Troy Astorino (@troyastorino) 's Twitter Profile
Troy Astorino

@troyastorino

CoFounder and CTO @PicnicHealth. Improving health with better medical records. @MIT and @ycombinator alum. Loves code, data, econ, physics, and crossfit.

ID: 379232769

linkhttp://www.troyastorino.com calendar_today24-09-2011 16:06:34

240 Tweet

305 Followers

111 Following

Nicholas Fabiano, MD (@ntfabiano) 's Twitter Profile Photo

Our study found that an emergency department visit for hallucinogen use was associated with a 21x risk of developing schizophrenia. đź§µ1/11

Our study found that an emergency department visit for hallucinogen use was associated with a 21x risk of developing schizophrenia.

đź§µ1/11
Troy Astorino (@troyastorino) 's Twitter Profile Photo

Fascinating. This is a story of rapid, dynamic evolution, cultural and genetic, through brutal and violent conquest. It’s a much shorter and more recent timescale than I would have considered relevant for evolutionary development, but it’s plausible that a selection mechanism

Jared Friedman (@snowmaker) 's Twitter Profile Photo

Lots of hot takes on whether it's possible that DeepSeek made training 45x more efficient, but Jeffrey Emanuel wrote a very clear explanation of how they did it. Once someone breaks it down, it's not hard to understand. Rough summary: * Use 8 bit instead of 32 bit floating point

Arc Institute (@arcinstitute) 's Twitter Profile Photo

Today, we're launching the Arc Virtual Cell Atlas, a growing resource for computation-ready single-cell measurements. As the initial contributions, Vevo Therapeutics has open sourced Tahoe-100M, the world's largest single-cell dataset, mapping 60,000 drug-cell interactions, and we’re

Today, we're launching the Arc Virtual Cell Atlas, a growing resource for computation-ready single-cell measurements. As the initial contributions, <a href="/vevo_ai/">Vevo Therapeutics</a> has open sourced Tahoe-100M, the world's largest single-cell dataset, mapping 60,000 drug-cell interactions, and we’re
Mike Knoop (@mikeknoop) 's Twitter Profile Photo

ARC-AGI-1 was designed to challenge deep learning. ARC-AGI-2 is designed to challenge AI reasoning. ARC-AGI-3 will challenge AGI we have not built yet. Each is designed to be easy, fun, and inspiring for humans. Today we're previewing v2 and opening donations to fund v3!

Troy Astorino (@troyastorino) 's Twitter Profile Photo

Wow! This is unexpected. Excited to see who they bring in The delta in viewpoint between the newsroom and the editorial team could be…a little extreme. The WSJ will feel like a singular voice in comparison

Elliot Hershberg (@elliothershberg) 's Twitter Profile Photo

Extremely clever new NGS tech from Roche 🧬 If it's hard to discriminate between nucleic acids accurately with a nanopore, why not synthesize a new polymer off a DNA template that is easier to sequence? It's an intuitively simple idea, but took *a ton* of creative nucleic acid

Troy Astorino (@troyastorino) 's Twitter Profile Photo

Congratulations to Lukas Biewald and the W&B team!!! 🎉 What an incredible 🚀 and outcome. Weights & Biases is already core infrastructure for the world’s AI transformation — I’m excited to see what you build with CoreWeave

roon (@tszzl) 's Twitter Profile Photo

"I would give the greatest sunset in the world for one sight of New York's skyline. Particularly when one can't see the details. Just the shapes. The shapes and the thought that made them. The sky over New York and the will of man made visible. What other religion do we need? And

François Chollet (@fchollet) 's Twitter Profile Photo

Today, we're releasing ARC-AGI-2. It's an AI benchmark designed to measure general fluid intelligence, not memorized skills – a set of never-seen-before tasks that humans find easy, but current AI struggles with. It keeps the same format as ARC-AGI-1, while significantly

Today, we're releasing ARC-AGI-2. It's an AI benchmark designed to measure general fluid intelligence, not memorized skills – a set of never-seen-before tasks that humans find easy, but current AI struggles with.

It keeps the same format as ARC-AGI-1, while significantly
Alex Tabarrok (@atabarrok) 's Twitter Profile Photo

Peter Marks Forced Out at FDA - Marks pushed the FDA to move faster—against every cultural and institutional incentive to go slow. His ouster is bad news. marginalrevolution.com/marginalrevolu…

Ryan Petersen (@typesfast) 's Twitter Profile Photo

Thousands, and then millions, of American small businesses, including many iconic brands, will go bankrupt this year if the tariff policies on China don’t change. 🧵

Troy Astorino (@troyastorino) 's Twitter Profile Photo

It sucks that people will now associate better typography with a lack of authenticity — I delved into em dash usage way before AI thought it was cool! option-shift-hyphen ftw

tobi lutke (@tobi) 's Twitter Profile Photo

I really like the term “context engineering” over prompt engineering. It describes the core skill better: the art of providing all the context for the task to be plausibly solvable by the LLM.