Before You Build: What Shutdowns Teach Founders

You're not meant to suffer for your startup. You will have times where you work really hard, you will have difficult moments. But suffering is not the objective. And if you find yourself suffering, it's worth stepping back and saying, do I enjoy working on this problem?

It's the week of June 2nd. Welcome to the Learning Corner, a Precursor Ventures podcast.

Welcome to the Learning Corner by Precursor. I'm Mia Farnham and I'm joined by our GP and Managing partner Charles Hudson, where we will be talking through articles and Goodreads that caught our team's attention over the past week. Our first topic is an honest and deeply reflective piece from Quinn Wang, one of Precursor's founders who shares the emotional and strategic lessons from shutting down her company, Quadrant I. Rather than focusing on market metrics, Wang writes about five patterns she noticed on her journey avoiding founder identity, chasing external approval over relying on startup playbooks, failing to realign the team as the mission involved, and deferring joy. Central to her message is the cost of ignoring intuition, intuition and the need to build in a way that honors core values. Now with her next venture, she's staying closer to the problems she knows, rejecting performative startup behavior and intentionally prioritizing joy and clarity. It's a really compelling look at what internal product market fit means for a founder and how being true to oneself might just be the real growth hack. So this post isn't just a postmortem, it's also a really internal reflection of the founder journey. So Charles, how often do you see these patterns show up in first time founders and how do you coach and encourage founders through these difficult internal struggles?

I think Quinn captured a lot of the things that I talked to founders about and about 2/3 of our founder population is doing this for the first time. And my LP's always asking me like what are some of the big differences between first time founders and second time founders? Is it recruiting? Is it fundraising? I'm like well those things matter. I think the biggest thing is this whole intuition and how to interpret advice. I find most first time founders tend to be overly deferential to more successful or more powerful people who give them advice. And that advice ends up being prescriptive even if the person delivering the advice didn't intend it to be. And I find second time founders tend to be better at taking advice's input, listening to it, weighing it, and saying 30% of this applies to my company, the other 70% does not. So we're not going to do it or we're going to take from this what's useful? I think there's a lot of first time founders who don't believe that they can or should trust their intuition unless it matches something that they're hearing from someone else. And you know, I tell founders, many of them early in our journey, think about it this way, there's a two by two matrix for these decisions you're going to make. There's was the outcome good and did you listen to my advice as a vc? If the outcome is good and listen to my advice, everybody's happy. I feel smart, we got a good outcome. If you have a good outcome and you don't listen to my advice, I'm like, wow, what a great independent thinking founder. I'm really glad you didn't take my input. If you listen to my advice and it doesn't work out reflexively, I'm probably going to tell you it wasn't advice, it was perspective or input. And if you don't listen to my advice and it doesn't work out, who cares? We're both frustrated. So the moral of the story is it's the outcome that matters, not whether you listen to my advice or kind of how you got there. Working on trusting your instincts as a first time founder is a really important skill. And I think the irony is that in the olden days when I started Inventure, there was a lot less content. So the only way to get perspective was to go talk to other people. So you kind of were forced to follow your instincts more often. And now on any topic you can find 10 or 15 really smart people arguing both sides of the same issue. What are you supposed to do as a founder? So I think that's one second. I tell people all the time like you're not meant to suffer for your startup. You will have times where you work really hard, you will have difficult moments, but suffering is not the objective. And if you find yourself suffering, it's worth stepping back and saying, one, do I enjoy working on this problem? Do I enjoy solving this on behalf of my customers? Is this something I really like doing? And second, am I trying to solve an insights problem with effort? And I think a lot of what robs people of joy is burnout. And for me, burnout happens most often when you have a problem or what you really need to do is step back and reimagine it or take a breath and instead you try to double down and kind of muscle your way through it. And that's how you, that's one of the many ways to get to burnout. And so the last thing I'll say is I wrote this post a while ago about why I don't like rebound startups and people who, you know, shut down on Friday and come back on Monday with a new idea. Having known Quinn for a while, I think some of these insights are born of perspective. And having been far enough away from Quadrani to have a clearer picture of what happened, her role in it, and why it happened. And I don't think it's easy to have these insights the day after startup shuts down because it's still too raw in most people, but not all people, but most people are reflexively defensive about what happened when it shuts down. And it's kind of hard to see your own role. And I'm speaking from personal experience, having shut down a startup myself. It wasn't until maybe a year later I was like, well, these are bad decisions I made that caused our company to have these problems. Yes, others were at fault, but like, let's be honest about my contribution.

