Robot evolution

[Ursa major]

And I keep saying this, but once again we are trying to compare apples with pears. Actually more like comparing apples with calculators!! I think this is a huge mistake. It is wrong to keep denigrating AI if it falls even slightly short in the comparison; It's just not reasonable.

Frankly, what is being called AI falls a long way short of being AI: the way it works does not involve anything remotely resembling thinking, let alone applying intelligence.

 

[Vertigo]

Frankly, what is being called AI falls a long way short of being AI: the way it works does not involve anything remotely resembling thinking, let alone applying intelligence.

That's a whole other argument and I agree completely. AI has effectively become a marketing term. I think the industry has achieved truly incredible things but we're still a long way from any real 'intelligence'. They are really more like expert systems that know how to do one or two things, sometimes very well, but are still a long way from general intelligence. I still think it will come but I also think it isn't helpful to criticise what has been achieved because it hasn't got all the way in one single step. We should rather be celebrating all the steps that are being achieved in the long journey to wherever it will eventually get. And I, for one, put no limitations on just how far that will be. Iain M Banks' Minds are, I believe, feasible but how long (as in centuries) it will take to get there is anyone's guess!

Also the argument as to how dangerous that final goal might be is again a whole other thing. But these early achievements mean, I believe, that Pandora's box is well and truly open now and will never be closed. So figuring out how to make it work with us rather than against us seems to be the way forward.

Another thing to ask yourself is how long did it take for us to evolve intelligence? When did the apes in the trees really start becoming intelligent? Be reasonable and give it time. And celebrate what we have achieved. [edit] Oh and try to ignore the marketers!

 

[Ursa major]

I agree that we don't know how far "AI" will go in terms of real machine intelligence** but, give or take very specific and targetted examples (such as those "AIs" helping analyse the shape of various molecules in the search for new pharmaceuticals, and those looking for, say, tumours in scans), I'm with SFF writer, Charles Stross, on them: AIs are another scammy bubble (as seen with too many blockchain "currencies"***) intended to do little more than make their creators**** rich.


** - As someone who thinks that our intelligence (and consciousness, for that matter) "emerged", I can't rule out the possibility that it (they) could emerge in other animals or sufficiently complex software/systems.

*** - On the one hand, each new digital currency can add to the amount of digital currency available; on the other, each digital currency has a (an eventual) limit, so it's more like a resource than a currency... which means that the most honest thing about them (whether fake or real, scammy or not) is the use of the term "mining" to describe how new digital "coins" are produced.

**** - Obviously, those using such currencies to hide their illegal activities can, and very often do, benefit financially.

 

[Robert Zwilling]

Data is a strange thing, and takes many different forms, both dynamic and static. It can be bottled energy, a physical resource, and contrary to what some people are saying, it can also be a commodity, maybe super commodity is a better description. Sometimes when people try to define it, they usually use what it is used as to define it, which gives it many different forms. It used to be a binary alphabet but now has values between on or off. Determining when it changes from raw status to information status is a gray area.