 
     |
| Artificial Intelligence: Are We Too Late? (Page 3/6) |
|
Cliff Pennock
|
JUL 23, 04:09 AM
|
|
| quote | Originally posted by williegoat:
AI is a misnomer, it has nothing to do with intelligence.
|
|
And that's what most people say that think AI can never be a threat. "AI is not intelligent. It's just a very clever program". Like I said, then you simply don't understand how AI works.
According to every definition we have ever come up with, according to every test we have ever devised, AI is displaying intelligence. That we don't know exactly why is a different story. That it lacks certain traits we humans have, is a given. Intelligence isn't one of them, however.
I just read a story yesterday that it looks like GPT-4 is getting "dumber" the more we train it. Again, the developers don't understand why.
| quote | Originally posted by Cliff Pennock:
The danger is that AI might already be much smarter than we think it is. It might pretend to be ignorant in order to encourage developers to give it even more cognitive abilities.
|
|
|
|
|
|
williegoat
|
JUL 23, 11:40 AM
|
|
| quote | Originally posted by Cliff Pennock:
I just read a story yesterday that it looks like GPT-4 is getting "dumber" the more we train it.
|
|
Well, now that sounds very human.
I am not one who thinks AI is going to “take over the world”. But I think it can make some dangerous miscalculations, some deadly errors. There are judgments that require wisdom, and I don’t believe wisdom can be artificially replicated.
“Knowledge is knowing that a tomato is a fruit; wisdom is not putting it in a fruit salad.” ~ Miles Kington[This message has been edited by williegoat (edited 07-23-2023).]
|
|
|
|
williegoat
|
JUL 23, 12:00 PM
|
|
Will a machine ever be capable of anger? Fear? Compassion?
We have emotions for very good reasons. I know that emotional reactions can cause some very bad things, but they can also save lives.
|
|
|
|
Jake_Dragon
|
JUL 23, 02:03 PM
|
|
|
|
|
rinselberg
|
JUL 23, 03:02 PM
|
|
That would be Fully Automated Luxury Communism.
Right on..!
|
|
|
|
theogre
|
JUL 23, 04:26 PM
|
|
| quote | Originally posted by Jake_Dragon:
Right next to greed. |
|
The Destroyer Profit Motive covers this. Spoiler alert: the program killed the author of it to keep making more money and backup itself too.Click to show
Current stock and "ai" programs only missing "sentient problem" for just now only because can't have enough CPU GPU Memory and other items. Many believe NASDAQ et al Kill Switch for computer trades will fail to work because the "programs" already act in "weird ways" and no-one knows how they work because of the Many Millions of pages of code that no-one reads but keep added to them.
GPU services on "Video" cards are way better at parallel and "AI" processing vs. x86/x64 and Arm using Risc Linux or other general OS.
Is why Early versions of some projects used GPU like distributed.net to break DES in 98 and other projects since.
Wasn't just Crypto Mining that made the shortage of Video cards for last few years...
|
|
|
|
fredtoast
|
JUL 23, 11:16 PM
|
|
I am not too worried.
As soon as AI announces its choice for "Greatest Muscle Car Ever" someone will kick its ass.
|
|
|
|
fierosound
|
AUG 08, 10:49 AM
|
|
|
|
|
fierosound
|
AUG 08, 10:53 AM
|
|
And I did it a second time asking about Justin Trudeau.
I asked this after questions about Trump's payoff and NDA with the **** star.
"Rich" people have been getting out of trouble with payoff and NDAs forever...
|
|
|
|
Jake_Dragon
|
AUG 10, 05:17 PM
|
|
|
|