She had a really good quote and it was. I was so focused on getting everything right that I let the company's performance dictate how I felt about myself. When things went wrong, I was discouraged. When they went well, I didn't feel proud or happy. I just felt relieved and immediately started bracing for the next shoe to drop. So in that mentality, you're never enjoying the journey and you're probably always just seeking input into her point. She just never trusted her gut. But I imagine as a founder, when you're bringing around all these awesome investors and advisors to the table, you're trying to arm yourself with as much information as you possibly can. How do you decide who to bring in and who not to?

I also think to that point, yeah, like this is somewhat related to, but it just sparked this for me. A long time ago, One of our LPs asked us like, what would happen if Precursor cease to exist? I'm like, what do you mean? He's like, what if you couldn't raise a fundamental business? I'm like, no, what do you mean? I understand the question. He's. I go, my whole identity is not this firm. I'm a husband, I'm a father, I'm a human being. There are other parts of my personality that are not being the founder, managing partner of this firm. Now, granted, it's a big part of my life and personality now, but it's not the totality. And I think one of the challenges for first time founders is sometimes being the founder of Xyz company becomes your whole personality and identity. And when you lose that, you've kind of lost yourself. And so I just told somebody this on yesterday before I saw you in New York. I started founders having a tough time and I said, well, a couple things. If this startup doesn't work out, you will be fine. You will rediscover who you are separate from being the founder. And second, never quit on a bad day. I was like, you're having a bad day. Never quit on a bad day. It's just a good rule of thumb. But I think for a lot of first time founders, being the founder and CEO of XYZ startup becomes the dominant part of their personality, which is what makes shutting down the company so traumatic.

Venture culture for a large portion in rewards hustle culture, which I think is a bit toxic in the end. I mean, you can watch Silicon Valley on HBO to figure out like they love to see the tents in the offices of people living in their place where they work all day. I think a large part of that performative. I literally heard that that is an actual thing happening like in San Francisco.

It's back.

No, it's back, but to the point of if at one point it doesn't seem like the company is going to go where needs to go. One, you need to be able to separate yourself from that eventually. And in this specific scenario, you like also need to find a place to sleep. You actually physically have to move so much.

I want to say that one thing is we have founders who just told me they were going to get a live work situation. They were going to live upstairs and work downstairs. I was like, please don't do that.

No.

And they're like, why? I'm like, you need actual separation between where you live and where you work. This isn't school. It sounds cool. It sounds convenient. It's like actually not a great idea in my opinion. And second, I've heard a bunch of these nine, nine, six stories in San Francisco. Again, you know, nine to nine, six days a week. And they're all AI companies. And I think some of it's performative, but I find when people work that way, it's usually driven out of fear.

Yep.

You know, a long time ago, one of my VC mentors was like, listen, if you have a company that's successful and he's, he's been around a bunch of companies, he's like, there were times that every one of my top performing companies where everybody was working seven days a week, but it was to keep up with business. It wasn't to do random code. It wasn't to hang out. It was because they had so much stuff to stay on top of. That the only way to do it was to work all the time. Music. And they didn't work that way forever. They worked that way until they could hire the people they needed to bring it back down to something sane. And it spiked again. And he's like, you can't do996 in perpetuity. That's something you do for a crunch or a push or for a season or for a moment. Not as a standard operating procedure. But I've heard more of this and it's all AI companies who I think, and I think it's anxiety driven, the feeling that if we're not working all the time, this technology is moving so fast, somebody's going to. Someone's going to outwork us and leapfrog us.

