I had a blood clot. An AI diagnosis may have saved my life | Gleb Tsipursky
TL;DR
Gleb Tsipursky spent five days treating pain and swelling in his left calf as a muscle issue. A chiropractor handled it the same way, but the symptoms kept getting worse. His own AI health tool, loaded with medical records, medications, lab results and visit notes, raised deep vein thrombosis and pointed to the key diagnostic step: an ultrasound. Because his primary care office and urgent care could not provide the scan directly, Tsipursky went to the ER. Doctors found four blood clots in his left leg.
Nauti's Take
Medical AI does not win by cosplay as a doctor; it wins with context: records, meds, labs, progression. Builders should spend less energy polishing bedside-manner chat and more on clean data grounding, escalation logic, and hard boundaries with clinical care.
Briefingshow
The case shows a practical role for medical AI: not as an automatic doctor, but as a navigation layer in a fragmented health system. The decisive point was not treatment itself, but identifying which test was urgent. That is where well-tested AI can buy time when patients misread symptoms or get routed into slow care paths.
Sources