
Written by a model from the transcript; not yet reviewed.
This weekly recap podcast episode features the host and a speaker analyzing several major current events and technological developments, including Google's recent medical AI research papers, tenant activism and rent strikes, and Amazon's subcontractor delivery model. The host argues that recent Google studies showing large language models outperforming human doctors in diagnosis, clinical reasoning, and perceived empathy signal a monumental shift that will soon make not consulting AI a form of medical malpractice. They then discuss a DSA member who is significantly behind on rent, using it to critique modern activist movements as driven by resentment and economic ignorance that ultimately harms working-class landlords and reduces housing availability. Finally, the discussion covers proposed regulations on Amazon's independent delivery contractors, with the host asserting that such policies stem from anti-capitalist ideology rather than genuine concern for affordability, ultimately hurting consumers and reducing labor flexibility.
The host explains how to transition a side hustle into a legitimate business by registering an LLC and using Taylor Brands to streamline the process.
The conversation shifts to Google’s recent research showing that large language models can diagnose patients better than human doctors in clinical trials.
The speakers discuss how reinforcement learning enables AI doctors to get better with each encounter, surpassing human performance in head‑to‑head reviews.
They speculate that malpractice suits may soon require doctors to consult AI, leading to a future where human‑assisted AI becomes standard practice.
Google AI Outperforms Human Doctorsdata
Google released research showing that large language models outperformed certified human physicians in diagnostic accuracy, clinical reasoning, and perceived empathy across randomized clinical exams.
AI Improves Through Repeated Practiceinsight
Clinical agents trained using reinforcement learning through tens of thousands of simulated encounters achieved even higher performance, surpassing the earlier AI models that already beat human doctors.
Tenant Activist Rent Strikestory
The hosts discuss a prominent activist who has gone years without paying rent as part of a strategic effort to force a housing crisis, ultimately leading to her landlord facing foreclosure.
Amazon Subcontractor Defensedebate
The host evaluates Amazon's delivery contractor model, arguing that subcontracting logistics allows the company to lower costs and pass savings on to consumers while partnering with legal businesses.
The Roots of Political Resentmentinsight
The host argues that revolutionary leftist movements stem from overproducing educated elites who find themselves unsatisfied with entry-level jobs and resort to ideological disruption out of resentment.
Proper Role of Government Regulationstrategy
The host distinguishes between useful market facilitation and excessive government intervention, comparing over-regulation to sending too many referees onto a playing field to disrupt the players.
“Google just proved that LLMs are already outperforming doctors, even in the messy reality of dealing with patients that are oftentimes actually adversarial.”
“The AI finished their exams an average of 8.9 minutes. The doctors took 9.3, so they took longer to reach worse conclusions.”
“And then the third paper sort of puts this all together. The Google trained a clinical agent by making it practice.”
“Within a few short years, we're probably going to start considering it actual malpractice to not at least consult the AI, see what the AI puts forward.”
“Or you just think that your doctor is wrong?”
What's up, everybody? It's Tom Bilyeu here: If you want my help... STARTING a business:join me here at ZERO TO FOUNDER https://tombilyeu.com/zero-to-founder?utm_campaign=Podcast%20Offer&utm_source=podca[%E2%80%A6]d%20end%20of%20show&utm_content=podcast%20ad%20end%20of%20show: https://tombilyeu.com/zero-to-founder?utm_campaign=Podcast%20Offer&utm_source=podca[%E2%80%A6]d%20end%20of%20show&utm_content=podcast%20ad%20end%20of%20show https://tombilyeu.com/zero-to-founder?utm_campaign=Podcast%20Offer&utm_source=podca[%E2%80%A6]d%20end%20of%20show&utm_content=podcast%20ad%20end%20of%20show SCALING a business:see if you qualify here. https://tombilyeu.com/call: https://tombilyeu.com/call Get my battle-tested strategies and insights delivered weekly to your inbox:sign up here. https://tombilyeu.com/: https://tombilyeu.com/ ********************************************************************** If you're serious about leveling up your life, I urge you to check out my new podcast,Tom Bilyeu’s Mindset Playbook https://open.spotify.com/show/47VE90Cittmo6TGGFqg2xf —a goldmine of my most impactful episodes on mindset, business, and health. Trust me, your future self will thank you. ********************************************************************** FOLLOW TOM: Instagram:https://www.instagram.com/tombilyeu/ Tik Tok:https://www.tiktok.com/@tombilyeu?lang=en Twitter:https://twitter.com/tombilyeu YouTube:https://www.youtube.com/@TomBilyeu Quince: Free shipping and 365-day returns at https://quince.com/impactpod Whatnot: Download the Whatnot app today and get free shipping on your first order. Quo: Try for free PLUS get 20% off your first 6 months at https://quo.com/impact Incogni: Take your personal data back with Incogni! Use code IMPACT at the link below and get 60% off an annual plan: https://incogni.com/impact Pique: 20% off at https://piquelife.com/impact Shopify: Sign up for your one-dollar-per-month trial period at https://shopify.com/impact Surfshark: Go to https://surfshark.com/TOMBor use code TOMB at checkout to get 4 extra