And I'm sure they're hearing about their competitors, which in this space is many people coming up every day. And they're probably hearing of those stories of them living in their office.

It's also the downside of having, you know, SF be the epicenter of AI is that like, you go to brunch on Saturday, everybody else is like, I can't come. I'm at the office. And they're like, well, am I doing this wrong? And I'm like, at brunch, maybe I'm doing this wrong. And you see it. It's just not theoretical. You actually see it.

I always think about this. A great manager that I love said, I don't know why people brag so much about how busy they are. Wouldn't you want to brag about how efficient you are with your time and how unbusy you are because it means you've freed yourself of unnecessary tasks? Needless to say, I don't think anyone's going to be pitching tents in Precursor's office anytime soon.

That is not our culture.

Our next topic is an article from Jason kuempler via the 404 Media blog, which discusses the challenges teachers are facing with the growing use of AI tools like ChatGPT in classrooms. Teachers are feeling overwhelmed, particularly when it comes to academic integrity, as students use AI to complete assignments. The article outlines concerns about the loss of critical thinking skills, the ability to discern misinformation, and the increasing difficulty of grading. While some educators are finding ways to incorporate AI positively, there's a growing anxiety about how AI could change the role of teachers in the classroom and the overall Educational experience. One noteworthy quote from a research chair at York University said, honestly, if we ejected all the gen AI tools into the sun, I would be quite pleased. How are we supposed to respond to this as investors who work to place capital in companies that not all the time, but ideally improve communities and the broader ecosystem overall?

There are two articles I read in the last two weeks that I enjoyed. One I think was in the New York Times, which was basically like, teachers fight back.

Yeah.

A lot of the narrative in the beginning was students are using AI to cheat or to do their homework and teachers are on the back foot. And that article in the Times was like, nope, teachers have kind of, you know, figured out some ways to sort of arm themselves too. And the reason I really enjoyed this article is, to me, a lot of the discourse on this topic has been flat. It's that students are using AI to cheat and teachers are mad because students are cheating. And I think for those who have the chance to read the article, a lot of what comes through isn't so much a fear of cheating, it's a fear of learning loss. Yeah, the teachers aren't so much like it. It's.

People have been cheating since the beginning of time.

Yeah, I was going to think writing answers on their hand, so. But I think what really came through with me is these people who are passionate about being educators are more worried about. It's that old. I mean, it sounds like a bad parenting hack, but what these students are robbing themselves of by cheating more than. And look, there were a few things that resonated with me. One person was like, how much time should I spend grading providing comments on an essay that someone just sped out of ChatGPT. I feel that way with some of the emails I get in my inbox. And like, this feels like somebody took my name and I was part of a big email list and they dumped it in outreach or dumped it in ChatGPT and sent me an email that's either creepily familiar or super vague. And I'm like, do I owe this person anything in terms of a reply? I don't know. Depends on the day. But I was just fascinated by the genuine concern that a lot of these teachers had about the implications. One that really got me was the Spanish teacher who talked about students not knowing the English word for the thing they were. And look, there's a quote. I'm on a advisory council at Stanford and every couple of meetings we read this quote about how students are not learning, they're lazy, they're terrible, and it was written 60 years ago. So this is a Kaufman dialectic between. But this does feel different because I wasn't around when calculators got introduced into math class. But I'm sure there was also a similar panic about people aren't going to know how to multiply, subtract, add, or divide and do math. This feels substantively different because the level of effort, I don't know. I just read it. I was like, I just really felt for the teachers in this article, many of whom want to create an environment for their students where they're challenged to learn. And getting an answer out of a system with very little effort robs you of the struggle of learning. And part of learning is like wrestling with things and coming to conclusion. My sense, though is it's going to be impossible to keep these tools out of the classroom. And it's probably not advisable because many of these students are going to go into jobs where the expectation will be that they have some facility with these tools. If I were a teacher, I'm like, look, I don't want to be the chatgpt police. I think this is also new. There's still work to be done. But it's funny, I just went to this session where I was hanging out with someone who's a little older than me, who's an alum of my college, and we were talking about our favorite classes. And his favorite class was the hardest class he took. And he's like, it goes my favorite class because it really stretched my mind. It really made me work hard to get to the answer. He's like, I felt like I worked the hardest and I learned the most in that class. And that's something that like stuck with me. And I. I don't know, I really identified with the struggle that the teachers in that article outlined they were. That they were going through.

