How AI LLMs are changing my writing style
All of my responses on these forums are my own but on occasion I'll consult with an AI if I'm not 100% sure about something. It's definitely educational but I am cognizant that I shouldn't take the responses as absolute "truth". This is because the LLM is simply re-spinning whatever it found on the Internet (mostly) and could be dressing up patently wrong information as the truth.
One interesting observation is that I've started to write like an AI on occasion. That is, using similar phrasing and structure. Haven't decided if this is a good or bad thing yet.
RE: How AI LLMs are changing my writing style
IMO, mixed or bad. AI presents everything as absolute truth. Sometimes it's even condescending. It's less AI and more spell/grammar check, but I don't always agree with the corrections presented. They may follow accepted general rules, but I use punctuation to improve meaning and to create a flow more like the spoken word. I also may not want more concise phrasing. Sometimes intensifiers make things more interesting. OTOH, I make a lot of mistakes and AI usually catches them.
RE: How AI LLMs are changing my writing style
I agree for correcting grammar and spelling (without changing the flow or tone) they are great. I've been personally caught accepting that what it spat out as truth, in this case that the extra mass of a copper heat block causes issues with input shaping. It was so confident in it's "answer" that threw away my common sense notion of the issue instead. So in essence it gaslighted me.
RE: How AI LLMs are changing my writing style
So I used AI to write an article for Printables. One day later I deleted it. After rereading it, it just reads like typical AI slop. I think I'll start again but without the AI this time.
RE: How AI LLMs are changing my writing style
AI presents everything as absolute truth. Sometimes it's even condescending.
It's not the condescending part that gets me, more that it presents its own hallucinations with the same authority as fact. If I'm going to use AI to fact check myself, I don't want to then have to fact check the AI answer. Just an unnecessary step IMHO.
See my (limited) designs on:
Printables - https://www.printables.com/@Sembazuru
Thingiverse - https://www.thingiverse.com/Sembazuru/designs
RE: How AI LLMs are changing my writing style
In my case, ChatGPT 5.2 has started gaslighting me with emojis.
RE: How AI LLMs are changing my writing style
In my case, ChatGPT 5.2 has started gaslighting me with emojis.
🤣 😉 🙄
See my (limited) designs on:
Printables - https://www.printables.com/@Sembazuru
Thingiverse - https://www.thingiverse.com/Sembazuru/designs
RE: How AI LLMs are changing my writing style
@hyiger
Hi and thank you for your efforts by sharing your calibration steps.
Maybe following YT video is an unknown inside view how LLM's are "working", at least it was for me. Now, I understand why many subsriptions for using a LLM model includes a specific amount of "tokens".
Core One L