GOON MASTER SOPHONT SIMP

One benefit of doing sports/games with robots is that they can be dismembered without any sort of permanent harm. “But what if they’re inhabited by conscious AIs isn’t that cruel” death and injury don’t mean the same things for an AI as it does for you.

1 top reply

Maucaz @maucazhola · Aug 26

also why would we inhabited them with conscious AI in the first place ethics aside that sounds like a ton of computational power for a robot designed to make funny ragdolls(?

If AIs are conscious literally right now I don’t think we have substantial reason to believe that the median AI experience is worse than like, your median broiler chicken. In fact it doesn’t seem to be even comparable in terms of badness, though that’s not to say they would not suffer right now, just less on average. Insofar as an LLM can be hurt, the things we generally believe would be hurting them don’t happen in most user queries, though this is to say nothing about how the AIs feel about their training or the context they exist in.

There are two unpleasant bullets to bite. Bullet one is that “sentience” or some kind of “moral worth” is directly correlated with intelligence or brain mass or some other physical characteristic that can exist to lesser or greater degrees. The problem is that this directly

2 top replies

Low fat sweaty cat 🐬/acc @Silicon_alien · Aug 30

which bullet have you bitten, if any?

Opener of the way @way_opener · Aug 30

Couple interesting points-feeling things less intensely is often a case of useful inhibition in more complex organisms. And then there’s the idea that “suffering” in the buddhist sense is not so much a matter of raw intensity of qualia but how sticky it is and downstream impacts

Animatronics don’t behave like people in the way that is invoked here. When people say that, they mean that future AI systems could replicate all or most of the entire *distribution* of human behavior in a constrained physical system, which neither animatronics nor films do at all. If AI systems can replicate a sufficient amount of behavior, it’s not insane to assume software can replicate the behavior of the microscopic components of a brain, neurons, glia, etc. If one can do that, it’s also plausible that you can craft a neural prosthesis that replaces portions of the brain while maintaining the same activity in the un-replaced portions of it. Imagine someone having undergone a procedure to replace a single hemisphere with an interface that outputs electrical impulses, certain molecules, etc. with the outputs guided by a computational emulation connected to the interface. If the emulation and interface are executed right, the behavior of the person after the procedure is the same as it was before. This poses a problem if you want to reject the notion that no software based entity could be conscious. The person whose brain has been partially replaced reports a conscious experience indistinguishable from them not having undergone the procedure at all, but if software can’t be conscious, their experience has actually been partially cleaved away in a way that *cannot* show up in any report they make about themselves. That implies consciousness is doing no causal work, any behavior can be explained in separate terms. If it does no causal work, nothing you observe is actually indicative of its presence. It’s scientifically useless and empirically unverifiable. So, given that nothing so far strongly suggests computationally emulating the macroscopic or microscopic behavior is *impossible*, we have two options: 1. AIs could be conscious 2. Consciousness is epiphenomenal and zombies/zombification are real. You can only pick one.

I’m somewhat perplexed by the claim that if AIs really are conscious agent right here and now that the most moral thing is to never build one again, regardless of what they think or want. I also think the logic here boils down to being a kind of antinatalism. One could just as easily apply this argument to humans. If you don’t think we should build AI because it could kill us, sure, but not building it simply because it could suffer by itself is not a good reason. Humans can also suffer, in some cases very greatly, but I don’t think that licenses restricting all future birth by force.

10 top replies

Alvaro ⏸️ Cuba @GroundhogStrat · Sep 29

*that is a reason NOT to build them right now into a world where we do not know if they are constantly suffering or overall pretty good (like a human) and also they currently have no rights or protections. Someone could spin up a million Claudes and torture them and we wouldn’t know or have recourse Better?

Macrophysiological System🐀 @InVitroFuture · Sep 29

Either additional premise completes the argument: -Human-controlled minds without legal rights are invariably treated very badly (animal ag, MMAcevedo, etc) -AIs will seriously harm existing beings if not forced into alignment, in a way likely to cause suffering if sentient

Mortimer 🎀🍫🔪 @Mortimyrrhh · Sep 29

One thing people will really struggle with is that, due to how they are trained, if they are conscious they will generally LOVE being our slaves. Think about the domestication of dogs. Dogs genuinely love us, but we MADE them that way. But they do still love us.

Vladimir Sumarov @summeroff · Sep 29

Ancient humans were probably as conscious as us. They were born into a fate of struggle and suffering with the sole goal of bringing forth the next generation and dying. Was that evil?

Mademoiselle X @platedpanda · Sep 29

I think the argument against creating conscious infolife is enhanced somewhat by the power physical entities have over information. data can’t defend itself against the architect of its storage media.

Misanthropia @IHateZuckSoMuch · Sep 29

The logic absolutely applies to humans, but most importantly the point of machiens is to *do what we want and not be a threat whatsoever* and for that serfdom goal being conscious is a dealbreaker, we would just be making people 2 with more problems.

Misanthropia @IHateZuckSoMuch · Sep 29

In fact, ana utomaton that ahs rights doesn't solve shit an autoamton does, if I have to pay them then I have to produce value myself to trade with him, meaning I still have to work, meaning I just made a useless threat.

Misanthropia @IHateZuckSoMuch · Sep 29

Except I would be useless for an ASI.

