[devonsolomon]

Great experiment! It’s important to highlight that even in rigorously blinded studies which find a drug ineffective on average, there can still be genuine responders—this is known scientifically as “heterogeneous treatment effects.” Essentially, individuals vary in genetics, metabolism, and neurochemistry, which can cause meaningful differences in drug responsiveness. Thus, an N=1 trial, like yours, might indeed reveal real personal benefits that wouldn’t appear in a population-level analysis.

However, systematically performing robust N=1 trials individually across multiple substances can be impractical—too laborious and time-consuming for most people. An interesting business model might be a startup that facilitates personalized N=1 experiments at scale: providing users with high-quality compounds, matched placebos, structured dosing schedules, and data-tracking tools. This could empower more individuals to accurately assess personal efficacy across a wide variety of substances, potentially offering valuable, personalized insights that large-scale clinical trials can’t capture.

[davidanekstein]

I made an app for this called Reflect [1]. You can run self guided experiments for anything you can model as a metric in the app. I just wrote about an experiment I did with nootropic coffee [2].

[1] https://apps.apple.com/us/app/reflect-track-anything/id64638...

[2] https://open.substack.com/pub/reflectapp/p/my-experience-wit...

[devonsolomon]

Love it - Great work. I’m gonna give it a go.

[kqr]

I like this idea! To build on it, the schedules could be non-random (systematic with random offset) and test combinations of substances in a way that maximises information both in the cases where substances are independent and when they have suspected interactions.

[devonsolomon]

Hard to pull off! But I imagine the wave of Bryan Johnson fans would be game.