Hi! I’m Jason Benn, and this is where I learn in public 📚

My goal with this website is for friends or strangers to discover that I am curious about similar things as them! If that's you, email me - I'm based in San Francisco (by 16th & Dolores) and am always down for "coffee" (whiskey). If you're too far away but still have strong opinions about any of the below then I also really appreciate reading recommendations.

If you're wondering where to start:
  • If you're looking to get into software engineering, here's how I learned to code.
  • If you like communal living, check out my work-in-progress Principles for a Great Community, based on interviews with people who've been doing this for ages.
  • Here's my newer, updated homepage, where I'm documenting some experiments in personal thought technology, which is really a pretentious way to complain about the fact that I don't remember 99% of what I read.

Things I'm learning about:

Communal living. I've been my best self since moving into my 20-person house. Would this be even cooler if we were even larger? We'd have enough social gravity to attract brilliant and fascinating people from around the globe, we'd have the financial capability to buy e.g., an old cruise ship (seriously!), and we'd be a lot more stable than the average urban community, which seem to be distressingly short-lived. How can we make a community that is supportive of big life transitions, like starting a family? How might we design spaces for both extremes of the introvert/extravert spectrum? This lifestyle is fringe - how can we normalize it for more people? What can/should we stand for and what difference can we make in this world? Seriously though, should we find an old cruise ship to buy and turn it into a floating city? Etc, etc.

Thought technology. It's always bothered me that I have forgotten most of what I've learned. What a waste of time! Hundreds of books, thousands of conversations, thousands of articles - most of them I wouldn't be able to describe a month afterwards. Something I've been thinking about recently is how to integrate everything I consume into a coherent, well-articulate worldview, and then to revisit and develop those views over time. My hypothesis is that a system that did this well would improve my thoughtfulness. The goal is to have more confidence in my beliefs, better conversations, improved judgment, broader and deeper knowledge about things I care about, better research and investment ideas, etc, etc.

Machine learning. I'm particularly interested in modeling uncertainty, natural language, interpretibility, fairness, and where ML intersects with computer science. My Man Crush Monday is Jeff Dean, the head of research at Google, because a) he said "If Google were created from scratch today, much of it would be learned, not coded" and "Around 10% of Google's 25,000 developers are proficient in ML; it should be 100%", and b) he wrote one of my top 5 favorite papers. I'm also the Lead (read: only) Machine Learning Engineer at Sourceress, and am currently organizing a 10 week course on MLE.

Format of this site

As I learn more about each of these scintillating topics, which is most days, this site automagically updates itself. Each of these topic pages have two parts: 1) a summary of my current, loosely-held opinions about that thing, and 2) a stream of my reading and writing on the topic - part 2 being the "learn in public" part. Clicking through links in the content stream will reveal the remainder of a post; posts range from short essays to bookmarks accompanied by notes and highlights (here's an example).

Newsletter: I periodically collect the top 5% of the things I've discovered and summarize them in a simple newsletter called the High 5. You can sign up here. 🙏

All opinions are my own, and may or may not be shared by colleagues, friends, or my future self. That's the goal!

Archived: things I used to spend more time learning about:

The future of work. An unemployment crisis might be right around the bend and we're shockingly unprepared for it. What kinds of work are most vulnerable to automation? How might we help the structurally unemployed? What technology related to finding and evaluating people seems inevitable? What technology related to learning employable skills seems inevitable? How and why does quantifying people go wrong? Will the bootcamp model survive and/or spread to other, non-tech industries? When and in what form will UBI come save our collective bacon?
> I now think this might just lower the quality of jobs, rather than eliminate them altogether.

Living well. What's the best way to learn stuff? How hard should I work at various points in my life? How can we build systems throughout our lives such that living in accordance with our values is as low-friction as possible?
> I still think most people are living far under their potential, and I've reformulated this as "Thought technology", above.