Need a better habit than doom-scrolling? Brain Gamer, the app Dr. Kumar designed, gives you nine quick games that work different parts of your thinking, knowledge, logic, and planning, in about the time it takes to wait for coffee. Play three free games a day, or go unlimited for a small monthly fee. A surgical robot trained on 100,000 cases can be smart on day one, maybe smarter than a surgeon with decades of grey hair to show for it. That should feel backwards. Every surgeon learns that older surgeons are safer than younger ones, because judgment and a feel for anatomy only come from years of cases and complications. So what happens when a machine skips that apprenticeship entirely and shows up already trained on more cases than one surgeon could see in a lifetime? Ravi's guest is Ryan Estes, a serial tech entrepreneur who runs a holding company in Denver with a portfolio of service agencies, software, and media brands. One of those brands is AI for Founders, where Ryan interviews startup founders building in AI, technology, and robotics, with AI in medicine as a recurring theme. He livestreamed years ago on Justin TV, before it became Twitch, and he says he's not a doctor and doesn't play one online, though he listens to Ravi's episodes at least twice to keep up. Most of what's broken in medicine isn't the medicine, it's the paperwork around it. Ryan points out that of your appointment time, you might only get about thirteen minutes with your actual physician, the rest eaten by clipboards, iPads, and insurers. Ravi has lived the fix on one side: AI scribes that used to take him as long to chart as the visit itself now produce a note in about a minute, so a slot that used to buy five minutes of face time now buys thirteen or fourteen. On the other side, he describes surgery requests denied in a split second, even for patients who can't walk, and his read is that insurers are counting on busy doctors not appealing. They also walk through where automation makes sense first. Acne, strep throat, and hypertension are treated with algorithms that are already written down, which is why Ryan uses an app called Nala Health, $4.99 a month, to get his daughter a prescription without needing a doctor on the other end of a text. Diagnostics are moving just as fast but in stranger directions: fifteen trained beagles sniffing a breath mask can pick up early stage cancers that billions of dollars in liquid biopsy research still haven't matched, and a wearable called Canary listens to someone's voice for signs of anxiety, depression, or the kind of subtle decline a caregiver might miss day to day. Robots are already in neurosurgery, urology, and orthopedics, placing screws through a two-millimeter corridor with sub-millimeter accuracy, though Ravi is careful to note they're human-controlled, the surgeon still lines up the shot and drives the screw. He admits he'd rather trust an algorithm that knows how to do a task as safely as possible than a human who makes mistakes, joking that his surgical colleagues are probably planning how they're going to lynch him for saying so. His actual position is narrower than that joke suggests: combine machine accuracy with the empathy, mission, and curiosity only humans bring, and medicine gets better, not replaced. Episode Resources Dr. Ravi Kumar's WebsiteThe Dr. Kumar Discovery on Apple PodcastsThe Dr. Kumar Discovery on Spotify In this episode, you will discover AI scribes now produce a note in about a minute, giving patients thirteen to fourteen minutes of real face time instead of five.An NHS case where an AI scribe dropped the word null from a report, turning a normal scan into what looked like multiple sclerosis.Why 93 percent accuracy, which sounds great on a test, is a failing grade when the other 7 percent includes a missed penicillin allergy.How requests for spinal cord decompression surgery can come back denied in a split second, even when the patient can't walk.Ravi's theory that insurers deny claims knowing a predictable share of doctors never bother to appeal.Why acne, strep throat, and hypertension, with their already-written treatment algorithms, are the first real foothold for AI in prescribing.How a $4.99-a-month app called Nala Health scans a face for acne and writes a prescription, with only the first hundred reviewed by a doctor.Why untreated strep throat, which can lead to rheumatic fever and heart valve disease, gets treated on suspicion alone.Fifteen trained beagles sniffing a breath mask that can detect roughly four stage one cancers, something billions in liquid biopsy research still hasn't matched.A wearable called Canary that listens to someone's voice to catch anxiety, depression, or a subtle decline in Parkinson's and dementia patients.Why human-controlled robots are already placing spinal screws through a two-millimeter corridor with sub-millimeter accuracy in neurosurgery, urology, and orthopedics.Why a robot trained on 100,000 surgical cases can be smart on day one, even though older surgeons are safer than younger ones.Connecticut's new ban on AI claim denials, which only protects people on the state's own health plan.Why Ravi isn't worried about ChatGPT inside Epic, given HIPAA protections and an expected opt-out from AI training.Episode Highlights [00:00:00] Welcome, disclaimer and Brain Gamer[00:02:22] Where this conversation will live[00:05:06] Utah lets AI write prescriptions[00:06:58] Asking a doctor friend for antibiotics[00:10:04] Untreated strep and its real consequences[00:12:45] Dogs that smell cancer[00:15:05] Wearables and richer diagnostic signals[00:18:35] Connecticut's narrow ban[00:20:25] Doctor, patient and the intermediary[00:23:11] ChatGPT, Epic and your medical record[00:24:40] Open Evidence and 93% accuracy[00:27:57] Trust, ads and industry credibility[00:30:55] An AI scribe omits one word[00:32:15] How scribes rewrote the office visit[00:35:15] Life before AI transcription[00:37:34] The scribe understands every accent[00:40:01] Robots already in the operating room[00:43:22] Older surgeons versus day-one robots[00:45:25] What robots can never have[00:47:52] Nine games for nine thinking circuitsKey Takeaways The biggest fixable problem in medicine right now is administrative, not clinical, since Ryan points out that of a typical visit, patients might only get thirteen minutes with their actual physician while the rest disappears into clipboards, iPads, and insurer paperwork. If you're denied a medically necessary procedure, don't assume the denial is final, Ravi says the peer-to-peer appeal is what actually gets most of his rejected cases approved. Treat AI medical answers as drafts rather than verdicts, since even a 93 percent accuracy rate leaves real gaps, like a missed penicillin allergy, and Ravi prefers tools like Open Evidence that refuse to answer rather than guess. Automated prescribing makes sense where the algorithm is fixed and the downside is small, but drugs like Accutane or tetracycline carry enough risk that a doctor belongs in the loop. The goal both men land on isn't machines replacing doctors, it's dividing the labor, machine...