Back to home page

DOS ain't dead

Forum index page

Log in | Register

Back to the board
Thread view  Mix view  Order
DosWorld

25.08.2024, 19:37
 

Mini Commander - an experiment to create software with AI (Developers)

https://www.slax.org/blog/27966-Mini-Commander-an-experiment-to-create-software-with-AI-.html

Sorry, a little bit offtopic.

---
Make DOS great again!

Carthago delenda est, Ceterum censeo Carthaginem delendam esse.

Rugxulo

Homepage

Usono,
30.08.2024, 10:24

@ DosWorld

Mini Commander - an experiment to create software with AI

> https://www.slax.org/blog/27966-Mini-Commander-an-experiment-to-create-software-with-AI-.html
>
> Sorry, a little bit offtopic.

Has anyone tried to rebuild this (for DOS)?

The Makefile, in particular, is pretty lousy.

I understand that hiring software developers is not a cheap pursuit, but relying on A.I., even if desperate, seems totally unreasonable. I wouldn't waste my time with it.

glennmcc

Homepage E-mail

North Jackson, Ohio (USA),
31.08.2024, 04:17

@ DosWorld

Mini Commander - an experiment to create software with AI

> https://www.slax.org/blog/27966-Mini-Commander-an-experiment-to-create-software-with-AI-.html
>
> Sorry, a little bit offtopic.

These things now-a-days that are being called AI... are not.

They are nothing more than search engines that grab stuff from their database
based upon the 'search criteria' of the question you asked.

A year and a half ago I posed this question to ChatGPT.
________________________________________________________________________

Can you actually write computer program code? Or when asked, do you
provide an answer containing code that was actually written by humans ?

Its response was...

As an AI language model, I don't actually "write" code like a human
programmer does.

When asked to provide code, I retrieve code from a database of previously
generated code snippets that were written by humans. The code I provide
is not my own, but rather was written by humans and stored in my database.
___________________________________________________________________________

---
--
http://glennmcc.org/

bretjohn

Homepage E-mail

Rio Rancho, NM,
03.09.2024, 03:29

@ glennmcc

Mini Commander - an experiment to create software with AI

> Can you actually write computer program code? Or when asked, do you
> provide an answer containing code that was actually written by humans ?
>
> Its response was...
>
> As an AI language model, I don't actually "write" code like a human
> programmer does.
>
> When asked to provide code, I retrieve code from a database of previously
> generated code snippets that were written by humans. The code I provide
> is not my own, but rather was written by humans and stored in my database.
> ___________________________________________________________________________

You are correct. "Artificially Intelligent" computers do not actually _understand_ what is being asked. Just like in the old days when computers could do a "screen scrape" to gather some data, AI just does an "Internet scrape". Computers have gotten very good at turning spoken words into text, and vice versa, but that is all they do. AI doesn't even believe what it tells you -- it isn't capable of believing anything.

Modern scientists mistakenly believe that "the mind (and consequently 'understanding') is simply an emergent property of the neurons in your physical brain", and that "intelligence" will simply emerge from the transistors in a computer. Neither of those things are true, and they never will be. Computers (or software) will never be able to "understand".

Here's an interesting article:

https://interestingengineering.com/science/chatgpts-performance-and-accuracy-has-decreased-over-time-stanford-research

When asked to list the first several prime numbers, ChatGPT has actually gotten dumber and doesn't even know where to "scrape up" the correct answer. When you ask ChatGPT a relatively simple math question that could be calculated with a relatively simple program, it isn't able to "whip up" a program to calculate the answer. ChatGPT (and AI and computers in general) doesn't actually _understand_ what math and computers and prime numbers and programs actually are. It can speak to you in a manner that can make you think it does, but it doesn't and it never will.

glennmcc

Homepage E-mail

North Jackson, Ohio (USA),
03.09.2024, 17:27

@ bretjohn

Mini Commander - an experiment to create software with AI

> You are correct. "Artificially Intelligent" computers do not actually
> _understand_ what is being asked. Just like in the old days when computers
> could do a "screen scrape" to gather some data, AI just does an "Internet
> scrape". Computers have gotten very good at turning spoken words into
> text, and vice versa, but that is all they do. AI doesn't even believe
> what it tells you -- it isn't capable of believing anything.
>

Thank you !

You are the very first person I've spoken with who also
understands that "AI" is _not_ AI but rather _is_ total BS ;-)

---
--
http://glennmcc.org/

Back to the board
Thread view  Mix view  Order
22074 Postings in 2036 Threads, 396 registered users, 25 users online (0 registered, 25 guests)
DOS ain't dead | Admin contact
RSS Feed
powered by my little forum