Fun wkend read by Alexia Jolicoeur-Martineau arxiv.org/abs/2510.04871… I was surprised to see "2" as the optimal # of latent features. The paper argues there is no reason to split z, maybe we can split it into memory and reasoning. Basically, build up some rules/facts in memory as you recurse

Narendra Modi indiatoday.in/amp/india/stor…

Zoom achieved a new state-of-the-art (SOTA) result on Humanity’s Last Exam (HLE): 48.1% — outperforming other AI models with a 2.3% jump over the previous SOTA. ✨ HLE is one of the most rigorous tests in AI, built to measure real expert-level knowledge and deep reasoning across

Zoom coming for Gemini was not on my 2025 list! Congrats to the team!

#TheGreatFlood is a great visualization of reinforcement learning at scale for human level intelligence. The mother functions as the primary agent, repeatedly iterating through a simulated environment with the objective of saving her son.The leak in memory is similar to the

A single chart to kill “college is dead” narrative.

I would add that there are ways to maximize in FAANG like positions as well. The trick is to align with passionate managers/leaders (often new hires). People who “stay put” often get promoted to middle management, stop up-skilling themselves, avoid confrontations and stop

The sales collapsed with everyone trying to cut MSG out of their food. The company Ajinomoto unfortunately had to pivot to something else foundational.

Voting for the next mayor who raises NYC's citibike speed back to 18 mph. Currently actively harming NYC GDP

Nadal is in new York 😱

NYC is becoming a hotbed for frontier tech. Hear from a few of the names leading the way including Joseph Krause (Radical AI) and Rob Cochran (Fauna Robotics) on March 31st @ SPC NYC moderated by our friend Eric Newcomer.

Protecting interest of one group doesn’t automatically mean harming other. He also didn’t say I will not let robotaxis operate period. Although, it will be a terrible decision. The question is asked at 15:00 x.com/nycmayor/statu…

Today, we released Lyra 2.0, a framework for generating persistent, explorable 3D worlds at scale, from NVIDIA Research. Generating large-scale, complex environments is difficult for AI models. Current models often “forget” what spaces look like and lose track of movement over

Probably the lowest piece of argument clinic in this week of The Economist. The piece poorly tries to wash away the history of Mughal rule over India in guard of giving us paneer and biryani? What? 🤷