolved over the years using everyone's favourite medium - memes! Let's take a trip down memory lane! [1/N]

Brigitte

brigittetousi.hf.co

05 December 2024 09:09

It's day 3 of the @hf.co & AI World EU data drop! Up on the docket, too fast & too furious cumulative likes in relation to model releases, highlighting 2024's furious competition and AI's rapid progress. 🏎️🏎️🏎️💨💨💨

Alexander Doria

dorialexander.bsky.social

30 November 2024 01:41

This is almost entirely Bluesky fault at this point. You can't simultaneously promote an open social network ecosystem and make very strong claim of data protection. Was going to blow up at any time, and Daniel did them a favor by exposing the dissonance soon enough.

Jeremy Howard

howard.fm

29 November 2024 08:58

A librarian that previously worked at the British Library created a relatively small dataset of bsky posts, hundreds of times smaller than previous researchers, to help folks create toxicity filters and stuff. So people bullied him & posted death threats. He took it down. Nice one, folks.

Lucas Beyer (bl16)

giffmana.ai

28 November 2024 11:06

The point of my comment is to inform people that robots.txt is a gentleman’s agreement, not a legal document.