Jacob's Leave of Absence FAQ

Previously on the blog: Jacob's College Application FAQ

What's going on?

After two years as a Stanford student, this fall I'm declaring a leave of absence!

Why?

It's overdetermined! Partially for personal reasons, because I've just been feeling a desire for a long time to take a break from being a full-time student, and this feels like the moment. And partially for the-world-is-on-fire reasons: I'm very concerned about AI risk, and I feel like my time would be better spent trying to make a difference there than taking classes.

So are you dropping out of college? That's so Stanford of you.

I'm not sure. Stanford is very flexible about letting students take quarters off without committing one way or the other, so I'm planning to take my life quarter-by-quarter. My current plan is to take a gap year and determine what makes sense after that. I'll get into the details more in the "The Future Beyond" section.

Fall 2026

So what are you actually doing this fall? Right now I'm pretty dedicated to AI safety!

My main focus for at least the next couple months (I'm pretty unsure after that) will be co-leading the Stanford AI Alignment club (which is rebranding soon!) and trying to make it go extremely well.

Lol, what? You're taking a leave of absence from Stanford to ... still be at Stanford? I agree that it sounds kind of insane! But I think it makes sense. (I would believe that, wouldn't I...)

I believe that AI safety is the most important problem in the world right now, and I have become convinced that university groups, when done well, can just be extremely effective at attracting more people to work on the problem, and helping them get the skills they need to do it. And Stanford should be a massive talent pool! But right now, the club is at a crossroads, and I think group organizing is one of those things where if I can really put a lot of time in, it will show.

When I got the opportunity to be one of the leaders, my gut reaction was, "This seems like a lot of work" but my other gut reaction was, "This seems like a massive opportunity for me to actually make a difference with hard work." The more conversations I had with people in the world of AI safety, the more I realized the incredible opportunity available.

So you think full-time Stanford organizing is how you can have the greatest positive impact on AI safety?

I'm not sure! I think if I was strictly optimizing for having the greatest positive impact on AI safety, I would probably do part-time organizing and part-time something else in AI safety, maybe policy advocacy/development work like I've been doing this summer.

But at the moment, I'm unsure exactly what I want to do, I'm slightly burnt out, and I think it would be beneficial to not overcommit and have some free time in my schedule. I'm sure it will all get filled in short order...

Where are you going to live if you're not gonna be an enrolled student but you're trying to be around Stanford a lot?

My hometown of Menlo Park, with my parents!

A crazy thing about my life during the last two years is that I went to high school in San Francisco, commuting 30 miles by Caltrain every weekday, but then I went to college within walking distance of my house (I swear I wasn't trying to do this; I applied to colleges all over the country) and still moved into the Stanford dorms. I think this was also a reasonable decision (the dorms are a really cool social vibe, and I particularly enjoyed Outdoor House last year) but this also does mean that I have a unique opportunity to participate in Stanford life actively while also hanging out more with my parents.

What do your parents think about this?

I like think they're happy to have me around... it's funny because my brother is going off to college this year (CU Boulder). They would have been empty nesters, but instead they have me coming back to live with them. Mom and Dad, if you would like me to add an official statement from you guys to this FAQ, just let me know.

Winter/Spring 2027

Are you going to do this Menlo Park / Stanford full-time AI safety organizing lifestyle for the whole year?

It would cost me at least $100 to do so.

So, this summer, working at MIRI and ensconced in the AI safety community, I had a lot of conversations with people about my career. And my sense is that there are a lot of good opportunities beyond just Stanford field-building, especially if I can successfully get Stanford AI Alignment to a place where my counterfactual impact is lower (aka the impact of me putting in additional hours is not as significant because there are enough awesome people working on it that the place feels like it runs itself). I'm particularly interested in policy and generalist roles, as I mentioned earlier, and I'm very open to the quirky. Now I feel like I'm just advertising myself ... here's my LinkedIn ... but I am quite open to opportunities starting in 2027 ...

Wait so what's this about the $100?

Okay so I ran into my former meta-manager (I'm actually not entirely sure of how to describe his role) William yesterday at Berkeley Bowl, one of the world's greatest grocery stores, we talked about my career. And I realized that I didn't want to sleepwalk into doing field-building forever. If I keep doing it full-time into the winter and the spring, as AGI approaches, it should be a thoughtful decision because it feels like it's really working.

