What I've Learned From Two Failed Startups

April 28, 2024


This won’t be a full recap of my businesses. Even if I tried, the amount of nuance I can remember of projects that took years to complete in real time is very small. Every time I fail at anything I try to find some kind of silver lining in it. It’s not always easy, but even that is an exercise in getting better at finding the positive outcomes in a negative situation. When it comes to running a business there is a lot that can go right, but there’s a lot more that can go wrong, and all wrong decisions can eventually lead to the death of your business if they are not corrected. This is my attempt at keeping a record for myself and the public of what went wrong so that in the future when I take another crack at it I will be better prepared. YMMV.

Failed Startup #1 - Torchlight Intelligence - AI Writing Assistant

My first startup was an AI Writing Assistant. This may sound quaint nowadays in the age of ChatGPT and LLMs, but back in 2019 this was almost foolhardy. I had seen the potential of GPT-2 and decided I would try to build a business off of it.

What went wrong:
  • Co-founders who cannot manage stress well make terrible co-founders, and terrible co-founders kill companies

I cannot overstate this enough. Shit will go wrong in your startup. If you go all in on a startup it will be one of the most stressful times in your life. If the person you are considering bringing on as a co-founder bulldozes conversations, shuts down, bullies others, ignores the problem, or does anything other than work with you to solve the problem under the most intense pressure of their life when the company is about to go under, move on. This person needs to be by your side (and you theirs!) through the worst of times working with you to keep the company alive. If you can’t affirmatively say “Yes” to whether or not they would do that, you haven’t thought hard enough about whether they are a good pick for a co-founder. I don’t want you to go through what I did. You can’t guarantee success if you choose someone else, but you can guarantee failure if you choose someone who doesn’t have a very good response to stress.

  • Co-founders who are not interested in going all in on the startup are better suited to be employees and make terrible co-founders, and terrible co-founders kills companies

What “all in” means changes depending on context. If all cofounders are working on it in your free time, then “all in” means dedicating that free time to building the business until it can support all the co-founders. Sometimes it can mean everyone quits their jobs from the beginning to give the startup the best chance at survival.

  • Did not talk to “my” users

I put “my” in quotations because I didn’t talk to users at all. I didn’t talk to anyone. Nobody did. This is what is known in the industry as a terrible idea.

  • Running a startup on financial fumes sucks

I need to have the freedom to purchase things to grow the business when I need them. I jumped ship from my low paying job into a moonshot idea (at the time). Thankfully I had some savings, but still I had to watch every penny like it would be my last. That kind of mindset absolutely influences business decisions. I think in the end though this could also be viewed as a positive from a certain angle of running a lean business, though I think in my case it was more like an impoverished business.

  • Have an ownership cliff clause

We never did anything with the split equity, but we absolutely should have had an earnings cliff of 1-2 years before any earnings were dolled out. Big oversight on the part of our business co-founder who, in his defense, was not familiar at all with “startups” and was more familiar with classic Wisconsin Based LLC’s (and in the housing and insurance markets, no less). To his credit though, he did put a really huge emphasis on dealing with the money and governance of things first before doing anything too in depth (how payments for the business are handled, how business decisions are made, banks/credit union stuff, etc).

  • Either advertising via Google and Facebook is not helpful or we just sucked at it

This is almost definitely a case of the latter. We probably jumped the gun on this and just used it as an opportunity to avoid talking to customers about our product.

  • Do not compromise on the name of the company. It is important.

I never really cared for the name of the company, we were just arguing over it for what felt like forever for what I felt like were incredibly stupid reasons, so I went along with it. I just wanted to build. This helped to set the tone of the co-founder relationship, which never got any better.

All this being said, there were a few things I did right the first time around.

What went right:
  • I went all in on the startup. Balls to the wall, building as fast as I could.

We got so much done in the first couple months. More than I ever thought possible, and I learned so much during that time. As far as progress on a startup you are building is concerned, there is no other way. Going full time on a startup generates more than linear progress than you would expect vs just going part time on it.

  • I got to experience what it was like to push one of my ideas to the extreme.

I will never forget the feeling of that first startup. The intensity. The vision. The scrappiness. The unhinged conviction to build something that nobody thought was possible, but we were doing it! We were really doing it! We were on the cutting edge of technology simultaneously discovering and building things that people didn’t even realize were possible! Even though I was sitting at a computer, inside my bedroom, doing nothing but writing code for 12+ hours/day 7 days/week, that was a point in my life when I felt most alive, and I will forever be grateful to my past self for quitting that awful job and taking the leap.

Failed Startup #2 - Robosymbiosis - Humanoid Robots to automate all manual labor on earth

I actually had this idea before my first startup, but I figured I wanted to get some experience running a startup first with something that would be lower capital. I also knew I was much more of a software person than hardware, so I figured I had a better shot at succeeding with a software company. After the first company failed, after I become homeless, after I got back on my feet at a high paying tech job, I figured it was time to take a crack at this humanoid robotics problem. Officially the company lasted ~2 years. Unofficially I was actively working on it for 3 years (~2021-2024).

Being the price conscious individual that I was (having just come from a funding starved company and homelessness), I knew I needed a good ratio between intelligent and affordable. Thankfully, there was a local high school robotics organization that was looking to expand into the region that was looking for mentors. I volunteered, began helping out, and used the stressful robotics season as a live filter/test of sorts to find the individual I considered to be best suited to hire. Thankfully I found one, and for the next 2 years we worked towards the titanic goal of building a humanoid robot to automate all manual labor on earth.

What went wrong:
  • I need to talk to users more (even more than the more I thought I needed to)

Like. Seriously so much more. This is an active process and a full time job, not a passive one.

