So, my weekend project was trying to learn #Python with the help of an #LLM. It's convinced me further that these models just aren't ready for production usage quite yet, but they're incredible for students, if you have enough generalized experience to be able to recognize when they're talking out their asses.

Still, felt like a better-than-90% accuracy rate and the ability to customize its answers and follow basic logic and context from question to question is really quite impressive. These models have already evolved far beyond what I thought their capabilities could be, and I look forward to seeing just how finely-tuned people can get them.