The key to "artificial intelligence" (AI) is the emancipitation of computers

To better understand this you might need to read EMAH (The Even More Astonishing Hypothesis)

Our computers' lack of agency can only be cured by an EMAH approach

Just like we limit our understanding of universe by creating road blocks such as "Big Bang" etc. (see "Origin" of Universe), we limit our computing by "stiff" programs without a true agency of their own. Only when we realize that the basis of what we call 'human thinking' rests on a nonstop flow chart without artificial stops such as e.g. input/output breaks, free will, mind, Homunculus etc, can we start developing truly intelligent* devices.

* The difference in intelligence between a brick and a human is quantitative, not qualitative (see EMAH).

It seems approximately equally hard to once and for all get rid of Homunculus as it's been to get rid of "God"

There is no Homunculus, yet researchers keep searching for him/her (also, isn't this language sex segregation stupid - in e.g. Finnish it's simply 'hän')

Imaging data suggest that conscious perception has little to do with the primary visual cortex

Masataka Watanabe (University of Tokyo): The ability to consciously perceive a visual target proved surprisingly unimportant, and shifts in target awareness had no clear or consistent effect on the activity of this subset of neurons (V1). "I was quite surprised that there was zero modulation of awareness in V1. Even in monkey studies where the [animals] showed only 10% of their neurons being modulated, [those researchers] were nevertheless observing modulation. By comparison, no such awareness effect was observed in the human subjects. Many researchers favor a model in which functions pertaining consciousness are widely spread among the whole visual system, including V1. The classical model, which assumes that the neural mechanism of consciousness is integrated into a narrow subset of brain structures, referred to as a homunculus, or 'little human', is almost defunct. However, a modern version of this model is under debate. It proposes that the neural mechanism of consciousness is a privileged set of cortical areas, a subpopulation of neurons, or certain neural dynamics (e.g. oscillations); while there are also visual systems that have nothing to do with conscious vision.

Klevius comment: The main problem here is that people don't read EMAH. The findings show that awareness ('observation' in EMAH terminology) is not a factor in visual perception. There's a confusion in the use of the word "awareness" because of the stupid dividion in observation/understanding. We can't see anything we don't understand! Read EMAH, dude!

Klevius hint to AI research: Forget about algorithms and read EMAH!

Although the intellectual work on EMAH was already done in my book Demand for Resources (Resursbegär 1992:28-33) it only got its name (the Even More Astoninshing Hypothesis) in 1994 alluding (etc) to Francis Cricks book The Astoninshing Hypothesis.