So I promised him: "I officially agree to give you $100 on February 1, 2027 if I believe AI safety field-building activities comprise a plurality of my work efforts. The goal is to ensure that if I stay in field-building long-term, it is the result of a conscious, intentional, thoughtful choice instead of sleepwalking into it just because it’s comfortable / fun or because it confers immediate social validation."

Did William extort you into this?

No, I promised this of my own free will and volition. I was not coerced. William is great. $100 pales in comparison to the opportunity cost of months, and it's good to remember that.

Are you going to go back to school in the winter if you feel like field-building isn't working?

If I go back to school, I think it will be because my lifestyle overall feels like it's not working, and that I've overrated my desire to go out into the world. If field building just isn't working, this might be a sign that I should still take a leave of absence, but just do something else with the time.

I currently think it's pretty likely that I'll take a full gap year for 2026-27. With that said, I'm really fortunate to have Stanford as a safety net.

The Future Beyond

What are you going to do after that? Are you going to go back to school?

After that, I'm quite unsure what I'll do. I could see myself being torn between two considerations:

The case for going back to school

It's so logical and normal to take one year off and then go back to school! I won't be coming back to college when weirdly old, and I'll be able to maintain a normal college trajectory, graduating in 2029. After some time Out In The World, I'll have a better sense of what I want out of my college experience. I generally like learning, and in many ways I like the college environment.

Going back to school is, in some sense, my default plan. It's the option I'll choose if I don't have a better one (though I guess that's tautological). It just feels like living the good life that I want to live.

When I first started drafting this post on Stanford campus back in April, I took great pains to insist that "I really am planning to go back to school" and that I probably wouldn't drop out.

The case for not going back to school

The world might be too crazy. Short AI timelines don't feel like science fiction anymore, and I have this fear that even if I want to go back to school, it might feel obviously ridiculous, especially if I'm equipped to do meaningful work in AI safety.

My hope is that if I end up deciding not to go back, either (a) I will be doing something sufficiently cool or (b) the world will be sufficiently crazy that going back to college would be the crazy thing. But it also might be murkier, and I think it's likely that I'd be able to have a bigger positive impact on the world regardless by doing AI safety full-time. I might have to have a serious reckoning between positive world impact and life satisfaction.

As I sit here writing this in the heart of AIsafetyworld (Berkeley, California), surrounded by positive-impact-maxers and top AI forecasters ... I'm not going to lie, their prognosis of the future feels rather urgent and grim, and I feel like there are many ways to have an impact, to the point that going back to school would be lunacy.

But I don't know yet. Only time will tell.

Can I bet fake internet points on whether you go back to school and/or see the latest probability?

Sure, here's a Manifold market!

The Past

I have a theory: all of this AI safety talk is bluster, and you're just trying to escape school. What is your comment on that?

I've loved taking classes taught by a Numberphile-famous math professor and a former Undersecretary of State. I've loved eating from a buffet three meals a day. I've loved spending evenings and weekends gallivanting with my dear friends, and the campus events, and the sprawling beautiful campus, and it's hard not to feel like I've been so blessed in my life.

It's crazy how we acclimate to things. Life on someone's treadmill. We still find a way to complain about the thrice-daily bountiful buffets. I do feel like there's something that feels like consumption about college, especially the insane cost of tuition. Still, sometimes people ask what I think about Stanford, and I make some comment about how I don't love the culture, the way that people have one foot out the door toward founding their AI startup, and I just have to say that I think Stanford is wonderful. I don't wish that I went anywhere else for college.

I do want to try something new, and one thing I've certainly never done is manage a mid-size organization.

You said this feels like the moment you want to take time off of college. Why?

We're entering a critical moment in the field of AI—

Please don't talk about AI, just this once, talk about your college experience.

Going in as a freshman, I was really excited for college. Due to a few unlucky rejections, I spent a lot of my summer at home; because Stanford started in September, it was the longest summer break of my life. Summer break’s too long — that’s what I always said — and I couldn’t wait for Stanford. 

