My friend the brilliant @kevin_volpp dropped dead
He survived (read abt it: https://t.co/UdybYJbUBr)
But every year 300-400,000 in the US don't
This week: our @nature article (+A.Schubert J.Ross @m_sendhil M.Lingman) applies AI to this medical mystery
https://t.co/TEwYLpLtjD
I am reminded of cosmetic surgery , which is a pure out of pocket cost to the patient.
Adjusting for inflation, the price went ⬇️ over 20 years whilst conventional medical care went up 118% by 2012.
https://t.co/QkyI0KoaFc
7am flight out of SFO
- Leave house 1h before flight leaves
- 13 min Uber to airport, views of the Bay
- TSA agent smiles asks if you like the Grateful Dead
- Ads for AI agents that cure cancer
- Multiple food options spanning global cuisines
- 2 min walk to gate
- Get upgraded to business class
- Depart 5 min early
7am flight out of JFK
- Leave house 3h before flight leaves
- 73 minute Uber to airport, bumper to bumper traffic
- TSA agent hates you
- Ads for underwear
- No food options besides Jamba Juice and hardboiled eggs
- 17 min walk to gate
- Get downgraded to seat next to bathroom
- Depart 2 hours late
The most frightening thing is not that we don't understand China, but that we don't understand China while being absolutely convinced that we do.
The leader of a major country writes 700 pages laying out his logic of governance, historical judgments, institutional anxieties, and methods of reform-and yet almost no one in the West bothers to read it seriously, while everyone continues to debate what "China really wants."
That, in itself, is deeply ironic.
Some people should resist cognitive laziness.
Building this robot from scratch in the west may be a 5 year exercise at best. The supply chain engineering needed to make it at this price will take 5 decades, however.
have you ever wonder WTF is in the subway air? so have i.
nyc subway PM2.5 levels are 200 μg/m³. above ground in NYC is 12 μg/m³. singapore’s underground subway is 24.1 μg/m³. imo NYC subway air is slowly killing NYC.
DeSciNYC connected with Jack Klein from NewYorkLab who has been investigating what is actually happening in the air below the city.
come here more and RSVP in next tweet.
Let's get @NYCMayor@bryan_johnson there too.
Two of the most important biomedical breakthroughs came from science of the Gila lizard venom (GLP-1s) and yogurt (CRISPR genome editing).
"The system that turned that lizard into a medicine is now being dismantled."
"Less support for scientists means strange questions no one will get to chase."
gift link https://t.co/JKEzQCGm75
I suspect many in Silicon Valley would be surprised to learn that detecting a cancer earlier does not automatically translate into a better outcome.
In some cases, we can identify cancers years before they would have become clinically apparent, yet the ultimate outcome may be unchanged. We've seen this phenomenon with prostate and thyroid cancers, where increased screening and detection have often led to overdiagnosis and overtreatment without a corresponding reduction in mortality.
Finding more disease is not the same as helping more patients. The question isn't whether we can detect something earlier—it's whether doing so meaningfully improves outcomes.
"If curing cancer were the only result of building ever more powerful AI systems, I would cheer for their arrival.
But the problem is that their impacts are much broader, and we are moving too quickly to ensure that these impacts are positive."
my brilliant sister, @2plus2make5
Every woman on earth should be demanding solutions for menopause - your heart, your bones, your connective tissue, beauty, mental health, intellect - And no one tells u about this
Whoever invests in this now will be rich. Please God spare the millennials
Even PI is testing World Models for evals!
- Closed-loop Pearson r = 0.929, MMRV = 0.119, beating three strong baselines (Ctrl-World, IRASim, Cosmos-Predict 2.5)
- 3 core objectives applied to Cosmos 3:
1) Forward-inverse dynamics consistency: jointly trains the model to predict frames from actions and recover actions from frames, anchoring rollouts to a physically plausible action manifold and counteracting drift.
2) Cross-view consistency: trains the model to inpaint each camera view from the others, keeping multi-camera observations coherent over long rollouts.
3) Test-time consistency: reuses the inverse-dynamics mode at inference to terminate that drift away from the commanded actions