Pinned Tweet
Every company has to answer one question now: when the models get better, do you get better or do you get replaced?
We built Solace so every model release is a raise.
2
5
578
It's not well understood that every aspect of healthcare in the US is a deeply political decision.
People attempting to "fix" healthcare without taking this into consideration have no chance of success.
A liquid painted onto a toddler's cavity can stop the decay. No drill, no needle, no sedation.
In a randomized trial, 830 kids under 6 with severe tooth decay got silver diamine fluoride or blue-dyed water, brushed on twice, six months apart.
At 6 months, 54% of treated cavities had stopped progressing, vs 22.5% with placebo. The trial stopped early because it worked.
The catch: it stains the decayed spot black for good, and it's off-label for cavities in the US. For a 3-year-old who'd otherwise need sedation to get drilled, plenty of parents would take that trade.
(JAMA Pediatrics, 2026)
2
3
354
Plain salt water beat steroids for carpal tunnel, and the relief lasted.
In a randomized trial, doctors used ultrasound to guide saline around the squeezed nerve in the wrist, freeing it from the tissue pressing on it. About 15 minutes, no anesthesia.
At 6 months, 86% of saline-treated wrists still had meaningful symptom relief. In the steroid-shot group, none did. Adding steroid to the saline didn't help.
It's one center and 39 people (63 wrists). If it holds up in bigger trials, the cheapest liquid in the hospital could spare some people a surgery.
(Radiology, October 2026)
4
260
A cancer vaccine built from each patient's own tumor just cut the risk of melanoma recurrence or death by 42% in a phase 3 trial.
1,137 people with high-risk melanoma that had been surgically removed got immunotherapy, with or without an mRNA shot encoding up to 34 mutations unique to their cancer.
At 2 years, 82% of the vaccine group was recurrence-free vs 70% without it.
Survival data aren't mature and it isn't approved yet. Still: a drug made for exactly one person, improving on a strong standard of care in a 1,000+ person randomized trial.
(ESMO 2026, Oct 2026; accepted in NEJM)
7
406
Doctors have always eyeballed whether a patient looks frail. An AI just turned that hunch into a number, from one photo of a face.
Trained on 40 million face photos, it was tested on over 38,000 cancer patients.
Faces that looked 5+ years older than their real age: 21% higher risk of death. In the model's highest-risk group, 52.5% died within 90 days, vs 2.1% in the lowest.
It's retrospective, mostly radiotherapy patients, and weaker in Black and Hispanic patients. Not ready for clinic. Still: a selfie as a vital sign is a wild idea.
(The Lancet Digital Health, October 2026)
3
355
Optogenetics won the Nobel Prize this month. Now it's giving blind people some of their sight back.
Patients with advanced retinitis pigmentosa got a gene for a light-sensing algae protein injected into one eye, then wore camera goggles that pulse light into it.
7 of 10 became more sensitive to light, 6 by at least 4x. About half of those tested got better at finding objects.
It's blurry, narrow and colorless, and it's 10 people. Still: a protein borrowed from pond scum is doing the job of the eye's dead light sensors.
(NEJM, October 2026)
149
Medicare patients have lived in a financial system they can't understand for decades.
It's called an explanation of benefits.
"We are the cows of the future... Just as cows can't understand the financial system that controls their lives today, we will not be able to understand the AI financial system that will control our lives in 10 or 20 years."
Watch the whole conversation with @DanielStach for @hpcivilizace on @Vedact24 here: bit.ly/YNH-HPC
1
199
Can’t wait for AI to cure cancer so another AI can explain why the treatment's not medically necessary.
1
3
229
Most companies don’t have a hiring problem.
They have a firing problem.
Keeping one obviously mediocre person for a year does more cultural damage than making ten imperfect hires.
3
274
Most founders dramatically overvalue intelligence and undervalue intensity. After a certain IQ threshold, I'd bet on the person who cannot tolerate losing every time.
4
288
Every healthcare startup promising to cut spending should have a slide titled “Whose revenue we’re taking away.”
One of the strangest things about modern health culture is watching people obsess over seed oils, microplastics, supplements, and glucose spikes while drinking alcohol 3 nights a week.
The alcohol matters more than almost all of it.
1
3
250
The health industry massively overcomplicates being healthy.
If you sleep enough, lift weights, walk 10,000+ steps, eat enough protein, stay lean, and rarely drink alcohol, you’re probably doing more for your health than 99% of people taking 20 supplements a day.

