"not very smart" and "intelligence" are pretty close to opposites. I lived every day in the historical antecedents for over 50 years. To this day people, I believe most people, believe that neural networks, advanced perceptrons, believe this "not very smart" idea that they have believed my whole life. As Jaime Carbonell told me, before he died of cancer, "Bob, the old ideas from computational linguistics about semantic and syntactic structures are dead. Just simple word prediction works so well that nobody cares." Jaime was a student of Roger Shank who authored "conceptual dependency theory back in the early 1980 that struggled trying to understand brain computation with conventional computer modeling, while people like me were being discredited behind our backs by the brain-as-computer assumption. Now, a few years after Jaime died, I read papers of people discovering "the old ideas" as if they are new. They were never new, and those of us, like Jaime, knew the stupid ideas would run out of steam. But the science is damaged by people who believe that neural networks are only what they see in the product engineering by commercial folk.
"not very smart" is a lie and it is important to say that it is a lie. Specifically, "correlation" is a term of art (r) and is not what these brain units computationally do in general. Some factually mimic correlation, many don't. But they are all doing the same functional neural function which can mimic and under some circumstances look mathematically like correlation, but the brain is not doing correlation. It is doing "re-cognition".
To this minute most neuroscientists I know believe they do not know where knowledge in the brain is stored, where in fact they see it, but through their bias about intelligence just want to see where the memory system is that the CPU talks to (sic.). "Where's the memory?" They ask. We know and have known for over 40 years at least. It's distributed throughout hundreds of billions of the neural computation units between their ears that they see and sadly, underestimate and therefore don't see.
Don't believe me? Here is a paper I published in 1983 sadly asking for a re-think of brain computation to an audience who believed the perceptron lies from MIT.
https://link.springer.com/article/10.3758/BF03203564
It goes on pretty much unabated. I find the science of lies (which I call Mendaciology) incredibly interesting because of what it reveals about what neural networks really compute and why neural intelligence is so much more powerrful as intelligence than anything that the 'computerists' believe is intelligent computation. Yet they do not look at the actual science. To this day.
Do not believe me? Here is a paper I wrote from 1974 based on TWO YEARS OF HARD EMPIRICAL WORK examining some of the best direct evidence of the computations we were seeing brains computing in real time.
https://files.eric.ed.gov/fulltext/ED097656.pdf
At that time the "real neuroscience" was called "sensory psychology" which involved "single cell recording" of sensory cells in cats and modelling computation with electric circuits. Neuroscientists being well funded looking for the analogue resistors, capacitors, and inductors but failing to see what was plainly in their faces in the radical differences among the'36' or so sensory cell types in the skins of cats. And sadly not knowing what to do with the apparently missing transistors. We were seeing vastly more about brain computation in what was then called "verbal learning and memory" in humans by their verbal behavior. Not to say that the sensory psychologists were not learning good stuff. It just wasn't particularly surprising or informative.
And as I said in my hopefully constructive comment, your article is very good on some of the historical antecedents but those are not what many of us have today and have had for 50 years or more. We are now STILL stuck in a dark medieval period where knowledge is trapped in corporate, 'non-profit,' and state secrets while the public is fed and forced to feed themselves with ancient history and fiat lies.
This is the truth IMHO:
I welcome the fact that you have stated and are willing to discuss your views. Structured open dialogue will keep us alive. It is truly, by the science of lies, as well as actual human history, the only way.
Our brains, inextricably, compute that way.
We write therefore we are.
Here is Gibbon. Who was right by the computational cognitive neuroscience we know today, 250 years ago.
