I thought a lot about a post @AJA_Cortes had about what would happen if you knew you would live to 100. It made me think more broadly in the context of longevity and where I think AI is going to make the most outsized impact in areas of health:
AI curing cancer is particularly relevant this week. Anthropic IPO’d on Thursday, and in the roadshow it sounded less like an enterprise software business (which is where most of its revenue comes from) and more like a pharma company. Dario Amodei said cancer, Alzheimer’s, heart disease and mental illness would be history in under a decade.
I don’t think a cure for cancer is happening in the next 5 years. What I do expect is enough progress to turn most, but not all, cancers into chronic diseases over the next decade, with a real possibility of personalized cancer vaccines after that.
Where I think the easier and more likely path to commercial success and revenue is through hedonism. Think about it- we’re hedonistic at our cores. Every consumer business orients to this eventually, and AI + pharma will be no different.
The building blocks are already here:
- GLP-1s for weight loss
- Myostatin inhibition for muscle. Regeneron’s trevogrumab cut the muscle loss on semaglutide roughly in half in Phase 2. Today that’s about keeping muscle, I’d bet it ends up being about adding it
- Dermatology: collagen stimulators like Sculptra are routine, and afamelanotide, FDA approved for a rare sun-sensitivity condition, is basically a tanning drug waiting for a cosmetic label. The unapproved version already sells online
- Massive advancements in hair loss, vision restoration (beyond just LASIK), reversing hearing loss
- Widespread adoption of peptide nootropics for memory, alertness, mood (more outgoing!)
- Sexual health: Addyi was approved for postmenopausal women last December, and tadalafil is a monthly subscription for a lot of men at this point
- STIs: doxy-PEP (an antibiotic taken after sex) cut syphilis and chlamydia about in half in San Francisco, and HIV prevention is now a shot twice a year. Gonorrhea is the holdout
- Contraception: a male birth control gel starts Phase 3 in 2027 and a non-hormonal pill for men is in early trials
Access and delivery for all of this will be everywhere, through consumer health models with pharma behind them. LillyDirect is the very initial stage. Lilly signing with Hims two weeks ago is the next one. I came across Layers, a business that is catering to this trend by @shivanijpatel that looks awesome. There will be many more.
Longevity gets reframed around hedonism too. The very concept of what defines a relationship could be rewritten if you can feel like you’re in your 30s and look like you’re in your 40s when you’re actually in your 70s
It’s a great heuristic. At Mirae, 62.5% of the team are former founders. YC backed companies too.
Was just putting together a note on @cognition, and about 167 of the employees, ~30%, are former founders themselves.
Wild. Very few leaders can attract that sort of talent. Truly the avengers - never seen anything like it.
1
108
Well it’s happening. I’m going to be in Vegas next week for the big HLTH conference. @dvasishtha said it best- unless you are raising or hiring, best the avoid.
On that note, hmu if you’re around.
2
2
129
There’s a famous analysis that shows a 3+ hr football game only contains about 2.5 minutes of actual gameplay. The rest is just time between snaps, TV timeouts and injury or called timeouts. Yet we turn in week after week.
Your doctor’s visit is kind of the same- the drive, parking and then the visit itself. 15+ minutes of forms to fill, waiting, getting vitals by the medical assistant, reconfirming those vitals and other information by the nurse practitioner or physician assistant and then the actual doctor visit, but wait- 10 more minutes of reconfirming details and administrative bullshit that you just shared. What the hell?
The role of clinical AI isn’t to replace the visit or interaction of the doctor, it’s to simplify all of the information gathering and interpretation for patient and doctor alike so he or she can give you 15 minutes of uninterrupted time. An EHR or app can’t do this. You need a neo-EHR, a third party that’s the conduit between patient, their data and the clinician and their EHR.
You need Mirae.
1
2
93
This problem is magnified in specialty care (eg., rheum, GI and derm) and the proliferation of new therapies coming online in the next 10 years will only further the need for true clinical AI to match therapy to patient at specific intervals of their care journey.
I feel confident in Mirae’s ability to impact all three projected opportunities to close the shortage gap through our specialty doctor “brain”
Covista analysis estimates AI could close 37% of the projected healthcare workforce gap by 2040 without replacing a single worker, largely through documentation, referrals, medication lists, billing, and other admin work.
62
Anuj Patel retweeted
Covista analysis estimates AI could close 37% of the projected healthcare workforce gap by 2040 without replacing a single worker, largely through documentation, referrals, medication lists, billing, and other admin work.
Our Mirae team is once again all together at our fffsite this week in <<REDACTED>> with our launch partner, one of the nation's largest AMCs. Some observations from talking with clinicians and operators across the system:
1. The access problem is real, and it's a serious financial focus for every provider system. Complex specialty care is seeing massive leakage, wait times are creeping up, and patients with complex autoimmune disease still can't get care inside of 2-3 months.
2. Technology is finding a very receptive audience in clinical care. No more lengthy implementations — your design partners are doctors, NPs, and clinical ops staff, and you're deploying in a matter of days. Clinicians-as-consumers is taking hold.
3. Matching patient cohorts to the right therapy can drive massive throughput and triage gains. This underpins much of Mirae's thesis, and our early assessment shows huge upside in capacity to treat autoimmune patients.
Looking forward to sharing more. Exciting week!
1
7
203
My wife just described 8Sleep as a Snoo for grownups and now I can’t get that image out of my head
1
83
A gym allegory:
There is a bar that wasn't unracked. But if you use that bar, it is now your responsibility to unrack it. There isn't some transitive property that allows you to forego equitique because you didn't receive the same treatment.
2
79
This is amazing, potentially foundational research. Unfortunately it, like many breakthroughs will struggle to make their way to bedside. It’s not feasible to biopsy patients at scale to determine the signal presented in this research. What you would need is a digital biomarker first, something that can be patient facing and free to deploy at scale with no marginal cost. Then, of those, a subset could get the type of analysis below.
We have built such a process at Mirae
Uncovered: early molecular warning signs of inflammatory bowel disease reveal the condition can begin before symptoms appear and continue silently even when patients appear well.
Published in @ScienceMagazine.
🔗wehi.edu.au/news/researchers…
1
4
287
Important, especially today, to remind everyone of the work we are doing at Mirae:
Precision medicine has traditionally been built from the molecule up: profile a small cohort deeply, identify a novel signature, many years later translate that to bedside. Mirae is building precision medicine in the opposite direction. Start from a continuous relationship with the patient, let AI define phenotypes from longitudinal data across a living population and deliver care against them today against the commercial availability of biologic therapeutics.
What is most important for Mirae at the end of the day is if we accomplish our ultimate goal of remission or a cure for a patient- the fastest and most effective way of letting patients resume their original, healthy day to day lives. We are making tremendous progress against this goal with multiple partnerships to announce over the coming weeks.
2
1
100























