Given his influence he seems to be worth the time that it takes to try to explain to him how he is wrong?

It does seem like the only way people can maintain this confusion beyond 10 seconds of thought...

The only way to approach general intelligence may be by emulating the human algorithms. The opinion that we are capable of inventing an artificial and simple algorithm exhibiting general intelligence is not a mainstream opinion among AI and machine learning researchers. And even if one assumes that all those scientists are not nearly as smart and rational as SI folks, they seem to have much headway when it comes to real world experience about the field of AI and its difficulties.

I actually share the perception that we have no reason to suspect that we could reach a level above ours without massive and time-costly experimentation (removing our biases merely sounds easy when formulated in English).

The "I feel" opening is telling.

I think that you might be attributing too much to an expression uttered in an informal conversation.

In fact among the first improvements that could be made to the human predictive algorithm is to remove our tendency to let feelings and preferences get all muddled up with our abstract thought.

What do you mean by "feelings" and "preferences". The use of intuition seems to be universal, even within the field of mathematics. I don't see how computational bounded agents could get around "feelings" when making predictions about subjects that are only vaguely understood and defined. Framing the problem in technical terms like "predictive algorithms" doesn't change anything about the fact that making predictions about subjects that are poorly understood is error prone.