Introducing 🤗 Diarizers: a library for fine-tuning speaker diarization models 🗣️ Improve multilingual diarization performance by 30% with just 10 minutes of GPU compute time! ⚡️ The first release comes with training scripts, datasets and a Google Colab 🚀 Check it out! ⚒️

sota RAG in 2024

Delighted to share our perspective piece “Language is primarily a tool for communication rather than thought” now out in Nature: nature.com/articles/s4158… (with long-term collaborators (and friends) steven t. piantadosi and Ted Gibson, Language Lab MIT) 1/n

cognitive scientist: so the lesson of Clever Hans is we need.. engineer: more horses cognitive scientist: engineer: stacked horses. parallel horses. pooled horses. horse dropout. RL with horses in the loop. cognitive scientist: engineer: Hans is All You Need

LCD Soundsystem - All My Friends (Glastonbury 2024) #LCDSoundsystem

ardaudiothek.de/episode/zeitfr…

crazy

New Data Colada post. Harvard's Gino Report allows us to re-construct what Harvard says is one of Gino's "original" data sets. We can compare that to the posted data to see how the data were altered datacolada.org/118

arxiv.org/pdf/2303.13988

So great to welcome Aleph-Alpha in the OSS/open-weight AI community! Impressive open-knowledge approach

This tool is great for finding the bsky accounts of people you follow on X. It's noticable that almost no bands I follow on here have bsky accounts ... come on music people, do better! github.com/kawamataryo/sk…

One thing I dislike on this site is how I control less and less of what is shown to me in my feed: TikTok-like engagement bait pushed down everyone's throat. I loved Tech Twitter, and looking for ways to find this community again. On Bsky, there are themed lists like this:

it's so funny that anthropic is partnered with palantir haahahahah lets gooo e/acc "thanks for your hard work, chuds" - dario to all of the shocked do it for humanity AI researchers as he ships AI to assassinate people more accurately

Sci-Hub and Library Genesis has been defined in the literature as 'Black Open Access' model. In the last few months I've been working on a paper, where I explain that Sci-Hub is actually 'Red Open Access' while Library Genesis can be understood as a subtype of 'Green Open

Great (and scary) visualization of 2024 daily temperatures compared to prior years by the BBC today. Evocative of the iconic Joy Division album cover from 1979: bbc.com/news/articles/…

open source everything

Closed source work makes it impossible to trust these claims. As far as science is concerned, the attribution of techniques in R1 should be solely to deepseek authors unless OpenAI open up their code base with verifiable commit history.

I don’t think you get to claim credit for someone else’s paper by saying “if we had published we would have shown we did it first”

Finally took time to go over Dario's essay on DeepSeek and export control and to be honest it was quite painful to read. And I say this as a great admirer of Anthropic and big user of Claude* The first half of the essay reads like a lengthy attempt to justify that closed-source

FREE. DEEP. SEARCH. No bullet points, no lengthy report, no ads, no login, no nonsense design. You ask, we answer. Pure & Deep. search.jina.ai