🌶️

Q: Best A.I. use case?

A: The A.I. Primary Care Physician.

In my life I’ve been genuinely harmed by the bad advice of:

Unavailable PCPs
Overworked PCPs
Compromised PCPs
Dogmatic PCPs
Dumb PCPs
Lazy PCPs

I see zero reason that A.I. PCPs should not fix every one of these problems, for every person 🤖🩺

playing /bracket, assisting /bracky, talking /greatbooks

Brilliant. The “UX” of AI = “Personality”?

In a similar vein I recommended to an AI edu startup that a great way to actually get into classrooms would be to let teachers name and train the AI character versions of themself. So “Mrs. M” is actually helping each one of her students on the homework she assigns each night, in the way *she* approves of.

+1 to this

would love to also see them with some personality - for example imagine a "Dr House MD" (after the tv doctor). . . . medical app

(you can actually test this rn with any LLM by asking a "medical" question and requesting an answer in a personality or character type)

BDFL of Quilibrium, Eng @ Farcaster, ex-Coinbase, always opinionated, never hydrated

/quilibrium

That is another sticking point for sure

But I've personally faced a ton of problems from literally just being pointed in the absolutely wrong direction of what steps even need to be taken

Would be awesome if we can get diagnosis nailed, and even more awesome if we can also streamline and optimize the execution

Would be great, if the PCPs had the same authority as a regular physician to also prescribe, otherwise you immediately hit the same brick walls when trying to get issues addressed that PCPs tend to refuse to help on.

Same

There’s this normative force that pushes us to see doctors as both a) unimpeachable sources of authority and b) all equally correct

As obvious and common sensical as it is, internalizing that there is a spectrum of capability - and a bandwidth limit - to doctors is a large unlock

you probably don’t think about liquidity enough // it is what it is  // build value, not valuation

| 🎩  |  ↑  |  ⌐◨-◨

Should integrate to Apple Health and just fine tune on PubMed. 

Would be incredible

Good god that’s crazy man. Personal stories like this make me hope this comes to fruition at warp speed!

I have T1D and I can’t even *find* a GP in Canadas freaking capital city. I’ve been on a wait list for at ***LEAST*** 5 years. I had to explain to a Dr. at a clinic here what “insulin dependent” meant a few months ago! I wish I were kidding! An AI GP would be a massive improvement on what we have to begin with.

✵ Create ✵ Share ✵ Remix ✵

Experimental 2D/3D artist. Data science Pythonista. He/They. Nerdy++; (Nearly) all work CC-BY/CC0/FOSS

https://mxvoid.com

Agree that this is the case right now; 

Don't see why this would always need to be the case – if we develop and harness the tech optimally

Hmmm. I think an AI-assisted PCP or Nurse Practitioner would be better. I've succeeded in too many prompt injection attacks that I think it's important for a human to still be in that loop somewhere.

Healthcare investor by day, web3/AI tinkerer by night. 🟪 Host: /venturecapital /theitinerant

Good pt – that single fact may be driving a lot of (but not all, imo) the deficiencies I noted

I dig @balajis.eth re: the need to empower certain jurisdictions to be more "experimental" re bureaucratic heaviness in medicine, so we get real data on what the cost/benefit of our super risk-averse approach actually is

Yes there's a ways to go for perfection, but it's really threat of lawsuits, despite informed consent / waivers, that is holding back further adoption in the US.

Agree with you that there are many people in healthcare with great intentions.

Unfortunately, these alone are often not enough to overcome not only bad incentive structures (as you say) but also the physical constraints of time, energy, and processing ability.

Overall I revere medicine as magic and healthcare workers as saints

BUT insurance incentives can also ruin the best among them 

Doesn’t even have to be any issue with PCP for there to be room for improvement

off & onchain operations
<--->
Sharing ideas over at /workflow <--->
Problem solving is my jam😎
🎩

seemore.tv/free/fc/adam-

Agree w you on current state, but I am projecting what “should be” for AI

The upper bound of AI performance is orders of magnitude less constrained than the upper bound of human

Bc the right AI can provide the kind of advice only POTUS gets right now, at world scale

This is just physically impossible for humans

Compromised, dogmatic & dumb. A.I can be all these things as well. 
Not that A.I can't be useful here, it just can't be outright trusted for accuracy.

Staying up late at night wondering what the unrealized possibilities of Farcaster may be.

Totally — A.I. can be help good professionals be 10x better and help bad ones to be replaced.

This will enhance current PCP’s resources as well.

All I can say is PCP is a hell of a drug.

The pulse of Farcaster | Visit the Ponder V2 Frame 📋 | By @ba and @cojo.eth |

Literally dealing with this now - it’s such a cluster. Feels like providers care about 10% as much as they did when I was a kid.

@survey would you trust your healthcare to AI? And do you trust your current pcp?

Agreed. 

I went in for some routine tests a few months ago. An automated email sent me my results.

I put the results in ChatGPT and it explained everything and gave me concrete steps to address any results off the baseline.

My PCP never bothered to call to explain my results lol.

Most of my best friends from college are doctors. They would probably be much happier if they could get rid of the 80% cases that could be solved by “RICE” or are not serious cases.