  • Not having a co-founder makes almost everything a lot harder
    • Motivation
    • Finances
    • Nobody to bounce ideas off of

Due to my previous terrible experience with co-founders, I decided to go it alone this time around. This made so many more things harder. The biggest thing for me was I had nobody to bounce my ideas off of, it was all on me. I had an employee who would faithfully execute any misguided idea I had for how to accomplish my vision. And man were there many. I also had to finance everything myself. That was rough, because over the course of those 2 years I spent over $35,000 of my own money on this project. And compounding both of these, I had nobody to really give me the morale boost I needed to keep pushing at the intensity level that this company needed.

  • Building a hardware company with a distributed team is really fucking hard
  • Throwing more money at the problem does not always make the problem easier

I had money now! That means I could fix problems by throwing money at them, right? Well…no. Not at all. That just wastes money. A lot of money. Sometimes it just takes grit and an obsessive determination that no money can buy.

  • Building a business takes a lot longer when you need the day job to support it

Yes, I was working a full time job while doing all of this. I do not recommend it. My day job was funding the startup that was failing without me even knowing it. I tried dedicating more time to the startup, but I put my day job at risk because of that, as so I had to refocus towards my own day job. I would have loved to quit my day job to work full time on my startup again, but I was smarter (in one way) this time, I wasn’t going to bet everything on something that I didn’t even have a single paying customer.

  • Payroll

Quickbooks sucks. Use Gusto. That is all.

  • How shitty it feels to tell someone who’s dedicated 2 years of their working life to you that it was all for nothing

A part of the reason why I let the company linger on for so long (last couple of months) without any progress in customer acquisition is because I was dreading telling the person who had generously dedicated 2 years of their life to my dream that they were being fired because I failed. It definitely hit me harder than him, but even though we are still on fantastic terms (he’s said he’s up for any other hair brained schemes I come up with), I never want to go through that again. I hope I don’t. He was in high school so it wasn’t as bad as it could have been, but what about next time? What if it were someone else with a family or a mortgage? I’m OK with gambling with my own well being, but when other people experience the fallout of a gamble not paying off, I take serious issue with that with an intensity that is difficult to convey through text. The only victim of my failures should be me, nobody else.

  • Many small individual but consequential problems make it really difficult to decide when to call it quits

There were many problems. Not sure what the MVP is since I wasn’t talking to customers. No co-founder to bounce bad ideas off of. I could pay for the company to keep running via my day job so it effectively had indefinite runway, so there was no real pressure to do market validation. There were a ton of other reasons, but it was essentially a really expensive hobby.

  • Use a shorter company name – 3 syllables max

I got to come up with the name all on my own but man was it a mouthful. 6 syllables. Ro-bo-sym-bi-o-sis. I picked it because the .com domain name was available. I became annoyed saying my own company’s name. Needs a short name next time.

What went right:
  • Governance

What I say goes. Easy enough. Though I think this may have been achievable with a great co-founder who thinks on the same band as me.

  • Learned some fun new hardware skills

I learned how to create basic electrical connections and how to think about electricity. This was previously a mental barrier to me. Turns out I just needed the right environment to learn in.

  • Management

I learned how to manage an employee. That was a great experience, and one I don’t know how I would have gotten otherwise. Every step I made in a social sense of the relationship with my employee, I think I made the right step.

What’s Next?

There’s going to be another one. I know there is. There’s no way there’s not. I don’t know what it will look like, but there are a few things I want to be a part of it:

  • Cannot be serious

My last two companies had a “Change the world” vibe to it. I would still like to do that someday, but I think I need to solve real problems at the ground level first. Something inconsequential that makes money. I need to get the hang of making money off of something that I solve myself. I have some experience doing something like that via DoorDash when I was homeless, but it’s not quite the same as creating value wholly on your own and delivering it to people. I need that experience.

  • Must cost as little money to run as possible. No reoccurring payments unless they have been PROVEN necessary

GitHub pages for the website. Maybe Zoho. No employees until absolutely necessary. Delay losing money for as long as possible. No email list – that just offers me an “out” for not talking to customers directly about their problems (this is bigger than it seems). No custom email, just use my default for as long as possible. The only costs allowed are the domain name. That’s it.

  • Stick with Gusto. That seems to be very reliable
  • Bench is......OK. Will probably look into someone else.

I used Bench to help keep track of my finances and help filing tax stuff. The people are friendly, the data (balance sheet, trends, etc.) is accurate to the best of my knowledge, the dashboard UI is nice, but their UX kind of sucks. There’s multiple ways to contact people, either through their dashboard chat box or email, and they actively encourage you to use both. This plus the fact that there’s maybe a dozen different people I’ve spoken to over the course of working with them makes it really difficult for them to keep things straight. It’s like talking to the first version of ChatGPT. Context gets lost, there is a ton of repeating myself, and there’s a general sense of frustration in trying to keep every single person up to date with how every other person has helped me and where I’m at thus far. I expect a lot more for a service that costs me ~$7000/yr.

I don’t know if this is just a fundamentally difficult problem to solve but I’ve never used an all in one accounting and tax service that I would recommend.

  • I need a co-founder

This is undeniable. I need a co-founder. Building a product is a full time job. As is selling it. I cannot do both and work a full time job at the same time. I might be able to do both if I quit my day job, but not all 3.

I am technical by nature. I can’t just stick with that if I want to have a successful company. Next time I need to go against my nature and start with the customer. Next time I need a paying customer before I start paying for a company. Next time I have some kind of traction before I begin asking people to commit to the vision. Next time will be different.

Next time.