Raw LLM Responses

I had a problem and decided to use ChatGPT to see if it could come up with a **complete** solution. It's solutions were naive and I had to explicitly direct it to use an optimal solution. Iterating on the code with ChatGPT was frustrating since it kept reintroducing bugs. The final solution did 90% of what I wanted but as an _engineer_ my time would have been better spent in understanding the problem intimately and gaining knowledge of the frameworks to solve the problem. AI will struggle immensely to provide solutions that can bridge the gap between something that is almost good enough and acceptable in business terms, where the code itself is less than 20% of the work. I hope someone makes a reality show soon with all the suits trying to use AI to solve a problem or replace a development team. It should be really funny.