Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[silverspawn]

I think the  movie would get much funnier if everyone had little health bars floating above their head.

[silverspawn]

Current structure of my sequence on C. metaphysics:

Post #1: introduce the four Models

Post #2: discuss simplicity, the justification behind it, and formalize it as well as possible. Solomonoff Induction doesn't quite work here for complicated reasons.

Post #3: apply this to our four models. Eliminate 1 and 2 (but frame the elimination of 2 as a reinterpretation of 2 as 3).

Post #4: introduce Marr's Levels of Analysis. Use this to split Model 3 into 3.1 (implementation level), 3.2 (algorithmic level), 3.3 (computational level). Eliminate 3.3 based on various arguments, many involving GLUTs

( so now the "introduction" is over, and we can talk about the hard part, i.e., matter vs. bits, Model 3.1 vs 3.2 )

Post #5: Introduce the Binding Problem and its basic properties, including its causal effect, why you can't ignore it, and why it's not easy

Post #6: Discuss in agonizing detail all the reasons why two people talking about an image aren't going to solve the Binding Problem for vision

Post #7: Introduce regular and convolutional neural networks for those not familiar, make the point that they really don't do anything different from humans talking about images they've seen, conclude that they won't solve the binding problem either

Post #8: review algorithmic theories of consciousness from the literature -- global workspace theory, higher order theory, integrated information theory -- with a focus on how they address the Binding Problem. (spoiler: they all fail miserably.) Interestingly  however, they all agree that feed-forward neural networks don't exhibit C., but integrated information theory would claim that this changes as soon as backward connections are added. Literally a single backward connection would give it minimal C.

Post #9: discuss in agonizing detail why this latest claim, that adding backward connections changes everything, is not plausible. I think this will include all the other problems with digital consciousness, but they all relate to the binding problem so there is no sharp distinction. Ask the readers to justify why of course more complex digital systems do still solve the Binding Problem

Post #10: go through all possible answers to this question, conclude that they all don't work. finally actually call the premise into question

( now the hide and seek is over and at last the competing hypothesis must be discussed )

Post #11: Discuss how a non-algorithmic theory would work -- with C. being about the electromagnetic field -- why we can't rule it out and what its advantages are. Also discuss the U-shaped information curve

Post #12: Go through conscious vs. unconscious processing and  how this checks out with AI research

Post #13: Explain how topological segmentation can solve the Binding Problem


This is not complete but planning further won't be useful since I'm only at post #6.

[silverspawn]

should all be done in at most a few months

[silverspawn]

Moment of truth, will resident evil #5 manage to be dumber than #4?

[silverspawn]

alas, no, it's surprisingly much better.

[silverspawn]

wow that was cool

[silverspawn]

So why are Nazis wearing pictures of skulls on their uniforms?

[Awaclus]

So why are Nazis wearing pictures of skulls on their uniforms?

Because they're the baddies.

[silverspawn]

Actually

According to a writing by Reichsführer-SS Heinrich Himmler, the Totenkopf had the following meaning: The Skull is the reminder that you shall always be willing to put your self at stake for the life of the whole community.

[silverspawn]

could think of prettier ways of symbolizing that tho

[silverspawn]

I've seen a lot of movies about Nazi Germany but Schindler's List is so much better than any other one idk why they didn't just stop making them after that.

[silverspawn]

"So if a system is comprised of four binary states that are related like this..."

This is the start of any information-based theory, and it already ignores a massive problem. Who decides that something is a binary state? There are practically infinite ways to draw boxes around little clumps of carbon or silicon, and again practically infinite ways of defining two different states for one such box.

Now you could say something like "all ways are in-principle equally valid, but only the obvious interpretation is relevant because all others produce trivial systems"... or something. But no-one actually says anything like this. Instead they just take it for granted that you can take bits as elementary units. This makes it unfairly difficult to criticize since you first have to do the missing step yourself and then explain why it probably doesn't work.

