You have probably been out of the game for a bit so easy to understand. Check out Serif Health that uses ML to parse TiC files into natural language readable info. Some of the biggest users are independent physician groups. Insurers as a group are the least interested. They know the prices.
@DutchRojas Great point! Check out Serif Health for great transparency data. The providers make it hard to find but Serif uses ML to make is easy to understand.
new ignition episode is live
how one reused password cost change healthcare $2.5 billion
part 1 of an 8-part healthcare security series
audio version on buzzsprout:
https://t.co/aUfyVeAcz3
change healthcare breach
sticker price: $872M
actual price: $2.457 billion
americans exposed: 192 million
all from one reused password
director's cut on substack with the stuff i had to cut for time:
https://t.co/DdudCikANq
one reused password
no MFA on a citrix portal
9 days of free roam in the network
$2.5 billion in damages
192 million americans exposed
if you're a doctor juggling 47 logins this one's for you
https://t.co/zfdzNI8dkc
@mark_k This is just a great question and something we talked about this week on a podcast: what will be the physician liability if THEY DON'T utilize AI tools in the future?
Brand meds sell at list price, then the supply chain adds markups. Same pattern with hospitals - they set prices, insurers pass them through. Both are cost problems at the source. Both have middle-chain markups.
The difference: you're disrupting pharma distribution with Cost Plus. Who's doing that for hospital prices? Price transparency data shows 300-500% variation for the same procedure in the same city. That's the bigger dollar problem for employers. I'd love it if you would expand CostPlus to hospitals!
new ignition episode
target's pregnancy algorithm was a 12-year preview of how healthcare AI is going to work for the rest of our lives
and almost nobody is framing it correctly
on the podcast:
https://t.co/38ksYjDu4D
genuinely cannot stop thinking about this one
your smartwatch can detect cognitive decline 6-12 months before you can. atrial fibrillation days early. depression weeks ahead of clinical scoring.
the science is settled
the actually hard question is who gets to see the model output
wrote it up:
https://t.co/oe6QRd6tpO
@TheAhmadOsman It all demands on your depreciation schedule whether you “make money” or not. And, what are you using the inference for. So much gripe about Anthropic pricing on here, but the delta between client pricing and local LM cost is huge. That is why we do it. Better margins.
target's pregnancy prediction algorithm from 2012 needed 25 features to know a teenager was pregnant before her dad did
your phone broadcasts 200-400 health-adjacent signals an hour now
a modern version needs maybe 6
this is the healthcare AI story nobody's telling correctly
https://t.co/fRGQaBRWoW