Matias Serebrinsky borrows the OSS's Cold War ideal-spy test—'a PhD that can win a bar fight'—to describe SciMed Ventures' founder target. He wants technical depth plus ruthless execution speed, and says credentialism matters less than those traits. The phrase captures the dual requirement: scientific understanding and operator instinct.
SciMed chose to accept more market and regulatory risk in psychedelics because the clinical science looked strong, Matias says. He points to uncertainty over payer reimbursement and pharma acquisitions as the tradeoff for underwriting less technical risk. The FDA's openness, including breakthrough designations, made that bet feel more reasonable.
Matias expects classic psychedelics may reach approval first, but says best-in-class compounds may be non-hallucinogenic. He cites Gilgamesh's AbbVie-linked program and Delix's positive Phase 1b data as evidence for this newer generation. The thesis is that first-in-class does not stay best-in-class by default, and he is bullish on the category.
Matias says old medical devices were single-indication assets, but brain-computer interfaces are becoming horizontal platforms. Miniaturization plus better algorithms enable closed-loop read-and-stimulate systems, and neurotech's circuit-level interface makes many psychiatric applications plausible. That shift is why he is excited about devices despite their historical reputation.
Matias warns founders often chase huge indications like Alzheimer's without a fast way to test their hypothesis. He prefers plans that validate in smaller, biomarker-defined populations first, then expand once the company is de-risked. He frames validation as the core early startup job, especially until better surrogate endpoints exist.
Matias says precision neuroscience is limited by messy, non-normalized data and by a missing business model. Wearables and other data exist, but few multidisciplinary teams turn them into a source of truth or deploy them with aligned incentives. He lists these practical gaps alongside the brain's underlying complexity.
Matias says many hot biotech companies use AI without being AI-first; the harder problem is delivery. For RNA and oligo therapies, he highlights crossing the blood-brain barrier with brain shuttles and getting compounds inside the right cell. Those enabling technologies are where SciMed is seeing interesting companies.
Matias says his time at NVIDIA gave him an insider view of AI progress—more compute, better data, better models—converging with neurobiology. That visibility helped convince him a venture fund focused on brain and mental health could work. He calls it one of the lucky breaks behind SciMed's launch.
Matias says storytelling is not just for fundraising. Founders must sell constantly to A-plus hires, pharma partners, and public-market analysts, and they need to transfer genuine excitement rather than merely polish slides. He calls it a critical, permanent founder skill because every stakeholder has other options.
Matias says SciMed's podcast and Future Neuro Founders workshop exist to inspire scientists to start companies. The podcast is also a selfish learning tool: he would keep recording even with one listener because he gains so much from the conversations. He calls the effort both selfless and extremely selfish.
