Raw LLM Responses

It’s very concerning to hear all of these side effects. Personally I’m using copilot with direct questions about things I’ve already designed in my head. Where I get stuck in technical solutions. Where my knowledge ends. My biggest success with it was on a complex game mechanic I’ve been working on for 8+ years but couldn’t get the technical side working. Copilot directed me to the solution partially: teaching me how to implement a tracking system, which was where I got stuck upon the most and thought me how to write the code for the necessary checks. Which deny players attempting to progress when they haven’t unlocked something or reactivating what they’ve already activated and earned the rewards for. Another big breakthrough was for work. Where I had to find a way to build up a system that could filter and organise a large amount of unorganised data. With that I was able to direct the filtered data through a general overview. Enabling the direction to analyse general performance per employee per specific task. Seeing how efficient someone is in a specific task. Using that information to handle peak hours better. In essence I’m cooking up the solutions myself. At least what the solution should look like. Trying to find the technical solution myself first but where my knowledge ends I ask copilot to assist and teach me how and why its solution does work. It sometimes is hit or miss. Think of broken code or misunderstanding my vision. Though generally it often is able to help me through it. I document the solutions with explanations in my own words. Copilot offers the raw codes, I do the framework and design additional supportive systems for my own overview. That been said. I’m noticing that my ability to connect the dots on a technical level are weakening. While my ability to design general frameworks and envisioned systems doesn’t. Overall I think it would be important for schools to make AI a base subject. In which you learn how to cross check and fact check it, how to ask direct questions, how / when to use it and when to avoid it