digitalFAQ.com Forum

digitalFAQ.com Forum (https://www.digitalfaq.com/forum/)
-   Tech Myths, Misinformation (https://www.digitalfaq.com/forum/myths/)
-   -   Best S-VHS VCR, ChatGPT's bad response? (https://www.digitalfaq.com/forum/myths/13250-best-vhs-vcr.html)

qwertz73 01-21-2023 04:00 PM

Best S-VHS VCR, ChatGPT's bad response?
 
1 Attachment(s)
Today I tested ChatGPT, and I asked what is the best S-VHS VCR.

Here is the answer :D

Attachment 16013

https://openai.com/blog/chatgpt/

Hushpower 01-21-2023 05:33 PM

Scary stuff.

qwertz73 01-21-2023 06:32 PM

1 Attachment(s)
Second question to ChatGPT

What is the best VCR for digitizing S-VHS ?

Attachment 16014

lordsmurf 01-21-2023 08:19 PM

Yeah, I'm not overly impressed. It's taken info from somebody like me, and just regurgitates it, often mushing it up in the process. It's more like "AP" (artificial parrot) that actual AI (artificial intelligence).

ChatGPT often takes a generic approach, and uses wimp words, or weasel words. Could, may, might, etc. That's what you say when you don't really know what the hell you're talking about.

Note how it started the answer like a student that was trying to evade the question: "In the beginning, once upon a time, the history of whatever". Answer the question, don't jawbone. Don't pad the answer, add words, hem and haw.

That Toshiba deck is low-end, not S-VHS at all.

ChatGPT is already being dissected, and an IT student has created a tool to detect fake writings. It's gaining traction in education, to thwart would-be cheaters. I forget the details, saw an article on it a few days ago (CNBC, maybe?)

It's the most advanced parrot to date, but it doesn't take long to find faults in the output text. It may get better, or may not. And that advancement may take another few decades. Microsoft is "thinking about" investing in it, but investors are already pushing back on that idea.

ChatGPT is something new and shiny, so everybody is gaga over it. Reality hasn't yet set in, in terms of limitations and weaknesses, but it will.

Interesting, yes. But not overly impressive.

MrPete 01-23-2023 09:49 AM

Note too: ChatGPT makes some truly AWFUL assumptions: the information in its corpus (textual data store) is true, authentic and complete.
Thus, it is incapable of humility.
  • It will literally make up an answer out of thin air rather than say "I don't know."
  • It has no idea about the uncertainty levels of its info or answers.
Yes, it refuses to answer some questions, but that's due to programmed in limitations to hopefully avoid giving out info like how to kill etc.

Bottom line, an interesting toy but NOT something to be trusted in any way.


All times are GMT -5. The time now is 10:42 PM

Site design, images and content © 2002-2024 The Digital FAQ, www.digitalFAQ.com
Forum Software by vBulletin · Copyright © 2024 Jelsoft Enterprises Ltd.