Twitter | Search | |
Nikhil Krishnan
Partnerships: Alum: Write Stuff: Social Experiment:
29,674
Tweets
1,479
Following
8,509
Followers
Tweets
Nikhil Krishnan 2h
If the patient moves closer to cash pay for those scenarios then doctors are more incentivized to actually help patients shop - a business would prob pop up that help DPCs contract/discover cheap commodity services
Reply Retweet Like
Nikhil Krishnan 3h
Replying to @alexadelman
I thought fax was short for “facts”
Reply Retweet Like
Nikhil Krishnan 3h
Replying to @singareddynm
question is whether people are going to abandon insurance and only use these clinics and go cash pay (otherwise you're effectively still paying those high premiums + deductibles) Some companies are going the DPC + Catastrophic insurance route, which is interesting
Reply Retweet Like
Nikhil Krishnan 4h
Replying to @chrissyfarr
Most underrated and underutilized dataset in healthcare is food consumption - one thing I like about the Apple Health app is you can actually integrate some of the food apps to see what you ate over time, but would love to integrate my grocery lists
Reply Retweet Like
Nikhil Krishnan 6h
Replying to @StuartBlitz
"I prefer the east coast because I like having seasons" is a widespread lie/delusion/stockholm syndrome
Reply Retweet Like
Nikhil Krishnan 6h
Replying to @mhdempsey
In theory a prescription dataset would be huge but it's also extreeeeemely valuable
Reply Retweet Like
Nikhil Krishnan 6h
Replying to @mhdempsey
definitely not the right person to ask here - actually one of the most difficult things isn't even individual datasets but actually linking datasets together knowing it's the same patient across them (imaging, prescriptions, etc.) but staying deidentified
Reply Retweet Like
Nikhil Krishnan 7h
Replying to @nikillinit
I'm a big fan of open-source data sets and think more companies should do it - competition should be on execution + deliver customer/patient satisfaction, not around hoarding data A company will eventually create a platform for that to happen, it's just a matter of time
Reply Retweet Like
Nikhil Krishnan 7h
Replying to @nikillinit
This is an area where multiple groups should actually come together and provide a more diverse/robust and real-time dataset, but many believe their data is proprietary and a moat hm, what if the government had bounties to open-source this kind of data 🤔
Reply Retweet Like
Nikhil Krishnan 7h
Replying to @nikillinit
If this becomes a standard training dataset, it's important to realize that we have to think about the biases of that reference set - especially in this case where all the data is coming from ONE hospital
Reply Retweet Like
Nikhil Krishnan 7h
This article about MIMIC, a de-identified medical record database that AI companies/research trains on, weaves together a few points I've been talking about recently (and a good read too)
Reply Retweet Like
Nikhil Krishnan 9h
Replying to @CortexIQ
I was messing around with IRL, but not quite the same. Would love what you're describing
Reply Retweet Like
Nikhil Krishnan 9h
Replying to @jamiehealth
maybe it seems less rude in the near future and more normal
Reply Retweet Like
Nikhil Krishnan 9h
Replying to @lifeofzing
no signaling is what I liked about snap! There were no external metrics I'm trynna get an invite into shoelace, someone hook me up
Reply Retweet Like
Nikhil Krishnan 11h
Replying to @The_OnlyHope
Thanks for being a part of it! I’m glad you’re enjoying it
Reply Retweet Like
Nikhil Krishnan retweeted
hope 20h
Subscribe and contribute to the decentralized Q&A platform that is nikhil's newsletter. I learn a ton from random people on the internet each week thanks to it.
Reply Retweet Like
Nikhil Krishnan 20h
Replying to @LukeThompson360
Maybe that feeling changes over time and it becomes more normal
Reply Retweet Like
Nikhil Krishnan 21h
Replying to @nikillinit
-another well done location based social network since there’s still a massive gap for that
Reply Retweet Like
Nikhil Krishnan 21h
Replying to @nikillinit
-calendar based social networks -no commenting social networks
Reply Retweet Like
Nikhil Krishnan 21h
Some ways I’d like to see social networks evolve going forward -time window constrained social networks like hq trivia -pay-to-use social networks -social networks that kick people off if they don’t contribute -pseudonymous social networks (with private 1-1 identity revealing)
Reply Retweet Like