We're Back! $1B GLP-1 Lessons. AI Funding Debates. China's 2030 Master Plan.
TL;DR
The “$1B one-person company” story got a lot messier once people looked past revenue headlines — TBPN argues Medv’s viral New York Times framing leaned on a projected $1.8B run rate, while key questions about valuation, margins, durability, FDA warnings, and lawsuits were missing.
Medv looks less like an AI miracle and more like an aggressive telehealth marketing machine — the hosts say the company outsourced core operations to Care Validate and OpenLoop Health, likely ran thin margins, and allegedly used 800 fake Facebook doctor accounts with names like “Dr. Tucker Carlzen MD.”
TBPN still thinks the one-person unicorn is possible — just probably not in telehealth arbitrage — they point instead to software and games like Balatro, built by solo developer LocalThunk, which sold 5M-7M copies at roughly $15-$20 and may have generated close to $100M in revenue.
AI’s current breakout pattern looks more like app proliferation than true solo breakout hits — the App Store saw an 85% jump in new apps this past quarter, but the hosts say they still haven’t seen the “Flappy Bird moment” or a vibe-coded product that actually reaches the home screen of mainstream users.
China’s industrial strategy is going full sci-fi, while the U.S. is learning from cheaper wartime hardware — the show contrasts China’s 2030 plan for flying taxis, humanoid robots, fusion, and brain-computer interfaces with the Pentagon’s low-cost Lucas drone, reverse-engineered from Iran’s Shahed and priced at $10,000-$55,000 versus $2M Tomahawks.
AI funding debates are getting sharper as model training costs explode — citing WSJ and The Information, TBPN highlights forecasts of OpenAI spending $121B on compute in 2028 and burning $85B that year, alongside internal tension over whether this is railroad-scale capex, software economics, or something entirely new.
The Breakdown
Back in Montreal, and immediately into moonshots and spiders
TBPN opens in full chaotic-house style: Easter greetings, a quick movie detour into Project Hail Mary, then a live check-in on NASA’s Artemis 2 stream. The crew jokes through a spider-on-the-mic incident while noting NASA had just posted that the astronauts were 250,000 miles from Earth — “the farthest any human has ever traveled.”
The $1B solo startup claim starts to unravel
The first big segment revisits the New York Times story on Medv, the GLP-1 telehealth company framed as the first “single-person billion-dollar company.” TBPN’s core pushback is simple: $1.8B in projected sales is not the same thing as a billion-dollar valuation, especially in a business where durability, margins, CAC, and legal exposure matter a lot.
Medv’s real business may have been ads, not medicine
The hosts walk through Medv’s outsourced stack — Care Validate and OpenLoop Health handling doctors, pharmacies, shipping, and compliance — and ask the uncomfortable question: what does Medv actually do besides marketing? They compare it to old nicotine and supplement playbooks, where companies could scale to huge revenue with sketchy tactics while still being worth almost nothing because nobody wanted the baggage.
Fake doctors, FDA letters, and the Puff Bar comparison
This is the spiciest stretch of the show. TBPN cites an FDA warning letter, a class-action anti-spam suit, and reporting that Medv used around 800 fake Facebook doctor accounts, including the unforgettable “Dr. Tucker Carlzen MD,” to sell compounded GLP-1s; the hosts say if you’re 70 on Facebook, that kind of ad could look very real. Their punchline is that this story seems far more about “pushing over-aggressive marketing tactics to the limit” than AI-enabled low-headcount scaling.
If the solo unicorn happens, it may look more like Balatro
After all that, they return to the original dream: maybe the one-person billion-dollar company still happens, just in software or games instead of regulatory gray zones. The example they love is Balatro, built by solo dev LocalThunk over about two and a half years, selling 5M-7M units at roughly $15-$20 each; they argue that if AI can compress dev and porting time, more solo breakout creators should appear.
More apps everywhere, but still no breakout vibe-coded hit
The App Store’s new app count jumped 85% this quarter, and the hosts say everybody they know now seems to have an app. But they still don’t see the AI-era equivalent of Flappy Bird — a solo-made tool or game that actually breaks into mainstream culture instead of just fattening the long tail.
China’s pig skyscrapers, master plan, and America’s reverse-engineered drone
In classic TBPN fashion, the macro section starts with China’s pork glut and “swine scrapers” — 26-floor pig towers built after African swine fever — before pivoting to defense and industrial strategy. They highlight the U.S. military’s Lucas drone, a low-cost attack system reverse-engineered from Iran’s Shahed and dubbed the “Toyota Corolla of drones,” then contrast that with China’s five-year plan, which reads like “Elon Musk’s fever dream”: flying taxis, 6G brain links, fusion, humanoid robots, and quantum everything.
AI money, labor, and the scale question nobody can answer yet
The final stretch stacks several funding and labor stories together: Anthropic’s $400M acquisition of Coefficient Bio, WSJ reporting on gigantic projected model-training spend, and The Information’s note on tension between Sam Altman and CFO Sarah Friar over IPO timing and burn. TBPN’s read is that nobody has consensus on how to model these companies — are they SaaS, railroads, utilities, or all three — while AI simultaneously creates jobs like “head of AI” even as automation fears keep hanging over the labor market.