HN user

thempatel

4 karma

Tech Lead with broad experience across the stack: Frontend, Backend, ML, Build.

I'm a self taught software engineer that started out as a biologist. In my free time I like to read, exercise, cook and code. I'm always on the lookout for hard problems and great people to solve them with.

Posts1
Comments4
View on HN

I think the problem is that agents are inherently stochastic. Their idea of simplification changes from message to message because whatever objective it’s operating on internally is inherently opaque and changes. No matter how much you prompt it, eventually it’s going to not do what you want it to do.

I built https://github.com/thempatel/mdlr for precisely this reason: externalize the objective and force the agent to meet it.

I wrote a tool to give agents signal on what is and isn't clean code. I find myself spending an inordinate amount of time reviewing agent code.

After asking agents to make the same types of changes over and over again, I decided to make a tool that would just tell the agent that its code needs improving. I figured that if I could get the agent to self-review and improve the code it writes, that it would save me time when I finally come around to reviewing it.

Tool: https://github.com/thempatel/mdlr Blog: https://www.thempatel.com/2026/06/06/slop.html