What does school look like if there's no such thing as a difficult class anymore because it's just become too easy for inputs?

I think it's ironic. I think we're going to go back to less technology. Yeah. And this is happening in colleges already. Hey, your final is going to be a handwrit written despite the, the decline in student handwriting. It's going to be handwritten essays and blue books with no computer. You're going to come in and you're going to have to write without an aid. It's going to be potentially old school Socratic oral examinations. I think people are going to find a way to get back to measuring whether or not you learned and it's probably going to be different than what it's been in the past, which has been a steady march toward more technology.

I think this time around, different from the olden days when Socratic seminars, blue books, that was all they had. I think now we'll have maybe technology probably fueled AI that could support the teachers to basically manage that so they can review and grade more things, be able to kind of even out the ratio between one teacher to like 40 students. So hopefully we can find that balance also.

I just think like a ChatGPT TA that's available 247 that understands all the material, that has infinite patience. That would be great.

Yep. No, I agree. I distinctly remember being in the sixth grade and I remember the kid next to me, I think it was in math class. He cheated on a test, but he had written all the answers on the inside of a Poland Springs like water bott model. Like he was caught. But I remember the teacher being like, you have to get kudo points for this. I'm not going to fail you today. But this is obviously unacceptable. Cheating has existed since the beginning of time. I just hope that we've created some sort of guardrails to retain some sort of purpose for education. I'll ask one last question. Obviously there are arguments of the pros that AI can exist across classrooms. I'm thinking maybe a world where highly personalized learning plans can exist for each student to meet them, where their learning abilities and learning speeds are. But at what point does this, this quote unquote leveling of the learning field maybe overstep a little bit?

I think different classes and different courses have different resource demands. It seems like I mentioned the idea of an always on TA answer questions and provide support to me seems like an unambiguously useful thing. Properly programmed and properly deployed, it's a way to help people get help. I can envision a world where not everybody takes the same test, where the test you take is a function of the things that you maybe struggled with during the quarter. And it kind of skips over the things that it seems that the AI seems to believe you've largely mastered. I can envision a world where the use of AI as an input in the class, how clever were you at quizzing and asking the AI for help? And I can even envision things where it's like, hey, here's an AI output. Find the things in here that are incorrect. I think we're just so early in grappling with this. We haven't figured out how to adapt and Right now I would say the simplest explanations are winning, which is people are just using this to cheat or people are. I think music is just to cheat per se. It offers a very fast off ramp to the lazy. But look, we've had Cliff Notes notes forever. If you didn't want to read the book, you read the Cliff Notes. And do you miss out on some of the richness of reading a text? Yeah, but if your goal is like I just want to pass my test, I need to know what happens in this book.

And I always got caught in a live discussion side when I do that. Well, our last piece is Bond Cap's very deep dive. It's literally 340 slides on how AI is transforming infrastructure, applications and markets along with where they see the most value accruing. They explore a few critical themes including how AI landscape is continuing to evolve from co pilots to autonomous agents, shifts in UX of software, and reshaping of how our companies are built and monetized. As foundational models become commoditized, defensibility will hinge on proprietary data, embedded workflows and distribution, not just on raw model performance. Bond also highlights that AI native companies are blurring B2B and B2C playbooks with novel go to market and pricing strategies that reflect the growing role of agents as real time like oracles. So like validators and interpreters across regulated sectors, it's an incredibly comprehensive body of work. So Charles, first question like what trend did you find? And specifically insightful.