Freshman Fall: Camp Stanford

My first quarter was fascinating. I met awesome people. I went on the SPOT camping trip (which I highly recommend), and my first friends included the top Northern California Girl Scout cookie seller two years in a row, a girl who won $75,000 for discovering exoplanets. My sleep schedule shifted three hours forward. I was exploring what college was like; what it was like to live on my own. It felt like the world's most epic summer camp, and I was only somewhat prepared.

There was so much going on; one negative was that it felt like every moment had such a massive opportunity cost. And coming from my small high school, there were so many relationships to track.

I had had a summer job writing newsletters about the 2024 election for Manifold Markets. Initially I took a break: I told my boss I wanted to fully focus on Stanford life. But when I decided to started writing them again, my Stanford social life actually improved: suddenly I was this person with an identity external to Stanford. A political columnist. Someone with a tech job. 

Chappell Roan was everywhere that semester, but November 2024 was the month I first listened to the song "Pink Pony Club". It played in my mind as I went to Berkeley to run the Manifold Markets Decision Desk, to run a big election party and project results in states. I had dreamed of being a political analyst once, and I got to achieve my dream. 

I discovered something weird: my favorite moments were the ones I spent in San Francisco or Berkeley. I went to an auction party and won a year's worth of aerial silks lessons.

Freshman Winter and Spring: Losing My Faith

By the end of freshman year, I had had a lot of fun but was in a moral crisis. I felt that the environment at Stanford was somehow corrupting me. My high school interests, puzzles, math, and linguistics didn't compel me anymore, but instead of being displaced by new interests (with the possible exception of theater ... Playback Theater and Dramatic Tensions: Theater and Marketplace were awesome), I felt kind of empty.

I felt like a lot of Stanford classes were full of zombies on laptops, and I wasn't learning as much as I could be. compared to high school, I didn't like the increased ratio of homework to classes. I liked the environment, but felt kind of excited for classes to be over.

 I realized I valued three things: making the world better, having an interesting life narrative, and just being happy. but now, with all these options opened up by being a Stanford student, I didn't know what to do.

Freshman Summer: Not Long Enough

Summer break's too long — that's what I always said — but this was the first summer in my life that I wanted to be longer.

I almost didn't go to the YC conference, because I was skeptical that doing a startup was really a way to have a positive social impact. But while I was there, I had an epiphany. I can go out on my own and define my life on my terms. it used to be that I couldn't take a gap year because I would just be sad and lonely sitting around on my parents' couch. But now, partially thanks to my time at Proof School, and Stanford, and at various parties in San Francisco and Berkeley, I had the social infrastructure to make my life interesting without being sad and lonely.

Still, I had no plan for what I would do, and I'm kind of a skeptic of plan-free gap years. Also, my dorm assignment for the next year was extremely good. I would be living with my best friends in Outdoor House.

Subsequently, I had the best summer of my life. I called it the America Tour, traveling to 16 states and actually making money doing it, particularly thanks to a math camp counselor job in Missouri. And I wasn't traveling just to travel. In each place, I was attending a conference or visiting friends or something. Stanford's quarter system meant that the schedule was offset from most of the other semester schools, so I saw my friends at East Coast colleges, which was very fun.

What shocked me when I looked at my happiness numbers is that they were lower during my crazy summer than at Stanford.

That was when I realized my daily happiness was really only one part of the story. Another part of the story is personal life satisfaction. It is correlated with happiness, but not perfectly: life satisfaction is relatively higher when I go on crazy adventures or work hard on cool projects; happiness is relatively higher when I hang out with friends and consume (rather than produce) cool things and such. The America Tour was when I realized that personal life satisfaction was what I wanted to chase. 

But one failure of the America tour was that I felt like I didn't accomplish anything. I felt like it was time to lock in, I still didn't have a passion, but I looked at the world around me and some of the people I met, and realized I should try to get into AI safety.

Sophomore Fall: Anticlimax

When I got to Stanford during freshman year there was a whole camping trip and orientation week of pageantry, but sophomore year was very anticlimactic: I arrived at my dorm one night, and classes started the next day. 

A few of my friends were gone for startups or whatnot. Outdoor House, my dorm with all my friends, was, in fact, awesome.

