exxos's Atari, Amiga & retro forum

You copy and paste code from the Internet?! On a Windows 7 machine! You love to live dangerously!

stephen_usher wrote: 19 Mar 2025 21:58 You copy and paste code from the Internet?! On a Windows 7 machine! You love to live dangerously!

Not since some several+ years back no. I was merely making people aware of the copy paste issue and did not say I was actively doing it! I don't copy or even look at anyones code on google or anywhere anymore. Just no point. In fact I barely even use Google any more..

exxos wrote: 19 Mar 2025 22:06

stephen_usher wrote: 19 Mar 2025 21:58 You copy and paste code from the Internet?! On a Windows 7 machine! You love to live dangerously!

Not since some several+ years back no. I was merely making people aware of the copy paste issue and did not say I was actively doing it! I don't copy or even look at anyones code on google or anywhere anymore. Just no point.

Could be worse... you could be asking A.I. systems to write code for you... Here be dragons. :-)

stephen_usher wrote: 19 Mar 2025 22:08 Could be worse... you could be asking A.I. systems to write code for you... Here be dragons. :-)

A.I. Has written LOTs of code for me . AI wrote my PCB BOM generation scripts last week and a test 68000 program ..viewtopic.php?p=125691#p125691

I really don't understand people's hatred towards the AI. It is a absolutely fantastic tool if you use it correctly and don't expect it to do absolutely everything for you. The amount of time it has saved me is nothing short of phenomenal.

exxos wrote: 19 Mar 2025 22:13 I really don't understand people's hatred towards the AI. It is a absolutely fantastic tool if you use it correctly and don't expect it to do absolutely everything for you. The amount of time it has saved me is nothing short of phenomenal.

It's just a pattern matching tool and depends heavily on the quality of the training data (and how much of it there is) as to whether the digital hallucinations are viable or not. In many fields, where the training data sparse the output can look reasonable but with actually be completely incorrect, and that includes a lot of programming unfortunately.

stephen_usher wrote: 19 Mar 2025 22:27 It's just a pattern matching tool and depends heavily on the quality of the training data (and how much of it there is) as to whether the digital hallucinations are viable or not. In many fields, where the training data sparse the output can look reasonable but with actually be completely incorrect, and that includes a lot of programming unfortunately.

AI isn't perfect, and its performance does hinge on the quality and quantity of training data. In areas where data is thin, like some specialized fields or even programming, it can spit out results that seem plausible but miss the mark. That said, I still find it a game-changer for efficiency. It’s like having a super-smart assistant that’s not always spot-on but saves me tons of legwork. I just double-check the tricky stuff or use it as a starting point rather than a final answer. It’s all about knowing its limits and playing to its strengths!