This post will be subject to change and evolution. It represents the starting point for me 'starting up'.
Last week was my final week at Anyscale. I've enjoyed my time at Anyscale and learned so much. I'm grateful to the team for the opportunity to work with such a great collection of folks.
The Next Step in the Journey
I'm starting up from scratch. No co-founder, no team, no funding.
I've realized that I crave the challenge. Even being past the "last day" at work, the energy sits within me to at least try this challenge.
It has taken a long time to come to this conclusion, as it was a long time in the making. But come to it I have and decided that there's no better time than right now.
bias: ai-based, lifestyle sized
I have, for some time, desired to start my own company. I tried in early 2012, but found better opportunities so it didn't go anywhere.
Right now I'm seeking to start a lifestyle business.
The first milestone is to get to $1M in revenue. Preferably from a saas based business but I won't exclude some education content.
Why lifestyle business?
The timing is great for this kind of business because of (a) the internet, (b) AI.
I want to target smaller scale opportunities. They are going to allow me to iterate quickly, as an individual. The longer it takes to act (and get feedback), the slower you're going to learn and lower likelihood of success. I want to target small applications and projects so that I can learn quickly. Along the way I will document progress and share observations in the space.
That doesn't mean it must be a lifestyle business. I just need to find something that I'd like to run with for an extended period of time.
I wrote out some of my core assumptions for this experiment.
The internet creates opportunity for independent entrepreneurs to build small (but lifestyle meaningful) businesses.
The leverage of AI makes individuals more capable than before.
The capabilities of generative AI, while limited and error-prone, provide opportunity for iterative creation (in the OODA loop sense). Apps will have new value-adds that were previously cost-prohibitive.
Many companies will attempt to solve this general problem. In the market, there will be a focus on larger opportunities. That means people will eschew smaller opportunities. "Lifestyle profitable" businesses can be meaningful for an individual but the opposite for a VC. 1
This opportunity window is time-limited as the technology becomes more commonplace.
Why no co-founder?
I don't have a co-founder I want to start with. That doesn't mean I don't want one. I just need to find someone that actually aligns with the values that I have and shares the same vision. I expect that this will change over time.
Why me? Why now?
That my general capabilities and experience is a competitive advantage in shipping and building.
That I can prove this out in a matter of months and don't require a significant build time / ramp time to be effective.
If I fail, then I'll go back to growth, but this is a worthy experiment for me - even if it fails.
Small niche businesses will not attract larger players because their markets and TAM are too small, as mentioned above.
General purpose tools (e.g., chatgpt) will continue to be powerful and draw attention. If I can use a narrow tool for specific use cases - you can be competitive.
In short, not everyone is a "hacker" or "tinkerer", some people just want it done.
- That I can't code or learn to code fast enough for the domains that I want to cover (which will involve learning new things).
- This will manifest as iteration speed and inability to ship.
- Resolution: Take a step back, evaluate an educational investment (e.g., training class).
- Inability to effectively market the AI products. If they're not well marketed and I can't get people to use them, then it's pointless and I'm missing a lever. This is an area that I know less about (e.g., how to use producthunt) and so therefore it might come up.
- This will manifest if I launch but get little traffic, repeatedly. I should expect to see variance in the things that I ship and their ability to get users.
- The core risk that comes from this project is myself and my inner demons and insecurities. This would come about because I lower my standards, accept mediocrity, and lose urgency and prioritization. It's going to come from consumption of media, of information, and not setting proper boundaries.
There is some risk of poor business choice or business model choice. This will arise in several potential ways:
- A business that is too costly to maintain - in this case I'll either (a) wind it down or (b) sell the business. The emphasis will be on selling it and I'll call this out as risks for the actual projects. A particular emphasis will be put on "operational" time spent. If a project is interfering with other operations or projects (that look to have better outcomes / possibility) then packaging it up and selling it may be the best option.
- A business that can't be profitable. In this case, the evaluation will have to be on (a) is this temporary and represents investment for the future / flywheel building or (b) is this just a bad business (similar to issue 1).
Shit happens. If something happens financially (personally). Or some larger world event that causes me to need money more quickly than I would otherwise. This means that I'll want to take a particularly conservative bent on my family's finances to take this risk.
- Shipping and iteration speed trump all.
- Do cool shit, write about it.
- Showing up consistently is better than intermittently.
- Complete but lower quality is better than incomplete with higher quality.
- Day 1 profitable (or line of sight to profitability) is the approach.
- Bootstrapping and owning the business is superior to raising capital for this class of business if you can do it effectively. VC funding can represent a fundamental goal misalignment.
- I will explore raising capital if, and only if, I have a significant opportunity and traction. That I can put the cash to immediate use to solve specific growth problems.
I'm building/working on two things:
- Tech proofs-of-concept
These aim to educate myself and others on what is possible. An example: "4 ways to run stable diffusion on your own and discussion of costs and tradeoffs of each." These projects are akin to a hackathon. This is about discovery, what are the limits of these actual applications or tools?
- Commercial Applications
These aim to prove or disprove the potential for a commercial entity. The emphasis here is less on technology exploration and more on pure application and value creation / capture. They're intended to be commercially viable.
These apps will be small scale businesses - at least initially. I aim to capture immediate and existing demand (instead of seeking to 'educate' a market).
Commercial applications should be 'complete' . They can be vaporware, e.g., landing page to check for demand if the build constraints or challenges are too high. Even these should be directly asking people for money or a deposit - even if it's refunded to the customer or user in the future.
I will write about or be prepared to write about everything I build.
The emphasis here is on (a) public commitment, (b) showing up in public, and (c) documenting learnings for others.
If I'm not shipping, I'm not learning. The emphasis here will be on shipping things and "completing them". This means that it becomes a practice, that there is real urgency in doing something repeatedly - not just idly standing by.
This is a huge risk for me. At least it feels like it. With that being said, I'll learn a ton and move extremely quickly. Cheers to the unknown!
1 Nikita Bier talks quite a bit about this on twitter. Raising VC money sets you on a particular path. It's difficult (if not impossible) to 'give back' the venture capital you previously raised.