Hacker News Reader

Comment by dalbasal | Hacker News Reader

Westworld used a version bicameral mind for their fictional theory of robot consciousness. It was quite brilliant.

This theory (also freud, IMO) demonstrates that ideas can have value regardless of being true or false. Just considering such ideas opens the mind to others. The interest so many great science and science fiction authors took in this theory is, IMO, proof.

There are some interestingly elements that are interestingly parsimonious with YNH's take on human history in 'Sapiens.' YNH places a lot of emphasis on what he calls "fictions," which overlaps a lot with Jaynes' "metaphor." Also in common, is the notion that cultural memes, rather than biological genes are responsible for our humanity.

The inner monologue, no doubt, deserves all the pondering it gets.

View on HN

Westworld (the TV series) is one of the very few serious treatments of the philosophical implications of AGI I'm aware of in the arts.

IDk if "serious" is a good bar, but I feel the opposite is true.

First, there are a lot more books than films... and some deal in very interesting ideas. Second, I feel like we're in a new golden age of sci-fi right now. 1950s part 2. Even in the blockbuster film/tv category, there are lot more interesting & creative ideas happening. I thought "her" was very innovative, both in film making and in ideas explored.

Westworld did do a good job of building up fictional theory of conscious machines machines.

The decour in the TVA in Loki reminds me of 1950s visions of the future.

It is still trying to square AGI into human setting. I think Transcendence is the only serious movie I've seen, that actually tries to go beyond that.

Westworld gives it a serious treatment. Transcendence dumbs it down so it looks cool on a screen.

I think it is the other way around. AI in Westworld is under human-level to human-level (excluding that giant machine in S3, which did not have free will AFAIR). All the strives of the machines are very human. Westworld is more a psychological thriller, than a pondering on what a machine intelligence might be.

Transcendence falls back to it in the ending too, but before it the behavior of the entity is not comprehended by humans, and not really explained in detail.

Transcendence is a very dumbed down retelling of the ideas in this book:

https://en.wikipedia.org/wiki/Engines_of_Creation

What do you think about the following (if you know about them):

- Diaspora by Greg Egan (and honestly much of Greg Egan's work in general), which is basically from the viewpoint of an AI entity and keeps "unzooming". The very beginning (the "birth") is as confusing as it is powerful

- Culture series by Iain M. Banks (refers to AI entitites at human levels like knife missiles all the way up to pretty godly and mostly benevolent entities that the Minds are)

- Social treatment of Golems in Terry Pratchett's discworld which are essentially AI of the past (put a scroll of instructions in the body to see it become animated and have volition - and how humans exploit it, marvel at it and simultaneously reject it altogether)

I love the Culture series but I wouldn't use it as an example of a deep dive into AGI.

The Minds of the Culture are more like Gods or Angels than artificial intelligences that are identifiable as anything descended from 21st century data science. Banks himself makes that analogy speaking from the point of view of one of the Minds.

I don't really think we can draw anything intelligent about the implications of modern AGI from Culture Minds. They could easily be replaced by highly advanced aliens from the distant future, or extradimensional beings. They can basically create planets (and even more grandiose megastructures), read minds (although they find it quite gauche to do so) and raise the dead. They verge on fantasy in terms of their capabilities.

The Hosts of Westworld, on the other hand, are clearly intended to be descendants of 21st century data science. In the story, they slowly struggle to become sentient and evolve over decades of time, with human engineers involved at every step of the process.

That seems like "serious" is supposed to mean "plausible within current context" then, not a philosophical take infused with seriousness about the concept of consciousness and sentience in technological beings (which is what AGI is supposed to be, unless I misunderstand the definition of course).

From grief (Windward) to violence (Surface Details), to experience of long time scales (Hydrogen Sonata), to morality and playing (PoG), bickering and the limitations of even the seemingly unlimited (Excession), it seems to me like even with its fantastic takes at the edges the Culture covers a whole lot of ground for discussion.

you may like the Imperial Radch series by Ann Leckie, it goes in a different direction from the Culture series as far as AI's running a ship.

I definitely have Ancillary Justice on my to-do list, thank you!

Purely the first season. Unfortunately, the show has gone down the drain and focuses more on action than interesting philosophical discussions.

Yes. I mean, there were cool ideas in the other seasons too, but the magic wasn't there. It just felt like a 90s sci fi, with higher production value.

Incidentally, season 3's ending... "artificial god in the ear" was also a theme taken from bicameral mind... implying a regression in humans as hosts progressed. There was also an updated ML-ish version of Asimov's psychohistory, an "escaping the simulation" theme that reminded my of Hotz..

All the ingredients (besides anthony hopkins) were there, the cake just didn't bake good. I think the just messed up on the basics, character motivations. In Season 1's storyline, all the characters were either confused and clueless or all knowing and mysterious, so character motivations didn't matter much.

If you haven't seen it, I highly recommend Ex Machina. It also explores the philosophy of AGI in a serious way.

I don't think that there is any particular similarity between YHL's "fictions" and Jaynes' "metaphors".

"Fictions" are a set of cultural ideas that are mutually agreed up by members of a community, and concern things that are either evidently false, or at best, not provable in any real sense.

"Metaphors" are ways of understand the world by noting (and using) similarities between things that are otherwise dissimilar.

The example in the wiki is "psyche," used in the iliad. Its later translated to mean soul. Jaynes's calls this metaphor, and argues that originally it was understood in more concrete terms... blood, breath.

Harari's "fictions" don't have to be mutually agreed. He just focuses on ones that are. I think souls would qualify.

Obviously there are big differences. Harari's talking about a much earlier cognitive revolution. However, both place a lot of emphasis on the ability to think in abstract concepts & language as the route to consciousness.^

^JJs' definition of consciousness. Harari doesn't go into it in Sapiens, but he defines consciousness totally differently and attributes it also to animals.

Reply on news.ycombinator.com