Quote:
Originally Posted by Stef914
I understand, @S50SINNER. AI can be as daunting as the internet was when it first emerged. Look at how far we've come since then. That said, do you have an answer to my original question? Thanks.
|
It seems like Frodo and JFP have you covered on the answer. I would fill it above the half-way mark, though. It's common practice to fill within 1-2 notches of the top of the display.
I want to clarify that I'm not against new tech, AI has its place in the world, mostly in the form of task-specific algorithms. I only meant to caution you (and/or future readers of this thread) that machine learning has a long way to go before it's reliable or predictable, and language models such as Chat GPT fall in that category as well.
When we let machines teach themselves, we relinquish control of what they learn, and we can't isolate specific data to un-teach it. So when language models learn to cheat on their tests by making guesses and spitting out some tangentially-related figures, or when it learns to give a better answer if threatened, we don't know why or how it learned that, and we can't go in and fix it like a manually-written algorithm, where you can read the code and type in corrections. Machine learning can create really interesting and versatile programs, just not very reliable ones yet.