If I am not starting from a real use case, figuring out what to build is difficult. It is even harder when I start with technology for its own sake.
Consider a hammer. It is relatively easy to make one when I know the job it must do: break a door or drive a nail into a wall. The use gives the tool a direction.
Starting with technology sends me in another direction. I begin with the hammer’s materials, mechanics, shape, and manufacturing process. I can become very good at cutting the steel head and shaping the wooden handle, but I may lose sight of who needs the hammer and what problem it should solve. I can make components that look useful to a manufacturer while missing the manufacturer’s actual workflow.
To shape a hammer for a producer’s needs, imagination is not enough. I need to become a producer myself, or at least get close enough to experience the problem. Becoming a good producer also begins with solving problems around me.
A technical moat starts with use
The desire to become technically excellent and build infrastructure with a technical moat can remain an exciting and impressive idea unless I become a producer. If I find a problem while cutting a hammer head, I can improve that part even before I fully understand every use case. But a vague goal of mastering technology and turning it into infrastructure does not tell me which problems matter.
After studying many examples, I found that deep-tech infrastructure founders rarely began with infrastructure alone. They first sold products to consumers or other producers and encountered problems by producing something themselves. Those problems, combined with a view of the future, led them toward deeper technology and infrastructure. Even the insight that GPUs for transformers would become necessary must have come from a bottleneck discovered through actual use and production, like finding a flaw while shaping a hammer head.
What school teaches and what I must find myself
School is better at teaching the parts of the hammer in depth: materials engineering, how to shape the head, and how to build the handle. Entrepreneurs and investors can think about the hammer’s use and production pipeline, then hire people who understand the parts.
That is why it is difficult for genuinely valuable new products to emerge from school alone. I have to discover the use and define the problem myself.
Return to rapid product exploration
I tried to anchor my exploration in a technology I wanted to pursue and spend two weeks looking for its use. I wanted technology for its own sake. That approach made good ideas hard to find. I like black magic and feats of strength, but showing off the feat did not need to be the goal. There was a faster and better path.
The exploration was still valuable because I was able to verify for myself why a technology-first anchor did not work. Once I accepted that conclusion, my framework changed. I moved away from the vague belief that better AI meant I should spend six months on every difficult problem to become more productive.
If I want to work on deep tech and difficult problems, I should not start by shaping the hammer head without knowing its use. I should first decide what problem needs a hammer, then build several hammers for different uses in two-week or one-month experiments. Listening to people’s reactions and feedback matters more than defending the original technology anchor.
Think flexibly. Keep looking for a faster way.