4/ And here's the part nobody prices in: when AI lets everyone produce the same thing, that thing loses value. Difference is what's valuable. Commoditized output deteriorates. The premium moves to what's scarce: judgment, taste, and being human.
1/ When intelligence becomes abundant, it stops being what differentiates people. The scarce inputs become everything else: drive, taste, trust, the willingness to act. AI doesn't flatten outcomes. It just moves the competition to a different field.
3/ Both the doom story and the utopia story assume humans are just cognition waiting to be amplified or replaced. We're not CPUs. Status, meaning, the enjoyment of the struggle, none of that is a compute problem. Humans stay humans.
@Anthropic You're 18. I don't blame the parents, it's the American dream to go to college. But for too many it's a delayed horror. That's exactly why I built the tool.
First time I used Claude this year, I built https://t.co/1YXayiLIus: a free tool to check if a college major is worth it. Now the government's cutting student loans for low-earning majors. Next: connecting students with real alumni for the story behind the numbers. @Anthropic
@Anthropic I graduated with $150k in student debt. Studied public policy at a top 10 university, and for me it was a fair investment. But I know countless people who went to private schools that added no value and were just ready to take their money.
Working in healthcare, it becomes evident AI is in an arms race with itself. Hospitals buy AI scribes to write cleaner notes. Payers buy AI to deny claims from those same notes. Same technology, opposite incentives. So who is AI actually working for here?
@Polymarket hospital CFOs treat pandemic prep like an insurance policy they hope to cancel. in healthcare it becomes evident: preparedness is a cost center until the beds run out. who prices the next one, and how?
@elerianm Hammack flagged the same thing: expectations are the transmission mechanism. in healthcare we see it when patients just accept the $75 copay and stop asking questions. what number is everyone quietly accepting now?
@chamath he's right that curing cancer doesn't need agentic AI. in clinics we see the opposite: the money flows to billing optimization, not the bedside. who do you think captures the value first?
@unusual_whales hospital CFOs are watching this too. half the expansion plans in healthcare just got shelved again while the cost of a hospital bed keeps climbing. how long before 7.5% is the number we all quietly accept?
@KobeissiLetter working in healthcare, it becomes evident fast: models optimize metrics nobody defined. AI drafts the note, nurse still has 30 patients. who exactly checks the study design?
Fed's Hammack just warned about the 'inflationary mindset' and she's right. Expectations are the real transmission mechanism. In healthcare we see this: once a patient accepts a $75 copay as normal, nobody renegotiates it. What number are we all quietly accepting now?
@paulg Healthcare is one of those 'extremes.' The demand is real, the training is brutal, and AI can't take your judgment call at 3am. But here's the catch: who's going to pay off the student debt for a decade of training?
@mcuban Working in healthcare, it becomes evident the contract is the product. Hospitals and carriers fight over network deals while the patient pays full price either way. How long before people realize they're the ones getting squeezed in the middle?
@alexandr_wang Working in healthcare, it becomes evident people want to be understood, not just answered. An AI that anticipates wanting is great until it's a patient who can't articulate symptoms. What happens when the wanting is invisible and the stakes are life or death?
@KobeissiLetter 23% of your money's value gone since 2020. Your paycheck buys 77 cents on the 2020 dollar and hospitals feel it too, supply costs exploded while reimbursements crawled. Who exactly got richer while your dollar melted?
@EricTopol@RBMD1982@nytopinion@AdamRodmanMD Working in healthcare, it becomes evident the AI is a great tutor but a lousy teacher. If trainees never build the pattern recognition themselves, what happens the day the chatbot is wrong or offline? Who carries the liability then?
That ER bill where the doctor fee is $200 and the facility fee is $2,000 isn't a billing error, it's the business model. The building bills like a hospital while you wait in a hallway bed. Who decided a hallway counts as a hospital stay?