Today after 12+ years at Y Combinator I’ll be transitioning to a Partner Emeritus position. This role allows me to continue to do office hours with the 1000+ companies I’ve worked with in the past decade while giving me the free time to explore new adventures. It also means that the w25 batch was my last batch funding new YC companies. The next adventure I’m excited to pursue (after taking the summer to relax) is how I can help government better serve its citizens. Thank you to the countless friends who have been pushing me in this direction for years. Government was the passion of my youth and I’m excited to reengage. Thank you to @ycombinator @paulg and @jesslivingston for changing my life.

I spent many years telling friends that our government is slow and hard. And that’s probably better than a government that is too fast and too effective. I still believe this 100% but damn slow and hard can be a pain in the butt. Just gotta keep motivated.

The amount of information available online about what is happening in SF is almost endless. The SF city government is refreshingly transparent. Ai makes this information accessible and understandable. I encourage everyone to pick their favorite topic and let their Ai agents go wild with deep research. You’ll learn a ton about your city! Also as a note for people slinging large numbers of sunshine requests. Give this a try! You learn a lot more than you think and city workers can spend more time working hard for the city.

I'm concerned by the relatively small number of people in government who are *incentivized* to improve its operations. It's like we expect government workers - out of the kindness of their hearts - to improve the institutions they run. That seems like a silly expectation and if that's what we believe, we aren't living in the real world. We should expect people to look out for themselves, their family, and their community. When we want more than that we need to figure out a way incentivize it. Btw - you can replace the word "government" in this note with non-profit, business, or any other human institution.

When an investor tries to use their dollars to direct the operations of an organization... that doesn't really work in startup world. Does it work in the non-profit world?

People expect new business projects to fail as often as they succeed — sometimes more often. They expect new government projects to succeed most of the time. That’s a strange double standard. Businesses and governments are both run by people, and people are fallible.

Check this out :)

San Francisco’s come back is real but we still have a serious problem. The unfortunately truth is that thousands of Fentanyl addicts live here. Most are not born and raised in SF but they have come to use drugs and are housed in our shelters and SROs. They take a drug that is 30-50x more addictive than heroin and as a result are not interested in drug treatment which is offered to them regularly. As long as these people live here and are not in treatment, the conditions on the street will be impacted by their presence. As long as they are here and using, dealers will figure out how to sell them drugs. I dont judge these addicts, they aren’t violent and I can’t imagine trying to combat a fentanyl addiction. But that doesn’t change the fact that their presence creates negative externalities that impact our city everyday. Too many people are talking about the symptoms and not the problem.

Experts with no data are just making guesses.

I’ve learned that a city that provides great services like San Francisco can be a magnet for drug users from the region, state, and country. When that city has cheap drugs, decriminalized drug use, warm nights, and the prevailing drug is 30-50x more potent than Heroin (Fentanyl) you have a problem that spending more money cannot solve. I wish I had an easy or compassionate answer. I don’t. Other communities who don’t have the resources or the caring citizens that San Francisco has will compassionately send drug users here to try to help them. And users talk to users about the best places to do drugs and the word spreads. Once again - San Francisco doesn’t have a crime or homeless problem. We have a drug problem. And Fentanyl is one hell of a problem.

I’ve been playing around with a new mental model. Think of government—federal, state, and local—as a set of big businesses. It’s not a perfect analogy, but it’s useful. Their policy decisions direct huge sums to for-profit and nonprofit organizations. Political science majors like myself start careers in government or campaigns, creating policy and directing government cash. Later, they can work for organizations receiving that money. This is how some monetize their expertise to achieve their “exit”—the ability to buy a house and pay for retirement. Millions of jobs depend on this system. It’s clear why changing policy is hard. Behind many policies is a massive amount of money and jobs. People don’t like when you screw with either. As a result, expensive policies that don’t work can survive indefinitely. Sound familiar?

You know - Sid @sytses told me a couple years ago that you can just do things. It’s advice I gave to founders everyday but needed to hear it myself. Turns out he is right :) Thanks Sid!

I can’t believe how often people ask me a question and my reply is - have you tried using AI? AI is nowhere near mainstream yet.

There is a lot of bullshit misinformation in politics. AI makes facts accessible without needing a team of PhD researchers working 24/7. I wish people in campaigns and government staffers spent less time responding / caving into the misinformation and more time accessing the truth and sharing it. There are more than enough real problems to solve. Solving / responding to fake problems is a waste of time.