Behold the Pigeon

[ReidG]

"You hear all the time about the wonders of AI, all the amazing things that it can do," says Ed Wasserman, Stuit Professor of Experimental Psychology in the Department of Psychological and Brain Sciences at Iowa and the study's corresponding author. "It can beat the pants off people playing chess, or at any video game, for that matter. It can beat us at all kinds of things. How does it do it? Is it smart? No, it's using the same system or an equivalent system to what the pigeon is using here."



A discussion of AI vs the humble pigeon.



https://bigthink.com/13-8/what-kind-...ligence-is-ai/


Pigeons are unlikely to get smarter, AI probably will, but still has some way to go.


And another view


"In this way our AI wonder-machines are really prediction machines whose prowess comes out of the statistics gleaned from the training sets. (While I am oversimplifying the wide range of machine learning algorithms, the gist here is correct.) This view does not diminish in any way the achievements of the AI community, but it underscores how little this kind of intelligence (if it should be called such) resembles our intelligence."









https://phys.org/news/2023-02-pigeon...elligence.html

[oska (John)]

"Auto-complete on steroids" is a very accurate way to describe ChatGPT, without waffling of course

Remember, when google released its "search appliance" way back when, the breathless called it AI too.

[xelasnave]

Why are folk so negative about AI?

Personally I look at it and most things really as something wonderful that clever humans are working upon...And this lack of respect and negative talk ticks me off...tell me why it should notvtick me off.

What type of a person is so narrow minded to attempt to denigrate this wonderful endeavour but casually placing it on the same level as a pigeon ...

It is obvious why no statues are put up for critics...what a miserable bunch who seem to only offer critisms and clearly praise is beyond them.

Maybe I take some time to appreciate the effort and work expended to get it to its current level ... you know just spend a little time thinking of what it must have taken to get this far...

Who is this pigeon guy anyways..what is his claim to fame? What has he done I wonder?

Alex

[xelasnave]

Hmmmm looking at the guy he seems like a self promoting looser seeking lime light by denigrating the work of others.

Wait till I see him in the hall.
Alex

[oska (John)]

It's the hype and alternative facts being objected to. Just to be clear many of the "critics" are the people at the forefront of the very advances being discussed, the scientists that developed much of it. And all they're saying is calm down, the hype merchants are way over the top this time, even for hype merchants.

This will probably give us better natural language search engines and simple support chat bots but further than that remains to be seen for many reasons. I don't think we should be calling these techniques AI because it's demonstrably not "AI", according to the scientists, but the facebook/twitter news cycle has already won that one.

[xelasnave]

Originally Posted by oska

It's the hype and alternative facts being objected to. Just to be clear many of the "critics" are the people at the forefront of the very advances being discussed, the scientists that developed much of it. And all they're saying is calm down, the hype merchants are way over the top this time, even for hype merchants.

This will probably give us better natural language search engines and simple support chat bots but further than that remains to be seen for many reasons. I don't think we should be calling these techniques AI because it's demonstrably not "AI", according to the scientists, but the facebook/twitter news cycle has already won that one.

Well there you have it ...the problem identified, well therefore why does not all this talk attack the hype merchants rather than the subject of their hype.

Seems very simple to me.

If these folk have any sense why can't they articulate what you have presented.

Why not say that the hype merchants have the brain of a pigeon ...

If there is a problem with the name why not change it rather than employ a derogatory attack... I still say the pigeon guy is interested in self promotion nothing more.

AND scientists are there to produce scientific papers not casual opinions.

Alex

[oska (John)]

Don't confuse the critics with the haters

I (over) simplified and tried to keep the technical terms out of it, but they were arguing the point, technically/academically, as far as I could see.

Nah, I think you're misunderstanding what's being said. The pigeon argument is about the nature/manner of machine learning in this current incarnation, an academic discussion if you like. There's absolutely nothing derogatory about it (as much as I read).

These are scientific discussions not "casual opinions" and even if they were, their expert knowledge means their opinions are hardly casual.

Maybe you're getting sucked in by the hype

The haters are getting upset about the hype or complete misunderstandings, not the reality, so most of their concerns are moot.

[xelasnave]

Originally Posted by oska

Don't confuse the critics with the haters

I (over) simplified and tried to keep the technical terms out of it, but they were arguing the point, technically/academically, as far as I could see.

Nah, I think you're misunderstanding what's being said. The pigeon argument is about the nature/manner of machine learning in this current incarnation, an academic discussion if you like. There's absolutely nothing derogatory about it (as much as I read).

These are scientific discussions not "casual opinions" and even if they were, their expert knowledge means their opinions are hardly casual.

