Why to start a start up

There are plenty of reasons to start a startup, from wanting to be rich (which we all secretly do, don’t lie to me dear reader), acquiring influence and even the noblest of us, that want to change the world. Most founders will mix and match the reasons, but for me, I knew I had to start a company because it felt like the highest leverage thing I could be doing with my time.

I’ve always been incredibly competitive, despite my best efforts to quell that part of myself, I found myself engaging in competition with friends, enemies and most importantly with myself. To start a company, to me, was the highest test if I was truly made out of the iron core that I believed myself to be.

Lee Kuan Yew built modern Singapore from the ground up, almost through sheer force of will. In the process of turning a backwater town into a modern first-world financial centre, Yew calls out what’s required of a founder:

> Whoever governs Singapore must have that iron in him. Or give it up. This is not a game of cards! This is your life and mine! - Lee Kuan Yew

In order to change the world, to have that level of impact, and to actually succeed as a founder, you need to have the iron inside of you.

The excitement of creating something from nothing and bending reality to my will was intoxicating, and I found myself staying up every night thinking about what I should be building. Building a company is a function of a founder’s ego coming into contact with the outside world,

The ideas, past year in summary

In 2023, I quit my job at Twitch, much to the chagrin of my parents but to the elation of my friends. I made all the stupid decisions of a naive first time funemployed founder: building instead of selling and talking to users.

My original idea when I quit was to solve on-call, the biggest pain point for me during my time as an engineer. Having spoken to other engineer friends, I realized this pain was massive and had the opportunity to significantly change the way we conducted engineering operations. Through late November and early December, I validated the problem with the users, but buyers were increasingly hard to convince. They viewed EngOps as a cost already accounted for in hiring, and sprint planning, meaning on-call was just part of the problem.

At the same time, I had applied and gotten accepted to YC, but was told to find myself a cofounder. I found a cofounder in a friend who had been let go from Twitch a few months earlier.

Pivot #1: In January, with our on-call idea dead in the water, we pivoted into what all our YC batch mates wanted: more sales! We figured that building an auto updating CRM filled with developers prospects was a good enough idea, and we got to approximately $25k ARR by the end of YC, selling to mostly non YC companies (those with revenue!).

Pivot #2: By May, we asked ourselves if we wanted to be on a panel in 10 years on how to sell developer products, and the resounding answer was No. Another pivot, but unclear on the direction, we spent a month searching for ideas. We explored consumer AI, meme generation and novel communication forms with AI. As we tried different ideas, we consistently found ourselves wanting an easy to use platform to build out our MVPs on top of - leading us to building a platform with nodes, akin to AWS for AI MVPs.

Pivot #3: By June, we’d explored and attempted to validate this idea, but it was increasingly clear that developers would prefer to use the underlying API. The APIs provided enough flexibility already without being too opinionated (same problem with AI frameworks - Langchain, CrewAI etc).

Pivot #4: I’d gone deep on technical topics again by July, and was just building stuff I found fun again, coming across fine-tuned embedders for retrieval systems (very similar to the recommendations work I did at Twitch) like ColPali and Chroma’s research. Super excited by creating embedding adapters, my research tests proved to actually work! We began selling, and had 3 early customers, but the vast majority of the people we spoke to said they were having bigger problems upstream, moving PDFs into a readable format.

Pivot #5: Converting complex PDFs into simple, LLM-ready data was a clear problem, but we found that while the need was strong, so were the number of companies in the same space. 4 companies from our batch also pivoted in the exact same problem space, with no differentiation between them! At this time, I focused purely on selling and we decided to verticalize into financial services (insurance, banks, lending etc). We got to $135k ARR by October, which showed demand, but I still struggled to see how we were differentiated and what the vision for the company was.

Righting the ship

Despite making six figures in revenue, I felt like we were still in the wrong space. Being in this headspace was a clear indication that we weren’t going to be successful if the founder didn’t believe in the company or direction. The second problem was the culture I’d created by allowing a lax attitude, and saying it was acceptable when deadlines were missed.

I decided to make two drastic decisions. Firstly, I decided to pivot the company. Finding a new direction, firing your customers after 6 figures in revenue is challenging, but it had to be done for the long term, building a billion dollar business. Secondly, [redacted]

Decision Making, Selling and Conviction

During the last year, the key lessons learned were:

1. Sell. Jesus fucking christ, just sell. As an engineer, I had a natural aversion to asking for money, but the reality is that you need to do this to build. This is canonical advice, but the way in which I sold also changed over the past year. This time last year, I was afraid of asking for $100 per month, ready to sign LOIs instead of just selling directly. A few months ago, I was negotiating 5 figure contracts, and I’ve learned a lot about what the difference between these two actually is.

2. There will always be hard decisions, but building a company affords you the opportunity to create the company that you want. Creating a culture of execution and having a high bar, leads to success. A company is nothing more than a group of people with a common culture, that culture which dictates everything.

3. Pivot with intention. The past year, it often felt like we had to find the next venture-scale idea within a few weeks if we wanted to continue. This forced us into a cycle of purely gradient descending into problem spaces we weren’t truly passionate about. On the other hand, starting out with a vision of the world, and iterating towards your vision is what makes you a reality-bender.