News (1) that a tiny reef fish can recognize itself in the mirror may have implications for how intelligence is defined, consciousness is understood and ultimately, how the contemporary darling of humans - "artificial intelligence," progresses out of hype and confusion. Humans, endowed with a massive quantum computer with almost infinite capacity, appear incompetent compared to life forms with limited resources. Assessment of intelligence has to consider both the output as well as the endowment - it has to be a ratio of these metrics. More tactically for humans, "IQ," has to be a function of initial conditions afforded to the individual, exhibiting intelligence. Biased standardized tests and IQ tests do not have a clue how this works.
A new measurement is sorely needed. Across the animal kingdom, if output is measured against brain endowment, humans are likely going to bring up the rear. It is apt, as they have been trying to burn the Oxygen out of their greenhouse before they suffocate, erect walls and hatred and smash a beautiful planet down, at the first opportunity. This is clearly not signs of intelligence, far from it. Recognizing self is an important leap in the intelligence spectrum. If a biological entity could do that with few brain cells, it would point to major pitfalls in our understanding of IQ. Self awareness is a necessary (but not necessarily sufficient) condition for consciousness. If we ultimately find aggregate consciousness across species is approximately the same, it will be damning for the human ego.
If consciousness can be replicated with few q-bits, as in the case of the reef fish, it may also mean that humans are on a wrong tangent toward "artificial intelligence." It would not matter if companies have infinite computing resources, that will not be sufficient to create intelligence. As the hardware, search and operating system giants have found out by repeated experiments, hooking up infinite processors and computers do not lead to "intelligence." It will certainly burn a hole in your clients' quarterly budgets but nothing beyond that. And, the ".ai" companies with stars in their eyes, doing everything to change the world, will do much less.
Let's learn from the reef fish - minimize resources and maximize output. That's the definition of intelligence.
(1) https://www.scientificamerican.com/article/a-tiny-reef-fish-can-recognize-itself-in-a-mirror/
A new measurement is sorely needed. Across the animal kingdom, if output is measured against brain endowment, humans are likely going to bring up the rear. It is apt, as they have been trying to burn the Oxygen out of their greenhouse before they suffocate, erect walls and hatred and smash a beautiful planet down, at the first opportunity. This is clearly not signs of intelligence, far from it. Recognizing self is an important leap in the intelligence spectrum. If a biological entity could do that with few brain cells, it would point to major pitfalls in our understanding of IQ. Self awareness is a necessary (but not necessarily sufficient) condition for consciousness. If we ultimately find aggregate consciousness across species is approximately the same, it will be damning for the human ego.
If consciousness can be replicated with few q-bits, as in the case of the reef fish, it may also mean that humans are on a wrong tangent toward "artificial intelligence." It would not matter if companies have infinite computing resources, that will not be sufficient to create intelligence. As the hardware, search and operating system giants have found out by repeated experiments, hooking up infinite processors and computers do not lead to "intelligence." It will certainly burn a hole in your clients' quarterly budgets but nothing beyond that. And, the ".ai" companies with stars in their eyes, doing everything to change the world, will do much less.
Let's learn from the reef fish - minimize resources and maximize output. That's the definition of intelligence.
(1) https://www.scientificamerican.com/article/a-tiny-reef-fish-can-recognize-itself-in-a-mirror/
No comments:
Post a Comment