I have finally started a new startup

This is the follow up to previous article.

In the previous post, I wrote about being stuck in the swamp of choosing a startup idea. This is the follow-up: after far too many prototypes, false starts and trips back to the drawing board, I have finally launched startup number two.

The answer, in the end, was not to find the perfect idea, but to find the direction I wanted to keep walking in.

What do you want to do?

This seems so obvious now, but the first question one must ask themselves is what they truly want to do? What are you interested in? Why are you planning to start a startup instead of just getting a job?

Going back to these very basic questions really helped me figure out the reason why I personally wanted to start another company:

• I want to work on technically challenging problems.
• I want to surround myself with smart, nice and hardworking people.
• I want to have the chance of an outsized reward for my efforts.

This immediately killed many of the ideas I had been toying with. For example, starting a YouTube channel reviewing AI papers went straight in the bin: I would be working alone. A portfolio of niche B2B apps that each did one thing better than an incumbent also failed the test: useful, perhaps, but not technically challenging enough.

What about a job?

This was the biggest question I had to face, why not just get a job? I could hit all 3 of the above mentioned goals in a hot new startup!

I realised two things are at play here.

The first issue was an ego I did not think I had. I was genuinely worried about telling people I might not do another startup. I had already told all my friends that "of course I am going again!". The fix to this problem happened in an interesting way, I was very down about not having yet settled on an idea and updated my CV (yes!), and I sheepishly told a friend, and they reacted as if it were the most normal thing in the world. So the realisation hit home, no one gives a crap about what I am doing; it’s just my ego.

The second is that I doubt I can get the jobs that I would want to do. Running your own company allows you to operate at the maximum level your own cognition allows. You can make your work as frontier pushing as you feel you are able, both across technical challenges and social ones. Working for someone else on the other hand will place you in a role, a pay grade, and the target becomes breaking out of the level you have been placed into. Also I do not have Oxbridge on my CV, I am a self taught engineer and have never worked at a FAANG.

The only application I sent out was to Netflix. They did not even furnish me with a reply.

Back to the drawing board

Ok, so now I know my goals, I know Netflix does not want me. Its time to build.

This part was the hardest, each idea I would start, I would hit a blocker and decide to head back to the drawing board. It felt like a game of snakes and ladders with many snakes and no ladders.

After enough false starts, I realised that all my ideas could be clustered into a single semantic node: "the future of the workplace". I was trying to figure out what the workplace will look like when we have a split between human and AI employees mingling at the water cooler.

This was a huge revelation. Yes I did not know the final product surface, but I did know what I wanted to work on.

From that realisation Viable Systems was born: an applied AI research lab focused on the future of the workplace. Built on the belief that organisation building will look very different in the next few years and I want to be part of the movement which designs it.

A lesson for the future

Whilst I am open to the possibility of needing to pivot (as is true of most startups), the one thing which I have learnt is that action beats thought. This sounds like a very strange thing to say, but the feeling of finally just getting a half-assed website and launch post out was like a weight lifted off my shoulders. I was no longer thinking in circles; I had picked a direction, made it public, and given myself something real to improve.

So if you are in the same position, do not wait until you have the whole answer. Pick the direction, publish the first version, and let reality start correcting you.