[silverspawn]

Resident Evil Retrospective because I like rating things

- #1 (--): I expected a dumb action movie, but it was genuinely good. Decent thriller, only a little stupid. 7/10.
- #2 (Apocalypse): This one is just a dumb action movie, and therefore worse. Not really possible  to get involved anymore. But still quite entertaining. Something like 6/10
- #3 (Extinction): Exceedingly dumb movie. A preposterous plot. Wants to tell a post apocalyptic story, but it doesn't feel like visiting a world, it just feels like a bunch of BS. At least the BS was entertaining enough. 5/10
- #4 (Afterlife): unfathomably stupid movie even compared to #3, but it's an improvement because now it's dumb enough to be genuinely funny. This one didn't feel like a movie but rather like a sequence of cut scenes in an RPG. Enjoyable and unique watching experience for that reason, 6/10.
- #5 (Retribution): Genuinely clever directing things here and the most entertaining setting. Also best characters and twists. Plot is only about as dumb as #2 while being far more enjoyable. Weak 8/10.
- #6 (Final Chapter): Reminded me a lot of Star Wars IX. Throws away all plot developments up to that point and comes up with something totally new. This could have just been a standalone movie with minimal changes. The bigger problem is that the new thing is total shit. Like in Star Wars. Also like in star wars it involves resurrecting a dead character albeit with a better justification.  Also like in Star Wars there is never even a moment where the good guys should have even a sliver of a chance. I don't know why some movies do this. You just sit around watching the inevitable sequence of miracles. No suspension of disbelief that can be broken. The twist and the incredibly forced ending are both dumb. 3/10.

[silverspawn]

In general, nothing past the first movie means much of anything because the writing just changes the status quo on a whim.

[silverspawn]

if I remember right, then forever ago on this forum, there was this discussion about consciousness in computers -- I believe liopoil with the watch tower avatar was involved -- and I believe my hangup was that the moment C. first emerges is weird. Basically because you can divide the creation of a system into arbitrarily many steps. If your system is just a bunch of code, it's just many lines of code. You can write it line after line.

Under that picture, it's weird that the system would flip from unconscious to conscious with a single line. But if no code is not conscious and the entire thing is conscious, there must be a point where exactly that happens.

And I think the way I later reconciled that problem is by arguing that you start with infinitesimal consciousness. So maybe there is a first line where it goes from 0 to not-0, but the not-0 is just an extremely slight amount of consciousness. This is then not too weird.

Today I would say that this solution does not work because binding is ultimately a binary property. If the system does visual processing, you cannot gradually make the visual qualia more and more bound; any two qualia are either bound or not bound.

You could have a partition that gradually includes more and more qualia. So you could start by having a qualia of two pixels, then of 3 pixels and so forth. But the step where you start with qualia x and quale y that are not bound (like one half of an image & one pixel) and bind them together is indivisible.

But if you stare deep into the mouth of the problem "bind qualia x to quale y", you will see that it's impossible to do by exchanging information. If I tell you about an image I've seen, this doesn't add the image to your visual field. It doesn't matter how vividly describe it. So exchanging a bit of information doesn't do it. And neural networks don't do anything else. They just exchange a lot of information. In fact, even saying "a lot of information" is giving it too much credit. All a neural network actually does is many many instances of exchanging one floating point number at a time.

^ {another way to get at the argument}

[silverspawn]

The other thing that gets super weird when thinking about this is that neural networks and the human brain both do processing at different levels of abstraction. There are some neurons that react to local features and some that react to global, high-level features.

So if anything, we ought to have several visual fields, some higher resolution and some lower resolution.

[silverspawn]

ALRIGHT SO

[silverspawn]

LET'S SEE WHAT PROFESSOR OAK HAS TO SAY ABOUT THIS POKEDEX

[silverspawn]

[silverspawn]

[silverspawn]

[silverspawn]

[silverspawn]

[silverspawn]

I rate this payoff 0/10.

[faust]

I rate this payoff 0/10.

You need to go to GameFreak building, I believe there is someone in there who gives you a reward for completing the Pokedex. (but it's not particularly satisfying either IIRC)