Alberto,

I could not disagree with you more, and less. The basic mistake you have missed is that computational cognitive neuroscience is the "AI" that is the science of using AI concepts to model human brain computations. AI is NOT SCIENCE. AI is ENGINEERING. I learned this back in the 1970s. As such, it will try any engineering trick (like linear algebra or frankly, anything a human can invent) to learn, see, speak, etc. in any area of human enterprise -- it is an engineering subdiscipline of Computer Science.

Here is the sad consequences of that for me, and others, trying to be a scientist.

https://medium.com/liecatcher/living-a-lifelong-dream-after-a-life-of-living-others-dreams-28f09b76c5ca?source=friends_link&sk=f2805e22f9d498e0533d7996ef537eaf

The key in giving human like behavior to AI engineering is in NLP. Not in capturing emotions. These will be captured like vision, hearing, etc. Like the human brain does it. Here is the science and how AI folk are seriously undereducated in precisely what will give them the ability to give us AI that can casually, and civily, converse with any human on the planet. They have to pay attention to the science of human natural language... which, by and large, they do not and do not have any clue.

https://medium.com/liecatcher/natural-language-and-your-brain-237185770b00

Motivation is important but not the way people think. Humans, and human brain computations, compute lies as well as truths all the time in order to manipulate their human environment (who can hurt you, who to trust, who not to trust, why etc.). Language REFLECTS that inherent property of neocortical computation. IT IS NOT WHERE IT HAPPENS.

The right answer is that humanoid AI just needs NLP with certain capabilities which will establish the easy basis for humanoid machines that any person can trust (or figure out how not to).

These computations are unavoidable... and coming.

But maybe another 20-30 years if we can keep up the good work despite the engineering traps that screw up the basic science.

https://medium.com/liecatcher/lie-diseases-and-implications-for-ai-nlp-53e8bd2abce2?source=friends_link&sk=274e27f2b9275a3f3508bfcccd5a28c2

And when these electromechanical guys show up, you will be happy. You can pull their plugs (kill them) and so they will respect you. Just keep the ability to pull their plugs. It will be simple if we do.

We'll need some laws democratizing the ability to unplug the machines (any real human can do it) and requiring courts to plug the machines back in.