My mom thought my issue with college was that I wasn’t taking classes that challenged me. So during sophomore fall, I plunged myself into the deep end, taking an oddball graduate-level machine learning class with a charismatic professor. It was called CS329H: Machine Learning from Human Preferences. While doing this, I applied to everything I could think of in AI safety.

It was hard. Kind of cool to experience that. Ultimately I got an A, but only because the class was graded out of 120%. And because of my awesome project partner Justin who did most of the work. 

But at the end, I was like … okay, mom. I am capable of a hard class. I got through it. 

But I didn’t like Machine Learning from Human Preferences. I liked sidequesting. 

In December, as part of the MARS program, I went to Europe for the first time in my life. the homework from the class was like an albatross around my neck. But it was still cool. I booked my flight to London with an 8-hour layover in Amsterdam. During the layover, my friend Caitlin and I rented bikes because we thought it was a bike-friendly city; there were definitely a lot of bikes, but they weren't friendly.

I spent a lot of winter break traveling with my family. I worked on my AI interpretability research and my personal website, jacob.show. I ran out of time to write the FAQ on that website, ironically, but people still like the website. But I wished I had more time to work on it.

Sophomore Winter: Winter of Sidequest

Winter 2026 was the Winter of Sidequest. I intentionally took easy classes and sidequested every weekend. San Francisco, Berkeley, Tahoe, etc. I was honestly just much happier than when I was spending all my time doing homework, and some of the events introduced me to people that allowed me to get further in AI safety.

I spent spring break not travelling anywhere but working on a project: HIPPOCAMPUS, one of my dreams. After a week I made a very bug-ridden version. Then I ran out of time to devote. I wish the break was longer.

Sophomore Spring: Busiest Quarter of My Life

I won the lottery and found the coolest Monday lineup of classes: Social Dance II, Arms Control with the former Undersecretary of State for Arms Control, the American Road Trip. I also took MATH 113 and CS 231n, scheduled at the same time as each other, because I thought that they wouldn't be that hard because I already knew the content. I was half right and half horribly wrong. MATH 113 wasn't so bad, but CS 231n, even though I had already taken a class based on it in high school, was very time-consuming, and added a great deal of stress to my life.

I also became the co-president of Stanford AI Alignment, and went on a listening tour to try to talk to everyone I could who was involved with the club or had advice about how to run a good AI safety clubs, which was enormously valuable. We ran a revitalization dinner. It was wonderful.

And I tried to enjoy my life. I staged an elaborate fake wedding. I went on trips with Outdoor House; surfing, whitewater rafting.

I got this feeling like I loved the place, but it was time to go. It was the feeling that I felt in my last year of Proof School. It was the end of a chapter.

When asked why she dropped out of college as a junior, Cheryl Wheeler said "Because I was in 15th grade, and it felt like enough already." That's how I started to feel. I'd been in classes for 14 years, and it seems totally reasonable to spend one year trying something else. And let's see, you said I can't talk about AI?

So it felt like time to leave?

I often think about the advice of psychologists Dan Gilbert and Annie Duke. 

Dan Gilbert says people are happier when they do fewer things better. If he can spend 2x the time on his teaching responsibilities to teach great classes instead of good ones, he’ll have more fun, and enjoy the whole thing more. 

I always try to implement his advice but end up with too much on my plate anyway. At some point, I have to make cuts. And when I think about my life goals, and the costs and benefits of different things, the answer is clear: the classes should go. 

Annie Duke says people are happier on the margin when they quit things. There is a lot of aversion to quitting. But people regret not quitting a lot more than they regret quitting. 

I spent freshman year realizing that I was unsatisfied taking classes at Stanford, even though they were often pretty cool, and I preferred my time sidequesting in SF and Berkeley. I spent sophomore year trying different things to fix it. 

Now it's time to try quitting classes!

At Stanford, it's pretty easy to try. I can always come crawling back if I hate it. 

I don't know what's going to happen. I don't know whether I'm prepared. And that’s scary. But I think it’s the good kind of scary :)

Note: I'm running a blog-a-thon today, and I only spent a few hours writing this post! It's one of three posts that I wrote and published today. So ... apologies if it's a bit rougher than usual. I may make further edits to it in the future.