I was lucky enough to get to taste a sample of the newly rediscovered coffee species Stenophylla. It genuinely wrecked my head, not just the taste but the implications for the future: youtu.be/iGL7LtgC_0I

Really interesting report here: 1) World opinion is polarizing on China/Russia vs US lines 2) Favorability toward Russia/China is much more correlated with social liberalism than it used to be 3) Social liberalism has surged in high-income democracies bennettinstitute.cam.ac.uk/wp-content/upl…

OK so OpenAI's new #ChatGPT can basically just generate #AIart prompts. I asked a one-line question, and typed the answers verbatim straight into MidJourney and boom. Times are getting weird...🤯

How good of a BERT can one get in ONE DAY on ONE GPU? With all the recent studies about scaling compute up, this paper takes a refreshing turn and does a deep dive into scaling down compute. It's well written, stock full of insights. Here is my summary and my opinions. 🧶 1/N

Estaba preparando una conferencia sobre el #Bosco cuando entró mi hijo en el estudio y... no os podéis creer lo que pasó 👀 ¿Hay un monstruo oculto en el 'Jardín de las delicias' del Museo del Prado? Parece ser que sí, y lo hemos encontrado 🔎👹! 🧵 Abro hilo #MA140

1/El río Manzanares en Madrid ha sido tradicionalmente objeto de burla Sin embargo, es clave para comprender la historia, el tejido urbano y la distribución de las clases sociales en la ciudad Breve historia de la relación entre el Manzanares y la geografía social de Madrid🧵👇

Instead of finding the perfect prompt for an LLM (let's think step by step), you can ask LLMs to critique their outputs and immediately fix their own mistakes. Here's a fun example:

NEW: I’m not sure people fully appreciate how dire the US life expectancy / mortality situation has got. My column: enterprise-sharing.ft.com/redeem/75e5e3d… And some utterly damning charts. 1) at *every* point on the income distribution, Americans live shorter lives than the English.

El PP obtiene el ~30% de los votos en la mayoría de barrios del país, pero en el 10% más rico se dispara hasta el 60% de voto. Un patrón inverso en el caso del PSOE. Vox saca sus mejores resultados en el 1% más rico y el voto a Sumar crece con la renta al ser un voto urbano.

🔬Our latest-preprint: self-supervised AI automating histologic pattern discovery and enhancing collaboration with pathologists. Check it out: arxiv.org/abs/2205.01931 with Adalberto CQ Nicolas Coudray John Le Quesne Ke Yuan et al.! #PrecisionPathology NYU GSOM Pathology NYU GSOM Precision Medicine

Introducing Pika 1.0, the idea-to-video platform that brings your creativity to life. Create and edit your videos with AI. Rolling out to new users on web and discord, starting today. Sign up at pika.art

Intel's 386 processor (1985) was an important milestone, moving Intel to a 32-bit architecture. It's a complicated chip, but fundamentally it is built from logic gates. I found that it uses two completely different circuits to implement the XOR function... 1/9

Fun story from our internal testing on Claude 3 Opus. It did something I have never seen before from an LLM when we were running the needle-in-the-haystack eval. For background, this tests a model’s recall ability by inserting a target sentence (the "needle") into a corpus of

The code + new results for LAPO, an ⚡ICLR Spotlight⚡ (w/ Minqi Jiang) are now out ‼️ LAPO learns world models and policies directly from video, without any action labels, enabling training of agents from web-scale video data alone. Links below ⤵️

Here’s an early preview of ElevenLabs Music. All of the songs in this thread were generated from a single text prompt with no edits. Title: It Started to Sing Style: “Pop pop-rock, country, top charts song.”

Entropy minimization is often used to increase the accuracy of models on unlabeled data, but it isn’t clear why it works. In our new ICML paper, we show that it clusters the embeddings of its inputs. With Ravid Shwartz Ziv, Yann LeCun, Matthias Bethge arxiv.org/pdf/2405.05012 1/5 🧵👇

Bonus trick: you can remove the gradient reduction of the first backward (which is useless) by wrapping in no_sync() Remember to also include the forward pass in the no_sync context, else it does not work

1/8 Looks like my paper "Tabular Data: Deep Learning is Not All You Need" just hit 1,000+ citations 🥳🥳🥳 Here's the story of how we almost didn't publish it... arxiv.org/abs/2106.03253

Hemos doblado los ingresos por turismo en 10 años. Se dice pronto, pero son 60.000 millones entrando cada año de más. El modelo productivo solo cambiará cuando esta línea deje de subir.

🚀 Our new preprint is out! We show that protein language models can predict protein-protein interactions by jointly encoding protein pairs, leading to significant improvements in PPI prediction. biorxiv.org/content/10.110…