Alex K. Chen | Notion

I live in Boston (socially adjacent with https://harvardst.co/) and also am very culturally SF. My overwhelming interest is in brain health and neuromodulation - though I am broadly interested in everything (esp complexity) and the “weak ties” ⇒ serendipity connection. I am a Venture Fellow at Healthspan Capital.

near-complete backup of crsociety.org which used to be very important to me: https://crsociety2-archive.pages.dev/forum/3-forums/ | lots of longevity content: forum.longevitybase.org

I sit at the periphery of everything — high betweenness-centrality, trading on the strength of weak ties. AI is making the illegible legible; this page is where I lean into that.

I am friends with many in augmentationlab.org and the directorysf.com grouphouses (+homebrew.nyc). I helped catalyze the early formation of a major SF grouphouse and catalyzed the early-stage life trajectories and careers of many.

I carried my team at https://flagship-hackathon.com/reviews/round-3.html

my core belief is that improving the human brain (by reducing all the “dumb factors that limit it”/reducing the amount of time the brain spends in “down states” + maintaining brain health & energy) is the best route to preventing human disempowerment (either by humans themselves, microplastics/pollution/noise/”group thinking”, the default mode network, or by AI). Neuromodulation for cognitive control (some of the famous authors in cognitive control include David Badre and Wilma [the Agile Mind]).

The attention residual. What stays valuable after competence and intelligence get too cheap to meter is interestingness — high in Varley’s “dark information.” That residual is what this whole page is about.

Some demos I am building:

I create neurofeedback apps (connected with openBCI cyton) to help modulate the bran out of drowsiness. My main passion (beyond breadth/serendipity/”why diversity is all you need”) is about using all the methods/modalities to make the human brain suck less (important for getting the brain to properly adapt to AI).

(the first iteration is here: https://www.linkedin.com/posts/edubaem_hardmode-hardmode-ugcPost-7436769050816344064-0vcY/).

A next iteration is this (though online demo uses synthetic data and not-real-time EEG data):

🌌 Aurora — live neurofeedback demo — 2-channel EEG driving generative music and visuals in a closed loop. Alpha reward, peak alpha frequency, aperiodic (χ) exponent, EMG artifact gating that freezes feedback rather than lying to you, and a small RL loop that learns which pitch/light combinations work for you. Runs in a browser; demo mode needs no headset.