Likes & Dislikes
A distilled view of Alex's preferences — drawn from the Longevity Pledge, Forbes essays, Forever.ai posts, and public interviews.
👍 Likes
Longevity science with measurable endpoints
Real biomarkers, proper trials, rigorous statistics — not vibes-based wellness.
GPU-accelerated deep learning on biology
The original thesis: compute × generative models × biomedical data beats intuition.
End-to-end platforms over point tools
PandaOmics → Chemistry42 → InClinico: target to clinical-trial simulation in one loop.
Publishing in top journals (Nature, Nature Biotech, Nature Medicine)
Peer review matters. First AI-drug Phase II data landed in Nature Medicine on purpose.
Multicultural teams and cross-border science
Labs and partners across US, China, UAE, Saudi Arabia, Hong Kong, Canada, and Europe.
Open-source benchmarks (MOSES, etc.)
Benchmarks beat marketing. Reproducibility is how the field actually moves.
Speaking at JPMorgan Healthcare, NVIDIA GTC, ARDD, WEF
Meeting capital allocators, compute providers, scientists, and policymakers in person.
Automated robotic labs (Suzhou facility)
Closed-loop AI + wet lab. Automation is the other half of AI drug discovery.
Healthspan > lifespan > peakspan
The frontier is peak productive capacity, not just years alive.
AI co-authorship disclosure
Co-authored a peer-reviewed paper with ChatGPT — and disclosed it openly.
👎 Dislikes
Vague anti-aging claims without data
'Longevity supplements' with no RCTs, no clocks, no mechanism of action.
Drug discovery without computation
Brute-force screening campaigns in 2026 — the world has moved on.
Short-termism in biotech investing
Aging and fibrosis don't fit a 12-month IRR — but the market will exist.
AI hype without peer-reviewed results
A demo is not a drug. Patents are not clinical evidence. Publish.
Siloed data
Proprietary datasets locked away from federated learning slow the whole field.
Regulatory paralysis on aging as an indication
Aging is not a disease bucket at FDA — this blocks the biggest prize in medicine.
Over-reliance on a single target or mechanism
Single-target bets lose; platforms win. That's why we build platforms.
Fear-of-missing-out on foundation models
Big generic LLMs ≠ domain-specific generative chemistry. Physics-aware matters.
Travel downtime
That's why it's Wi-Fi, papers, and keynote drafting at 30,000 feet.
'First mover' claims without Phase II data
Many claim first-in-class AI drugs. Very few have in-human efficacy.