News for hackers. Stuff that matters.
Re:Past Performance is Not Indicative of Future Results (2020) by swayvil on Saturday July 31, 2021 @12:28PM (#28019539)
I see no path from "observation" to "model" that does not involve an arbitrary (aesthetic? Nonrational, human-necessitating?) choice.

This would suggest that "general" AI is impossible.

ON THE OTHER HAND

There is a variety of general AI, called an "optimizer". It starts with something better than a void. Maybe that's the path we should be looking at.

Re:Past Performance is Not Indicative of Future Results (2020) by Reimersholme on Saturday July 31, 2021 @12:36PM (#28019602)
Well, human thinking relies on prior models/filters for understanding the world as well so that would invalidate us as having general intelligence too?
Re:Past Performance is Not Indicative of Future Results (2020) by username90 on Saturday July 31, 2021 @03:07PM (#28020787)
Human thinking includes building new models/filters for understanding the world, not just applying old ones. And that isn't used for learning, we do it all the time when solving any kind of challenging problem or even for simple problems like trying to recognize a face. Computer models might never compete with human performance unless they can learn how to solve a problem as it is solving it, because that is what humans do.
Re:Past Performance is Not Indicative of Future Results (2020) by swayvil on Saturday July 31, 2021 @03:58PM (#28021116)
I am on the same page.

To talk about the models some more...

There's this big mass of models. And it's got all kinds of sections. Special sections that we learn about in school. Special sections called "science". Sections that we invent ourselves. Sections that we inherit from our parents, religion, etc. It's partially biological. Partially cultural. A massive library of models, mostly inherited.

You move in relationship with the mass in different ways.

You can create new models. That's what basic science is. Extending the edge of the mass. Naming the nameless.

You can operate freely from the mass. Creating your own models or maybe operating model-less. Artists, mystics, weirdos.

You can operate completely within the mass. Never really contending with unmodelled reality. The map and territory become one. Like in a videogame. I think that's the most popular way.

Re:Past Performance is Not Indicative of Future Results (2020) by swayvil on Saturday July 31, 2021 @01:41PM (#28020144)
Those relied-upon models may be acquired nonrationally.

Via aesthetics etc.

Or, in the case of the optimizer, I think the human equivalent would be desire.