Thoughts On: Five months treating bugs like patients and coding agents like a medical team
Using a teaching hospital as a model for running AI Agents is a novel idea.
"The problem with coding agents is that although they’re more than capable of writing code, it’s hard to trust them to produce production-grade code that is well-tested, maintainable, and within the desired scope. Agents are relatively cheap to run, but that's little comfort when their code is going to land in customers' hands." ... "Speed matters, but quality matters much much more."
I think any experienced developer feels the same way as the above quote. It's one thing to do it fast, but if we shoot the client in the foot, it's unacceptable and a great way to lose a client.
Other ideas are great:
"No code before a plan, no plan without review."
I see people jumping into implementation without spending the time to determine how the software would be used. This is a difficult thing with the AI development. One one hand you can say this is a waste, on the other hand continuous improvement using rapid customer feedback can be a benefit as well.
From a business owner perspective, no one cares if it's written in Rust/Python/Javascript/C++ or whatever language. They only care that it runs on the platform they want (mobile, ios, android, mac, linux, windows, ...) and that it does what they expect it to do. If the code behind it is a mess, that can be okay if AI can read it and deal with it.
From someone who started development writing everything by hand this idea was always something difficult to deal with. It is a good trait to want to take pride in you work as you craft a new program. However, when you don't even look at the code anymore, where does the pride of craftsmanship come from? This is going to be an ongoing difficulty for existing developers. For new developers who are starting their development with AI, I wonder if in 10 years when they are all through college and in the workforce, if anyone will even open a code editor anymore.