Hao and I stopped working on Tolus in September.
The site is still up and we’re leaving it there for a while, so dormant is the right word, not dead. But we’re not building it. Which is a slightly weird thing to write on a blog where half the posts are, one way or another, about why Tolus has to exist.
Here’s the short version. We ran the pilot, we talked to a lot of teachers, and we found out that we believed in the idea and hated the customer. To be clear, the teachers were great, and most of them genuinely wanted this to exist. But the teacher isn’t the customer. The customer is the school, the institution that actually signs the check, and selling to it was shit.
The unit of time in education is the semester, and the unit of money is a budget that was set before you showed up. Tyler Bosmeny said as much on stage at Startup School, long sales cycles and broke customers and procurement, and I wrote it down and nodded and kept going. Knowing a thing and having lived a thing are not remotely the same.
The thing is, we still believe in the idea. That’s what makes it uncomfortable. If we’d found out the idea was wrong this would be an easy post. Instead we found out that you can believe in the mission completely and still not want to spend the next 10 years selling it to the people you’d have to sell it to, at the speed they move.
Around the time we were working this out, Diana Hu (YC partner) reached out, and we ended up talking for a while about, essentially, the value of living life. She said, roughly, that Hao and I are talented and that we rush into our ideas, and that we should take our time. Not take our time building. Take our time finding the idea. Get seasoned, live some life, and when the one that feels right shows up, go all in on that one.
Which stung, mostly because she was right and we knew it. We’d just spent 7 months proving it at full price.
So I’ve been thinking about what rushing actually means, because it’s an interesting word for it. It isn’t about speed. Moving fast is fine, moving fast is the whole point. Rushing is putting the clock on the wrong thing. Believing you have to find the startup idea now, that everyone your age already has theirs, and that if you’re not pushing something by 18 you’re behind.
The last post was partly about how a weekend around people raising rounds at 18 moved my sense of normal. That was useful. It’s also exactly the thing that makes you rush.
What we were missing I’ve started calling seasoning, for lack of a better word. It’s the stuff you only know because you’ve done things. Not knowledge, you can get knowledge from a book. Seasoning is knowing you hate the customer before you’ve built the product. It’s knowing what a funded company looks like from the inside before you try to run one. It’s having seen enough that when a new idea shows up you have something to compare it to. The right idea only looks right next to enough wrong ones.
We had basically none of it, and the way we got our first unit of it was by building a company around the wrong customer for 7 months. That’s the expensive way. The question that’s been bugging me is whether there’s a cheaper one.
There is, and I kind of stumbled into it. A couple of weeks into Berkeley I went to a random event, ended up talking to Dennis Zax (cofounder of Naive) for a long time, and I’m now a SWE intern there.
I almost didn’t take it. I’ve always believed that working on your own thing beats working on someone else’s, basically every time, and I still mostly believe that. What tipped it was that by then I was already unsure about Tolus, and seasoning was the exact thing I knew we were missing. Diana saying it out loud settled it.
So I want to be precise about why I’m there. It’s not for the money and it’s not because I want to be a software engineer in 4 years. It’s because Naive is a company a couple of years ahead of where we were, run by 2 people who dropped out of the same school I just started at and already sold a company before this one, and I get to sit inside it and watch how it works. That’s seasoning at a fraction of the price.
I can’t say anything about what I actually work on there, and for this post it doesn’t matter. What I can say is what surprised me, which is how informal the whole thing is. I own my project end to end. Nobody is above me on it, nobody checks in on a schedule, the etiquette is just get your shit done and tell people when it’s done. The entire company runs on Discord lmao.
And it works, it’s a Series A company with real customers and it works. It’s the same thing Hao and I had, 2 people and an idea, just with more money and a much higher bar for what counts as done.
Also, and this is the part that actually convinced me: 4 weeks in I already have a few startup ideas that came directly out of being there. One of them I’d have taken very seriously a year ago. I’d probably be building it right now.
I’m not, and the reason is the actual lesson. It’s a good idea. It’s also an idea I’d be building because it was in front of me, in a field I don’t especially want to spend the next 10 years learning. A year ago that wouldn’t have registered, because I’d have had exactly 1 idea and nothing to compare it to.
The thing we’re looking at instead is in robotics. Part of that is the market, it’s one of the few that’s obviously bigger in 10 years than it is now. The bigger part is that it’s the first time in a while I’ve been excited to learn the field I’m working in. Tolus was like that at the start. Then it turned into sending 100 cold emails a day, and the 100th cold email teaches you nothing the 10th didn’t. Being back in a space where I’m learning at high velocity every day feels great, and that feeling turns out to be a pretty good filter.
Worth noticing where all of this came from, too. A YC partner who reached out. A cofounder I met at a random event. A senior from my high school over coffee, in the last post. I wrote a whole thing in July about how the people are the product and then spent the next 2 months watching it be true.
So if the lesson is get seasoned first, the obvious follow-up is: for how long? Diana’s advice was take your time, and the uncomfortable version of that is whether there’s a point where taking your time becomes the mistake. I think there is, and the way I’ve started picturing it is two curves.
The first is seasoning. It climbs fast at first and flattens later, because the 10th thing you’ve seen teaches you less than the 2nd did.
The second is how much risk you can actually afford, and that one only goes down. At 18 there’s nothing to lose. Every year after that adds something you’d be putting on the line, rent, a relationship, a job you’d have to quit, eventually a family. So there is a clock. It’s just not on the idea. It’s on how long you can still afford to go all in once you’ve found it.
The right moment is where the two cross. Too early and you’re free but clueless. Too late and you’re wise but stuck. Here’s where we are on it:
We picked at the far left because it was the only idea we had. Everything since September is moving right on purpose. The plan is to go all in exactly once, somewhere near the crossing, and not before.
Not much after, either. That’s the part I hadn’t thought about until this fall. Seasoning flattens out and the risk budget never stops shrinking, so waiting forever is the other way to get it wrong. There’s a window, and the whole game is noticing when you’re in it.
This goes further than startups. Any big commitment made because you felt a clock has the same shape. Picking a major, picking a city, picking a cofounder. The clock is almost never where you think it is, and the thing that actually improves the pick is having seen more. So go see more, as cheaply as you can, while you can still afford to act on it.
Anyway. The next thing is probably in robotics. It’s the closest anything has come to feeling like the one, and that’s all I’m going to say until we’ve talked to the people who’d actually use it, because I’d like to not write this post twice. Taking our time, apparently. We’ll see how long that lasts.