It was a lot. I decided I wasn't going to put it through an LLM and ask it to tell me what's exciting. I'm just going to, you know, click through it myself.

First question like what trend did you find specifically insightful?

It was a lot. I decided I wasn't going to put it through an LLM and ask it to tell me what's exciting. I'm just going to, you know, click through it myself. I felt better reading this because a lot of the things I've been struggling with in AI I felt like this highlighted the cool a couple of things. The global nature of competition in AI and I think think being in the us we talked mostly about the US companies but this really is a global phenomenon. I think the scale of dollars that are going into both the hard infrastructure and the software side are like really, really, really unprecedented in terms of the velocity. All these questions about the business model and there are a whole bunch of sections on just like business model which I read and I was like oh cool. The Things I have questions about are not answered in this presentation, but they are framed well. So I was like, okay, I feel slightly better that like the things that are giving me pause, I'm not the only one. But sometimes just like seeing some of those charts about just the amount that are being invested on the infrastructure side, they're just staggering. Also, the rate of change in adoption for some of this technology is also just staggering. And I think it just put a lot of flesh around feelings I've had that this really is a super dynamic world and things are moving so fast. On the days where I'm just like, I'm not sure what to do, I was like, oh, this is why.

Right?

This is why.

That AI adoption piece outside the us that's where I found metrics that were the most staggering. From the slides it said it took 23 years for the Internet to reach 90% non US usership. In contrast, a leading USA based LLM reached 90% non US usership in three years. So how much of this adoption is due to the product versus the underlying Rails, the Internet software? We've talked about this a little bit in previous podcasts.

I mean everyone stands on the shoulders of someone. The fact that the Internet is in many places meant that ChatGPT could be in many places. And because LLM, if you think about it before localization, internationalization was so hard and LLMs are actually pretty decent at that off the bat. So you have a product that almost out of the box, works in lots of languages, which is, I don't know, it's the thing I kept looking at was tying it back to something we talked about a few weeks ago. If the scale of outcomes for the winners in venture continues to grow and get bigger, you can actually invest hundreds of billions of dollars of private capital to create companies that are worth trillions of dollars at the turn of. We just have never seen it before. Yeah, we've seen investing hundreds of Even with like the last cycle, we saw investing tens of billions of dollars in companies like Uber to build companies that were worth tens of billions of dollars on new hundreds of billions of dollars to produce companies that are worth trillions of dollars in value. That's like a new order of magnitude step function for the industry. And we're doing it all without, I would say, high confidence that we figured out the business model and monetization piece.

Yeah, that's the key part. The key takeaway is the monetization piece has not been figured out. It looks like consumer but monetizes like SaaS like you say it all the time. AI hasn't had its adword moments yet. When will that happen? It could be a year from now. It could be 10 years from now. I don't know.

Because I feel like once people realize, okay, ads are going to be a big part of the business of the Internet. The only question was how quickly can we get advertising dollars that are being spent elsewhere to move to the Internet. And I think people forgot the Internet will allow you to just actually spend more on advertising than you ever could do. The old it's not just a shift, it's actually an expansion of the pie. I kept reading this deck hoping to figure out, well, what's going to be the inflection. They talked about, like the moment of reckoning for Amazon, the moment of reckoning for Uber. And I'm like, it doesn't feel like the moment of reckoning for open air anthropic is happening anytime soon. The only thing I can imagine is some macro event that makes capital far more expensive for those companies. I don't know what that event would be and I hope it doesn't happen because it would be bad for everybody.

Yep, I agree. Well, we are all out of time for today. Thank you, Charles, for stopping by. Please like and subscribe and we will talk with you all next week. Thank you. Bye Bye.