I will explain how I use LLMs for coding and writing stuff

first of all, I do NOT support generative AI being used to replace human artists and other creatives or the big AI corpos lying and harvesting data from users, thats why I prefer open-source models usually

coding

since I was little I was very passionate about coding and computers and all kind of tech stuff in general, I have been coding since before LLMs became mainstream and I could do lets say most of the basic stuff, some sort of "jack of all trades, master of none". As many I started with web dev (self-taught + took some courses + w3schools resources) which I still do to this day and most likely will pursue a career in. then ChatGPT launched.. at first I only used it for homework like many ppl of my age (which I may or may not still do) until I realized it can actually code which was MINDBLOWING for me back then

fast-forward to now, I use AI a lot, it is truly better than me now when it comes to actually writing code quickly, BUT I do NOT just vibe code and call it a day, I always plan the project myself beforehand, write the documentation myself, write parts of the code myself and most importantly I review EVERYTHING that the AI produces and test everything end-to-end as I do not trust LLMs fully just yet

I mostly use models such as deepseek and glm and for reviewing plans/ideas just chatgpt (very generous free limits so hard to not use it)

writing

I sadly used to use AI for writing a lot, often not even going over the text before posting it or wtv, but now I write everything myself (which is very fun to do, actually) and have AI just tell me if I had any typos or parts that could be improved and I make the necessary edits after.

what I wont use AI for

• Generating images/videos which can be done by an artist
• Important and risky decisions about code
• role-play (yuck)

as always make sure to use AI responsibly and to value your privacy and not overshare stuff lol