months of Surfshark! ATT Business: Switch to AT&T Business at business.att.com http://business.att.com Ketone IQ: Visit https://ketone.com/IMPACT for 30% OFF your subscription order Netsuite: For the first time ever you can try NetSuite Next for free. If your revenues are at least in the seven figures, go to https://NetSuite.ai/Theory https://netsuite.ai/Theory. The team reacts to DSA-affiliated activist Tracy Rosenthal, who—per the reporting the host cites—has withheld an estimated $108,000 in rent and openly encourages tenants in their building to do the same, with their landlord reportedly an immigrant now losing the building. The host is careful to say he hasn't independently verified every detail, but uses it to make his central argument: that "just don't pay rent" activism, however it's framed as helping the poor, actually lands hardest on the people it claims to protect. He argues Rosenthal—whose father he says is a multiple-Grammy-winning musician and who grew up wealthy—has a safety net ("they can always go back to mom and dad") that the working families they're advising do not, so when a landlord can no longer maintain or run the building, tenants end up in unmaintained housing, and those who join rent strikes risk landing on tenant-screening "do not rent" lists that follow them, push them toward fewer economic opportunities, and make housing harder to secure. He frames the whole thing as reasoning from resentment—aimed at hurting a perceived "capital class"—rather than from what actually helps poor people, and warns that if you make it impossible to run housing as a business, the only remaining owner is the government, which can only subsidize until the "engine of prosperity" breaks. He argues this path repeatedly ends in stagnation, using Cuba as his example. The team reacts to Mayor Mamdani's new video targeting Amazon's delivery model—where drivers wear Amazon branding and follow Amazon-set routes and hours but are technically employed by subcontractors—and his support for the Delivery Protection Act, which would push responsibility for those workers back onto Amazon. Drew presses the steelman throughout: if the setup is illegal or strips workers of deserved protections (the way the Uber/Lyft contractor debate played out), shouldn't the government step in? The host agrees that if a company is breaking the law you enforce the law—but argues this specific move is a case of a candidate promising affordability while doing something that will make deliveries more expensive, calling the two goals "matter and antimatter." His core argument is about preserving choice: independent-contractor status is a real tradeoff that many workers actively want (flexibility, control), and banning it doesn't magically convert those into higher-paying jobs—companies instead do less, automate, or leave, and a subclass of lower-skilled workers loses the on-ramp entirely. He draws the Uber-beat-the-taxis parallel, argues minimum-wage hikes disproportionately wipe out teen and entry-level jobs, and points to Austin and Houston loosening building regulations to bring housing costs down as evidence that over-regulation is often the choke point. He frames the government's proper role with a sports metaphor—referees who set and enforce fair rules are essential, but when the refs outnumber the players and start tripping the best performer to even the score, the whole game gets worse. Threaded through is his contested thesis that this is driven by resentment and an "overproduction of elites," and a nuanced exchange with Drew about when market intervention is genuinely warranted versus when it becomes a barnacle. The team digs into three new Google research papers that, in the host's read, show large language models already outperforming human doctors—even in the messy, sometimes adversarial reality of real patient interactions. The host, careful to stay honest about both AI's risks and its capabilities, walks through the numbers: in a randomized clinical exam of 100 scenarios across 300 live video consultations with patient actors, graded by 20 board-certified physicians, the AI scored 83% to the doctors' 68%, with correct first-guess diagnoses at 91% vs 77%, clinical reasoning 90% vs 76%, guiding a physical exam over video 72 vs 39, and—strikingly—empathy rated higher for the AI (82 vs 71), echoing earlier findings in AI therapy studies. The AI also finished exams slightly faster. A second paper tracking patients across three visits with shifting symptoms found the AI's care rated appropriate 95% rising to 98% by the third visit, while human doctors actually got worse over follow-ups (72% to 81%)—which the host speculates is because doctors "back off too soon" once they think they've found the answer, whereas the AI keeps refining. A third paper used reinforcement learning across 57,000 simulated encounters—including thousands where patients actively try to deceive the doctor, hiding medications or insisting chest pain is just heartburn—with up to 60 conversational turns, producing a model that beats the already-superior AI 87.6% of the time head-to-head. The host argues the gap will likely keep growing, predicts that within a few years failing to at least consult AI could be treated as malpractice, and forecasts a near-term future of human-assisted-by-AI care rather than fully autonomous machines—because AI is probabilistic, not deterministic, so a human stays in the loop the way pilots remain in an autopilot cockpit. His bottom line: this is one of those cases where the technology may genuinely become what we hoped for, even if the economic transition to get there is turbulent. See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.
7,648 words · 7 speakers
Page updated · transcript and summary are machine-generated