Welcome! I’m Jason Benn, and these days I’m mainly known as the founder of the Neighborhood, a one-square mile community of communities in SF that aims to recapture the vibes of a university campus, but for all generations. The Neighborhood is part of a local movement composed of coliving (one big rented house), cohousing (multiple adjacent owned houses), third spaces (community centers), and good vibes (I bump into friends once every 40 minutes around Alamo Square, Lower Haight, NoPa, Duboce, and Hayes Valley).
When I co-founded the Archive coliving house in 2017, I didn’t suspect that it would have such an impact on me. Living around people that inspire me destroyed my limiting beliefs even as we all grew more connected. But I also knew I’d never be able to raise a family in a coliving house. So, after the Archive, I began to experiment. As I stumbled upon little successes and educational failures, my aspirations have grown. I share those lessons here in the hopes that others will also be inspired to live closer to their friends!
If you'd like to subscribe to updates, follow me on Twitter or send me your email:
Thank you! I'll take good care of it.
Here are my past updates:
Nov 22 🤯 #8: The Summer of Spaces
Selected writing about community. My favorite posts, essays, and tweets about community that are representative of my thinking at the time:
Jul 23 🛠 The unconference toolbox
Nov 22 🧭 You’re invited to Explorers Club
Jun 21 😤 I quit my job...!
Jun 20 🗺 A mindmap of my community beliefs
My engineering career, in reverse order
Before this expedition into community, I was a Machine Learning Researcher at Generally Intelligent, where I helped build a self-supervised system that solves infant cognitive development milestones from object permanence until we reach human-level general intelligence. I also designed and prototyped a collective intelligence tool for AI research called Exobrain.
Before that I was the Founding Machine Learning Engineer at Sourceress, where I trained the first models, built the model pipeline, hired and trained a data labeling team based on case studies I conducted of successful ML organizations, and then automated the entire process. These ML models worked well and we eventually sold our code, models, and data to a handful of other organizations.
Prior to that I was a software engineer at the nomadic university startup Minerva, where I implemented the grading system and the classroom product, which is essentially a pedagogically-inclined Zoom.
During these 8 years I alternated between full-time and part-time teaching myself how to code. I actually started my career in a dead-end finance job with no technical skills!
“Jason Benn’s transformation is nothing short of astonishing.” — Cal Newport, Deep Work (p8-11)
In 2013, at 23 and motivated by a quarter-life crisis, I explained to my parents via PowerPoint that I thought it was a good idea to read programming textbooks in the basement until I got a job as a programmer. Fortunately, it only took six weeks of studying before I discovered Dev Bootcamp, which 3x’d the rate of my progress. Seven months later, I had my first job. This was the year I discovered my agency.
Ever since, I’ve found it quite enjoyable to learn. Between 2013 and 2021 I averaged 5-10 hours of studying per week. I designed and then completed curriculums in programming (1 year), computer science (3 years), machine learning engineering (3 years), and AI research (2 years). My guiding principle was to simply maximize my usefulness, so that I could negotiate for working conditions that are most linked to career satisfaction: autonomy, mastery, and purpose.
I’m glad I never saw this doc back when I was getting started, though. It would’ve scared the crap out of me.
Another direction that has me excited is building an Ivan Illich-style “learning web” around AI tutors. Studies have shown that the average tutored student performs better than 98 percent of non-tutored students. When you look at historical geniuses, you can’t help but notice that most were essentially raised by their tutors. Incredibly, and just in the last few years, large language models have become powerful enough to be this for everyone.
I think edtech way underrates the importance of colocating students, and that community structures like the Neighborhood will have an important role to play. Education is a function of motivation, and motivation is often social!
Contact. I think the future of making friends is online first, offline later. That's why this page exists! In case it wasn’t clear earlier, I'm based in San Francisco.
Inner work. San Francisco has a classic techie → hippie character arc and maybe I’m not as unique as I thought. I’m more into emotional fluidity than meditation, and I know it’s working when I feel a greater capacity for love.
Tools for thought. Genius is overrated. Thinking quickly, which is a talent, is trumped by thinking well, which is a skill.
I’m also fascinated by worldview building. How might one incrementally build up an original body of thought? I can't stand that I don't remember 99% of what I read.
Drumming. I played so much Rock Band in college that I was left with no choice but to become a real drummer. I peaked when I recorded a rock album with my dad called The Basement Sessions.
Books. My favorite spot to read is in my 6-person hammock, which I usually set up in Buena Vista Park. I love to organize book clubs and reading retreats.
Art. I was obsessed with drawing and painting dinosaurs as a kid. I still dabble.
🐿 Here's a Google Photos album.
Productivity. They say that gentlemen don't give advice. As everyone knows, I'm no gentleman. Here are some strategies that have made me more effective.
Personal experimentation. My body is a
temple testbed. One month I drank nothing but Soylent.
Legacy. Tina Kim and I unofficially broke the Guinness World Record in the undeniably prestigious & competitive "Most lemons caught blindfolded in 30 seconds", with 35.