>“Learning” is the proper term. It takes input and applies it intelligently to future tasks.
To me "learning" is loading up the memory.
"Thinking" is more like applying it intelligently, which is not exactly the same, plus it's a subsequent phase. Or at least a dependent one with latency.
>Machines can learn, machine learning has been around for decades. Learning doesn’t require biology.
Now all this sounds really straightforward.
>Machines don't learn. They encode, compress and record.
I can agree with this too, people are lucky they have more than kilobytes to work with or they'd be compressing like there's no tomorrow.
But regardless, eventually the memory fills up or the data runs out and then you have to do something with it, whether very intelligent or not.
Might as well anthropomorphize my dang self. If you know who Kelly Bundy is, she's a sitcom character representing a young student of limited intellect and academic interest. Part of the schtick was that she verbally reported the error message when her "brain is full" ;) It was just an observation, no thinking was required or implied ;)
If the closest a machine is going to come is when its memory is filled, so be it. What more can you expect anyway from a mere machine during the fundamental data input process? If that's the nearest you're going to get to organic learning, that'll have to serve as "machine learning" until more sensible nuance comes along.
Memory can surely be filled more intelligently sometimes than others, which should make a huge difference in how intelligently the data can handled afterward, plus some data is bound to be dramatically more useful than others too.
But the real intelligent stuff is supposed to be the processing done with this data in an open-ended way after all those finite elements have been stored electronically.
To me the "learning" is the data input, or "student" phase, and the intelligence is what you do with those "learnings" if it can be made smart. It can be good to build differing scenarios from the exact same data, and ideally become better at decision-making through time without having more raw data come in. Improvements like this would be the next level of learning so now you've got more than just the initial data-filling. As long as there's room in the memory, otherwise you're going to have to "forget" something first when your brain is already full :)
I just don't think things are ideal myself.
>journal on extra-terrestrial aliens would prove that the word "aliens" is used to mean "extra-terrestrials"
Exactly, it proves that the terminology exists, and gives it more meaning sometimes.
The aliens don't have to be as real as you would like, or even exist, nor the intelligence.