Maybe you're getting sucked in by the hype

The haters are getting upset about the hype or complete misunderstandings, not the reality, so most of their concerns are moot.

Good...so you agree with me?

Alex

[AdamJL]

Originally Posted by ReidG

"In this way our AI wonder-machines are really prediction machines whose prowess comes out of the statistics gleaned from the training sets. (While I am oversimplifying the wide range of machine learning algorithms, the gist here is correct.) This view does not diminish in any way the achievements of the AI community, but it underscores how little this kind of intelligence (if it should be called such) resembles our intelligence."

Our intelligence is also based off predictive models. Your brain is constantly making models of the world. It's got so many feedback loops reinforcing those models and hundreds of times a day it will ignore data contrary to those models.

And let's be clear here. The AI that we are building towards is really AGI. AGI is just human level AI. We're not there yet.
But there's another capacity above that, and it's ASI. Super Intelligence. It will be so far beyond us that we probably can only guess at framing what it is and what it does.

Here's a great article about the "upcoming" AI revolution, written in 2015!
https://waitbutwhy.com/2015/01/artif...olution-1.html
https://waitbutwhy.com/2015/01/artif...olution-2.html

[oska (John)]

Originally Posted by xelasnave

Good...so you agree with me?

Alex

On the haters, sure. The reasoned opinions of scientists, not so much

Keep in mind ChatGPT passing exams is a parlour trick. The structure and nature of the training data (vast quantity of previous exams and their Q/A format) plays into the strengths of machine learning (fancy pattern matcher) implying a general knowledge that is only relevant to human learning. And this parlour trick is computationally expensive to execute and vastly more to create.

Anytime these systems have been given access to "everything" they go off the rails almost immediately. And that is kinda interesting itself given the way that parallels eerily to certain tribal groups.

[xelasnave]

Originally Posted by oska

On the haters, sure. The reasoned opinions of scientists, not so much

Keep in mind ChatGPT passing exams is a parlour trick. The structure and nature of the training data (vast quantity of previous exams and their Q/A format) plays into the strengths of machine learning (fancy pattern matcher) implying a general knowledge that is only relevant to human learning. And this parlour trick is computationally expensive to execute and vastly more to create.

Anytime these systems have been given access to "everything" they go off the rails almost immediately. And that is kinda interesting itself given the way that parallels eerily to certain tribal groups.

Why should a scientist have an opinion that wecshould think is relevant?

Opinions are belief based if they have something to say produce a paper...

My observation simply is a lot of effort has gone into this, ever what we see now is rather wonderful and sure there will be areas that can be better but I object to the general negative responce. Same with " smart" cars ... very few step up and say well aren't they clever and won't this be wonderful down the track...

I am not a grizzler and frankly I have plenty I could grizzle about but I always look for the bright side and appreciate everything... So I guess I am grizzling about grizzlers so I will shut up and go back to ignoring all negativity I find in the world...but it plain sucks.

Alex

[AdamJL]

Originally Posted by xelasnave

but it plain sucks.

that it does, Alex. It's everywhere!

[oska (John)]

Originally Posted by xelasnave

Why should a scientist have an opinion that wecshould think is relevant?

Opinions are belief based if they have something to say produce a paper...

And that's your opinion

[ReidG]

Belief is not important to a scientist even though everyone has some beliefs. What is important is what can be demonstrated to occur.


The method of science is that someone makes what they believe is a discovery or an improvement in some area and they publish their results along with enough information to others to try to replicate their results.


Some here may recall the publication of a process for cold fusion by Martin Fleischmann and Stanley Pons. https://www.nature.com/articles/d41586-019-01673-x


The short story is that they believed that they had discovered a way to get fusion at room temperature and quickly published their report. The rest of the scientific community rapidly expressed their concerns and no one was able to replicate their results which were ultimately put down to experimental technique errors.


Plenty of criticism of the claim but no one descended to personal criticism of either the researchers or those who disagreed with their claim. This is the essence of the method of science.


Just this week we have seen a modern version
https://www.abc.net.au/news/2023-02-...ares/101950476


Google was in a rush to announce a new AI and came a little unstuck when it failed to meet a simple test. Google is obviously going to improve its performance and it may very well become important but the method of science will be applied to them and anyone else and they will always be expected to prove their claims.


So let it be with automated cars and anything else, if you cannot show that you have got it right then the world should not be criticized for pointing out that the results do not match the claim.


As one of the articles in the original post points out there is no guarantee that we will ever have fully autonomous cars however much we might desire that. Many of us are old enough to possibly benefit from a fully automated car but wanting and believing do not produce such outcomes. The proof that it is possible is for the future and we all wait with interest.