Conviction and taste

While writing the post about compressing reality, in which I argued that it was a valuable skill for the future, it opened a side quest of finding other traits that will become more or less of a requirement for startups in the age of AI.

At the root of every startup, and of business in general, is capital allocation that creates compounding results, where the goal is to create as much value as possible from a fixed amount of capital over time.

Historically limited to managers or C-level roles because of their long-term planning responsibilities, this capability is now being pushed by AI into individual contributor roles.

A single individual with AI can now easily output more than a team of three people from a few years back, and we are still at the beginning of that transition. I would not be surprised to see a single engineer be responsible for the output of a team of twenty when the UX around context switching improves.

Now, obviously using AI costs money, so you need to account for that: if that engineer’s token bill exceeds the cost of the team, that does not make sense.
I would argue that we saw this play out in the past: the cost of any utility (e.g. water, energy, internet) decreases over time as the underlying infrastructure is built and scales; Moore’s law exists in both the real and digital worlds.

Given a few years, I’m pretty sure we will have capable AI running cheaply anywhere in the world. The only question left is what to do with it.

This is where I think two traits will be important for any builder.

One is conviction: the capacity to have a vision of what the future should be, especially when there are so many things to do and most are only a few prompts away. Many people have blank page syndrome; few can write down a vision for the next ten years.

The second is taste: the capacity to build a solution that solves the problem in the simplest way possible. That does not mean it is built perfectly from the start, but the end goal should be. If a simpler solution exists, others will find it and replace you.

Obviously no one can come up with everything at the start, and you should not aim to. One might say that it is equally important to be able to course-correct and refine both your conviction and taste over time.