AI discussion (split from Canva thread)

[R C-R]

26 minutes ago, William Overington said:

Do we know when that was recorded please?

As @Alfred said, it was 8 months ago, but If you had just clicked on the link I provided to the YT video, you would have seen that mentioned just below the video in the section with the number of views, the hashtags, & so on; & would not have to ask.

[Alfred]

20 minutes ago, Alfred said:

It was eight months ago.

Psychologist and cognitive scientist Gary Marcus made some interesting observations about the interview immediately afterwards.

Another article by Gary Marcus also makes for interesting reading.

[William Overington]

28 minutes ago, R C-R said:

As @Alfred said, it was 8 months ago, but If you had just clicked on the link I provided to the YT video, you would have seen that mentioned just below the video in the section with the number of views, the hashtags, & so on; & would not have to ask.

Ah, but that date is when the video was added to YouTube not necessarily when it was produced.

For example, I remember this programme when it was new in the 1990s, yet the video is listed as added to YouTube three years ago.

So, here as an example, yet please enjoy as a serendipitous experience. Later in the video there is some wonderful singing from Edith Prock.

William

 

[R C-R]

41 minutes ago, Alfred said:

Psychologist and cognitive scientist Gary Marcus made some interesting observations about the interview immediately afterwards.

Also worth a read, if for no other reason than noting where they agree vs. disagree about what AI is, how it works, & so on.

EDIT: It is also worth considering that (like the name implies) 60 Minutes is a one hour long US TV show, that with commercial breaks there is only about 45 or so minutes available for program content, & the show covers several different topics, so the 13 minute interview could not begin to cover all that Marcus wanted it to include.

[Alfred]

13 minutes ago, William Overington said:

Ah, but that date is when the video was added to YouTube not necessarily when it was produced.

A fair point, but if you follow the other links that I subsequently provided you’ll see that they also date back to October last year.

[R C-R]

13 minutes ago, William Overington said:

Ah, but that date is when the video was added to YouTube not necessarily when it was produced.

But obviously I hope, the interview was recorded at least 8 months ago.

[William Overington]

2 minutes ago, R C-R said:

But obviously I hope, the interview was recorded at least 8 months ago.

Well, yes. 😁

William

 

[SrPx]

What I extract is that Gary Marcus (psychologist and cognitive scientist) strongly disagrees on the capability of AI to end up with a conscience in the sense that we have (well, science have not been able yet to fully fine tune the concept of what the human conscience is, a bit presumptuous to think we can "build" one). Also, some consider that Geoffrey Hinton is not the 'godfather' of AI, or at least it is a shared crown (Turing and Yann LeCun (who, btw, disagrees with Hinton about conscience, etc)). And these two (Hinton and Marcus) have been disagreeing on the emotions and awareness possibilities in AI for 30 years. If Geoffrey is after a particular agenda, is not something I could find material about. But what is clear to me is that they both agree on one fact, which was what I first heard from Hinton some time ago, and with which I also totally agree: That this Pandora box is more dangerous than we think it is, in the sense that we are not a civilization that can act fast enough controlling well the bad (human) actors handling dangerous and nasty stuff. It is specially important when a tech is this level of disruptive (I'm kind of "old" and I don't remember something similar in my lifetime). Indeed, these actors tend to be much faster than our good policies (one of the reasons why I get beyond shocked hearing people denying the need of very strong regulation, ASAP). It's clear to me that he is absolutely right about the risks (and that he truly saw those risks coming and so left Google due to it), not so much in the capabilities he attributes AI to end up feeling emotions or having an actual conscience, even less when he is attributing some of that to current  "AI". But about the risks, IMO, he's right on the money.