Instead of banning the agents, we better make sure that agents are acting on our behalf and under our instructions.
Most app will have agents buying and selling on behalf of. We cannot stop that.
Guardrails for agents are better than banning.
Vibe coding taught me the hard way. These principles confirmed it:
AI made it easier to create software. It didn't make production easy.
1 - Done does not mean DONE.
2 - The code gets over-engineered, and you will not know about it. Unless you have a large team reviewing the AI's work. Which defeats the whole purpose.
3 - AI reviewing the code can over-review. It chases unrealistic edge cases and kills your project.
4 - Maintainability is not native to AI. It builds code nobody can maintain later.
I built a machine around the agents that takes care of these issues, and I can safely say that vibe coding made me a better builder of solutions (not a better coder).
I built a trading algo that gave me a 29% return over the year > $500 month.
I applied the lessons I learned while building major trading solution like Murex and Calypso.
And now finalizing an open source development harness that I hope will be successful and will drive consulting business in AI. Check www.nomorehappypath.com
Most algo's, especially ones made as a 'side project', are actually just directional bets and not true algorithms. And there are cheaper ways to take directional bets thus most of these are actually losing money - if the author just understood their directional bet, they could buy relevant stock/options to have the same risk profile for cheaper. But compiling algorithms back down into their directional components is an entire field in quant finance, akin to something like DS&A for CS.
Which is the joke I was referencing. When you take a directional bet and win, that isn't proof you're algorithm 'worked', it proves your directional bet won. Akin to someone who "vibe codes a website" and then sends you a localhost url. It's proof you're just making shit up and not aware of what it is you're building or how it works.
I agree with you on this one; the compiler output remains the same if run the next day, but AI code might not. That is why you need a strong, independent reviewer who reviews the AI code and freezes it after verification, so that next time you do not have to ask the AI to regenerate the code.
reply