1. The Venture Clock Ticks Fast
Read and weep. Giovanni Colella, a psychiatrist and entrepreneur, tells the tale of Kintsugi, a company that built a platform to analyze speech for signs of mental illness. The technology was “extraordinary.” The FDA reviewed it in good faith. Seven years and $30m later, Kintsugi failed. Why?
When you force a dynamic system [an AI biomarker, like Kintsugi] into a framework designed for static ones [what the FDA was built to review], you generate friction that is not incidental. It is structural. And structural friction, when it compounds over years, becomes fatal. Venture capital runs on a seven-year horizon. FDA clearance for a novel AI biomarker runs on a different clock entirely. There is no synchronization between these two systems…The clocks run at different speeds, and when the capital clock runs out before the regulatory clock finishes, the company disappears.
The friction leads most startups to take a worse path: “They build a wellness application instead of a medical device. They strip the clinical claims from their product. They avoid any language that would trigger regulatory scrutiny. They ship quickly, iterate in the market, and generate revenue while their more scrupulous competitors are trapped in a regulatory waiting room burning money.”
2. Gene Therapy’s Sticker Shock
The Incidental Economist offers a stark assessment of gene therapy access: “they represent some of the greatest medical breakthroughs of the last 20 years. And yet, in the United States (U.S.), many patients who could benefit from them never receive them—not because the science fails––but because our health insurance system does.” What gives? A USC policy paper offers an answer:
The U.S. system is built around chronic treatments paid over time and thus is poorly structured for one-time therapies whose value accrues over decades. Cell and gene therapies (CGTs) create extreme stress on the current reimbursement system by combining very high up-front prices with long, uncertain benefit horizons.
3. Moonshots with Layovers
Forget the cancer moonshot. Rep. Jake Auchincloss (D-MA) wants an Alzheimer’s moonshot – not only for the destination, but the stops along the way:
We got semiconductors in part because NASA was like, “we want to go to the moon. It’s so hard. We’re just going to, and because we’re taking on the absolute hardest challenge, we’re going to ratchet up a bunch of other associated technology and science with us”. And by trying to cure Alzheimer’s, I think we would be astounded by the number of other really smart and cool things that happen along the way. “Oh my goodness, we accidentally cured ALS too. That’s the kind of thing that happens when you take on the absolute hardest problem.”
4. Influencers, Meet Regulators
The FDA has set a new standard against “false and misleading claims.” The agency issued a warning letter to ImmunityBio over a set of remarks its Chief Scientific and Medical Officer made on a podcast. One example from the warning letter:
The podcast, which features the host introducing Dr. Patrick Soon-Shiong as the Global Chief Medical and Technology Officer of ImmunityBio, makes representations about the efficacy of Anktiva such as: DR. SOON-SHIONG (13:27): “[Interleukin-15 (IL-15) is a molecule that] stimulates the natural killer (NK) cell and the T cell…the most important molecule that could cure cancer…nobody could figure out how to get IL-15 into your body with a single jab, and that is Anktiva.”
5. On-Demand Ad Boards
Pharma Leaders, a substack, offers a delightfully jaded view of AI’s progress in “disrupting” pharma. Spoiler alert: it isn’t. As the author says, “The tools are new. The problems are old.” But as the piece assesses how different pharma companies stack up, one idea stands out – and it suggests how much could still change:
Your medical team wants to run a KOL advisory board. Good idea. You need external scientific input before a launch. Standard process. Industry guidance puts the minimum planning time for a pharma advisory board at eight to twelve weeks, with the full end-to-end process typically running three to six months. That’s the baseline. Most teams don’t beat it.
[One startup has built] a platform that creates AI simulations of healthcare professional archetypes, trained on over 200 million medical publications and 400,000 clinical trials, that can run virtual advisory sessions. The idea is to generate and stress-test hypotheses first, then bring actual KOLs in to validate the most commercially relevant ones.
Caution is advised: “It only works if you have clear questions to begin with. If you don’t know what you’re trying to learn, the simulation just helps you generate bad insights faster.”