Max Hodak: Average Is Not Good Enough

Video summary

Max Hodak explains how Science’s Helix infrastructure, hiring system, and retinal implant work turn speed into startup advantage.

At Startup School 2026, Science CEO Max Hodak argues that startups usually fail less from bad technology than from weak operating systems. Science’s retinal prosthesis has completed major clinical trials, and one patient used it to read a 300-page novel, but Hodak focuses on the infrastructure behind such work: procurement, budgeting, experiment costing, hiring, and performance feedback. A $3,000 power supply can be rational when delays cost a team $100,000 per week, while one wafer iteration may cost $40,000. Science’s Helix software connects purchasing, manufacturing, quality, and recruiting. Its hiring funnel uses companywide voting, phone screens, homework, and interviews; 17% reach phone screens, with judgment, horsepower, and agency as core criteria. Hodak’s central lesson is that iteration speed determines success, founders cannot delegate judgment, and there is no universal startup playbook.

Chapters

1. 0:00Infrastructure at Startups: Max Hodak’s Brain-Computer Interface Background
2. 0:51Science’s Retinal Implant: A 300-Page Novel Read by One Patient
3. 3:25How Your 17th Employee Buys Things: The $3,000 Power Supply
4. 6:56What Does an Experiment Actually Cost?: A $40,000 Wafer Iteration
5. 10:18Building a Rigorous Hiring Process: 17% Reach Phone Screens
6. 13:03Judgment, Horsepower, and Agency: Four Hiring Stages
7. 16:06Rethinking Performance Reviews: Helix IGEN Reviews
8. 18:51Rate of Iteration Separates Success From Failure: Weekly Versus Monthly Learning
9. 20:38You Can’t Delegate Your Judgment: Differentiated Decisions Create Outliers
10. 22:47Action Produces Information: Unblocking a Company’s Operating System
11. 25:10Q&A: Exceptional ability, engineering hiring, and the Neurolink lesson
12. 29:36Q&A: Judgment, consciousness, and interdisciplinary BCI engineering
13. 35:11Q&A: AI-native operations, regulatory automation, and custom software
14. 39:31Q&A: Custom infrastructure, neural longevity, and anti-collusion reviews
15. 42:58Q&A: Funding Neurotech experiments, reaching revenue, and independent judgment
16. 47:46Q&A: BCI power, implant packaging, and Science’s retinal implant origin
17. 52:15Q&A: Science Integrates Transhumanists and Clinicians Across Six Countries