Really felt this.
The part that hit me most is the shift from “I want to build the bot” to “I want to understand the environment the bot operates in.”
I’m going through something similar from a different angle. For the last months I’ve been building execution intelligence infrastructure around EVM systems. Not a trading bot, not a searcher setup, but infrastructure that tries to understand what is happening around liquidity, runtime behavior, ownership context, participants and state changes.
And honestly, the deeper I go, the more I feel what you described here.
The hard part is not one transaction or one strategy. It’s that everything around execution keeps moving. Liquidity changes, assumptions change, runtime context changes, incentives change. A system can look simple from the outside, but once you try to reconstruct what is actually happening, the whole environment becomes the real problem.
Your point about the expertise becoming the asset really resonated with me. At some point it’s less about “can I run the bot?” and more about “can I understand the execution environment well enough to help systems make better decisions?”
Different side of the same problem, I think.
Would be nice to connect sometime. This is exactly the kind of work and thinking I’m trying to go deeper on.