Was fighting gsoap and Onvif for a few days, getting everything to work together like it should. Decided to use both ChatGPT and Gemini to try and help me out since documentation and examples are scarce.Probably took over 100 attempts before I got everything right. Each time one of the LLMs would suggest something, I'd try it, but it would lead to more errors.Inexperienced developers would probably continue from that 2nd, 3rd, 4th failure. They wouldn't revert the changes in between. So in the end, you'd end up with a total mess of a codebase. Things changed all over the place that have nothing to do with the original set of errors being fixed.This is why I don't usually accept PRs for my open-source projects that were written by LLMs. Often the real fix is just a few lines, but the PR contains hundreds of unrelated changes.#ChatGPT #Gemini #LLM #AI
Related
🥼💰🧊 NIH freezes funds to Harvard and four other universities, but can’t tell them#AI Q: 🚫 Should research be political?🔬...
🥼💰🧊 NIH freezes funds to Harvard and four other universities, but can’t tell them#AI Q: 🚫 Should research be political?🔬 Scientific Research | 🏛️ Federal Policy | 📜 Academic Freedo...
A short list of things that capitalists and entrepreneurs the world over have tried hard to convince us to buy and becom...
A short list of things that capitalists and entrepreneurs the world over have tried hard to convince us to buy and become dependent on, in the past, and assured us had no downside ...
📰 Leica Cine Play 1 Review: Pricey but Worth Every PennyLeica’s first home entertainment projector is pricey, but like t...
📰 Leica Cine Play 1 Review: Pricey but Worth Every PennyLeica’s first home entertainment projector is pricey, but like the company’s cameras, the image quality is worth the splurge...