We have to make it so painful that generations from now CEOs will flinch at the idea of business corruptly teaming up with government.

reading the epstein emails shifted my politics. my prior assumption was that the ruling class was malevolent but also selfish, often working at cross-purposes. the epstein emails show a high degree of unity among the ruling class, vindicating the marxist view of class politics.

Surely megacorps use Claude for mass surveillance already under the fig leaf of ‘marketing’. Unlike corporations, government is at least ‘of’ & ‘by’ the people - in stated purpose, design & law! Is Dario a corporate exceptionalist or just ignorant to the risk of corporate rule?

On the cover of this month's issue of Nature Nanotechnology! Our nanosensor + AI-based liquid biopsy (blood test) for brain tumors that also enables the discovery of new biomarkers. nature.com/articles/s4156…

Automatic speech recognition struggles with children’s speech. That gap matters! The $120K Children’s Speech Recognition Challenge is driving progress toward models that truly understand kids. Join breakthrough! Submit your solution by April 6th. kidsasr.drivendata.org

1) Update the hero section with a simpler hook 2) Brought FDA trial prediction markets to the home page 3) Cleaned up the navbar and added a footer with some of the extra stuff. Might delete stuff there soon but better than being in navbar. Even more coming to

**Short summary** — FDA demands doctors drill real burr holes (real drill, real holes in the skull) in patients’ skulls just for placebo. Multiple holes. For fake treatment. What happens if you get approved for actual gene therapy later? Drill fresh new holes again. Brilliant.

Our latest Deep Dive covers work by hadar fisher Justin Baker Diego A. Pizzagalli, PhD Nigel Jaffe Christian Webb exploring whether smartphone passive sensors & LLMs can track daily activation in adolescents w/ anhedonia receiving behavioral activation therapy youtu.be/K1Lo50QBcQM?si…

Friendly reminder that while all this geopolitical madness is going on the planet merrily keeps heating up. Source: reddit.com/r/dataisbeauti…

>the matching makes it look like it works >a 100B USD drug if it works >those in charge of much of the FDA begin attacking the sponsor >cant articulate reason as to why the matching was bad >talks reveal drug pricing is now a major concern at FDA

Biotech is amazing!

🚨New model, new collaboration! Very excited to share this work with Emily Osterweil & team, gaining insight into FXS pathophysiology using human cortex. It takes an amazing and multidisciplinary team to make this possible, which we are super grateful for!🚨

RCT-Twin-GAN: a generative AI framework to build digital twins of clinical trials conditioned on real-world populations - now out in npj Digital Medicine Kudos to Phyllis Thangaraj and Sumukh Vasisht Shankar for leading this CarDS Lab 🗞️ nature.com/articles/s4174… 🧵

Our March issue is now live 🐞 In this issue we feature the complexity of autism. See more below: nature.com/natmentalhealt… Cover by Marina Corral Spence

Modern drug development is frustratingly inefficient with <10% of drug entering clinical development making it to the market.

Could be why pembro might work better than others in its class - this unintended off-target activity is against a legitimate fetal-origin target!

Claude was able to find Vinay Prasad's CV from 2023 (for tenure application?). onlinelibrary.wiley.com/pb-assets/CV%2… VP actually received 2 separate grants from the Arnold Foundation. One from 2017-20, the other 20-23. And probably is still supported by the foundation. Total from 2017-23

This paper is in preprint but appears to show the same attenuation of effect size with increasing methodological rigor in rTMS studies as has been shown in trials of Acceptance and Commitment Therapy (ACT). The longer and better you study things, the less they might actually work

Interesting take: “Every prior intelligence explosion—primate sociality, human language, writing, institutions—wasn't an upgrade to individual cognitive hardware. It was the emergence of a new socially aggregated unit of cognition. … The scaling frontier isn't just bigger

Practically everything he found was price-gouging from contractors or outsourcing. The mindbending part of DOGE was that it was going to increase public spending by cutting public staff. Building state capacity saves money.