I kind of disagree with looking at cogsci. I believe "organic imitation" has run its course. If functions are what's wrong then we need to fix the functions. Those tin cans are never going to be conscious, so let's focus on what's left- Performance. https://davidhsing.substack.com/p/what-the-world-needs-isnt-artificial
Thanks for the referenced article. It's a great explanation.
My response to the 'cognitive science approach' is about progress (I claim AI is possible since brains can do it). What if the idea of progressing with AI fails with the computational model, that is somewhat an arbitrary approach mimicking the digital computer? If we don't consider alternatives to computations, and we have tried computations for 70 years, perhaps there are better models?
I've learned a lot by reading the works of cognitive scientists from linguistics, neuroscience, psychology and the like. Those thinking through how to build machines that emulate humans can learn something from them IMHO.
Nope, "AI" not possible. There's no such thing as a "design without a design." An artifact that "does things by itself" is an oxymoronic concept. Not mentioning that, there's underdetermination of scientific theory leading to aphorisms such as
-Correlation does not imply causation
-The map is not the territory
-All models are wrong
etc. Making "reverse engineering the brain" a fool's errand.
There are plenty of ideas left in mathematics and engineering, no biology text needed. As I've pointed out in another article, modern superscaler processors don't take any of their innovative ideas from biology... It's just good ole fashioned ingenuity working around roadblocks, not imitation.
I kind of disagree with looking at cogsci. I believe "organic imitation" has run its course. If functions are what's wrong then we need to fix the functions. Those tin cans are never going to be conscious, so let's focus on what's left- Performance. https://davidhsing.substack.com/p/what-the-world-needs-isnt-artificial
Thanks for the referenced article. It's a great explanation.
My response to the 'cognitive science approach' is about progress (I claim AI is possible since brains can do it). What if the idea of progressing with AI fails with the computational model, that is somewhat an arbitrary approach mimicking the digital computer? If we don't consider alternatives to computations, and we have tried computations for 70 years, perhaps there are better models?
I've learned a lot by reading the works of cognitive scientists from linguistics, neuroscience, psychology and the like. Those thinking through how to build machines that emulate humans can learn something from them IMHO.
Nope, "AI" not possible. There's no such thing as a "design without a design." An artifact that "does things by itself" is an oxymoronic concept. Not mentioning that, there's underdetermination of scientific theory leading to aphorisms such as
-Correlation does not imply causation
-The map is not the territory
-All models are wrong
etc. Making "reverse engineering the brain" a fool's errand.
AI as a concept is a dead end with no way out, impossible both in principle and as a practical exercise. Explanation here if someone wants it https://towardsdatascience.com/artificial-consciousness-is-impossible-c1b2ab0bdc46
There are plenty of ideas left in mathematics and engineering, no biology text needed. As I've pointed out in another article, modern superscaler processors don't take any of their innovative ideas from biology... It's just good ole fashioned ingenuity working around roadblocks, not imitation.