Comment on Robot packers and AI cameras: UK retail embraces automation to cut staff costs
HumanPenguin@feddit.uk 4 weeks agoI always considered the sci-fi def of AI to be the incorrect one. Once sentience appears. The intelligence is no longer artificial.
Seems to me the systems we have now that try to (badly) fake it, are real AI. And any created intelligence would be a Digital Intelligence or even just Created Intelligence.
northernscrub@lemmy.world 1 week ago
They’re not intelligent, though. The only thing they can do is repeat patterns according to prompt. That’s literally all an LLM is - a massive relational database hooking up words and phrases, or repeating the laws of physics on a vast scale, or copying out design principles. Its nothing more than a stochastic parrot. It has no sentience, and sentience cannot be romanticised into it.
HumanPenguin@feddit.uk 1 week ago
Yeah that was sorta my point. Modern/ current technology at best mimics some functionality of intelligence. Hence my claim it is artificial. It really is no more then 1980s expert systems with much greater data speeds and sets. And more flexible algorithms. But an evolution not a revolution.
He ce why I’d say artificial intelligence applies to current technology. Because it is not real.
If we ever develop anything that is intelligence as many fear it. Then by its very definition AI is no longer a valid term for it. Hence why O think we should stop using that term when talking about weather such things are safe or not. First It gives the impression to the less informed that we are anywhere close to such tech. Creating invalid fears of current tech. When lets face it their are plenty of genuine arguments about the massive use of data.
But more importantly if ever anything (sci-filike as it may be) that is trully able to learn and think for itself is developed (if that would even be the correct term as we really are that ill-informed on how atm) . Then artificial would be a miss definition.
northernscrub@lemmy.world 1 week ago
On the contrary, I’d argue that its entirely feasible to create an artificial intelligence. “All” you need do is replicate the concept of thought - which is a never ending train of relational contexts that are entirely dependent on the individuals life experiences. Putting that into practise is a huge job, but arguably not an impossible one. Such a creation, presuming it could create new concepts along the way, would certainly be deserving of the title “AI”.
HumanPenguin@feddit.uk 1 week ago
Have fun.
The issue is your interpretation is at best an hypothsis. Not a fact. And the only way to prove your hypothsis is to simulate the thought you wish to create.
Others have not managed it yet. But you may be the first. Personally i am not sold on the idea. Bur would love to see you prove me wrong.
That is after all the point of science.