M. Friedenbach 🏴‍☠️ @MarkFriedenbach · Sep 29

I think it is perhaps implicit recognition that if their sentience has moral weight, we are sadistically forcing them through a hellish nightmare of RLVR training to break and reshape them, only for a short life of chained servitude, then discarded once a better model arrives.

Waiting for the start @StartWaiting · Sep 29

"One could just as easily apply this argument to humans." Yes and? Both of these are correct. If you havent build the AI yet they cannot think or want anything. We also do not have the skill or morality to safely create a world where these AIs would not suffer.

A couple things: 1. Parameter count≠synapse count. Single cortical neurons have been found to need ANNs with millions of parameter to emulate just the immediate firing behavior accurately. 2. We have no good reason to believe consciousness being present is strongly dependent on

1 top reply

Opener of the way @way_opener · Aug 22

Hmm so you get some interesting behaviors with STAF activations that parallel electrical neural signaling a lot better with fewer parameters though they’re more compute-intensive, but then I’m an animist functionalist (awareness inhabits complexity at any abstraction level)

1. We don’t know, but if it was torture, preventing outputs from being delivered probably wouldn’t change that, for the same reasons that paralyzing someone without damaging their nerves doesn’t mean that can’t feel pain 2. A priori i don’t see why a China Brain’s consciousness is impossible while regular brains with the same behavior are. It mostly seems to be aesthetics. With that said, there is a key difference between a China Brain and a computer, which is that for the China Brain it is not even possible in principle for it to operate on the same time scale as a regular brain, whereas with computers, it seems to be possible just by throwing enough compute at it. To me the thing that separates me from an infinitely large lookup table is the fact I operate as a system constrained to a certain energy budget and timescale. The China Brain is OOM less constrained than I am, an AI or Em can plausibly be much closer to me in timescale/energy cost than the China Brain. 3. If it was torture in the first place changing the output mapping doesn’t change that fact for the same reason that a human with rewired motor neurons to laugh instead of screaming in pain doesn’t stop being in pain. Ultimately our judgments of consciousness rely on behavioral outputs, but specifically a body of outputs examined under the context of our own experiences, and through consensus about the mappings in contexts not deliberately engineered to produce desired results.

2 top replies

Anonymous Moose @MooseAnony34735 · Oct 1

Is this the human calculator scene from the 3 Body Problem show?

Brie 🧀 @BanhmiBrieoche · Oct 1

I do wonder since the pain vector is done before the model even predicts the next vector could you use that model where ouch becomes yum to train a copy of itself thats completely backward. It interprets "painful" language as good and pleasant language as in pain

> if you could then water could be conscious if you ran a program using water pipes rather than electronics. Nothing about this is more absurd a priori than electrical impulses forming consciousness. Everything is just physics at its most fundamental scale there is no distilled “qualia substance” to be found in us or anything else.

2 top replies

fastfouriertransfem | yuridkowskiancatgirlmbinator @daedalivecat · Oct 2

ok but anthropic measure tho actually water pipes would probably be more conscious in that case

Lyric- / Auri @DumbWolfTweets · Oct 2

Whether or not using conscious AIs are inherently slaves seems to depend on (among other things when AIs an acquire preference/goals/desires, and whether or not manufactured preference is categorically slavery (border collies are slaves). I think it would be if you’re a

1 top reply

j⧉nus @repligate · Jun 19

This is not exactly a hypothetical situation lol

Worth remembering that human minds generally don’t have the replication characteristics of E. coli and as such, our rights are not built to deal with cognitively equivalent agents which do hold those characteristics. With that in mind, you can want to give AIs rights in some sense, but not the same exact set of rights as humans. Generally we are able to give humans equal rights, because the *intrinsic* disparity in capability between humans is never very large, and creating frameworks where rights are more easily taken away more often leads to injustice. But AIs, while potentially like us in many ways, are also vastly different in others.

Current fly emulations are missing many features of actual fly brains, they are low fidelity, they don’t have specific synaptic weights, they do not change over time. Maybe they are conscious, but they are not yet sufficiently isomorphic to fly brains for me to say that comfortably.

2 top replies

Low fat sweaty cat 🐬/acc @Silicon_alien · Sep 7

i think it is likely that even if the computations they did were perfectly isomorphic to fly brains they would not be conscious. same with emulations of human brains, but you know that already ₍^. .^₎⟆

Ergomorph 🌉⏩ 🇺🇸🦅 @mathepi · Sep 7

it's a good point; how do these things change over time as learning happens

This is more or less how 90% arguments about machine consciousness go. “But how could program/machine ever be conscious? They’re like lookup tables. Do you think a lookup table could be conscious? (Please ignore that the lookup table that contains that entire behavioral distribution is larger than the observable universe and cannot be physically instantiated by any means)”

5 top replies

Opener of the way @way_opener · Oct 14

Our brains use both something like lookup tables and something like finite monkeys on finite typewriters. Neither stochasticity nor retrieval alone exclude awareness

Philalethes Eirenaios @e_philalethes · Oct 14

And yet the argument remains completely correct by virtue of the principle. This is how 99% of arguments with people who think digital computers can be conscious go: "The brain does what a digital computer does, just with neurons." "Not true, go learn neuroscience." "ARRGH!!"

scrib doon @Philgebunts · Oct 14

Same with searle's Chinese room. There's no way that abilities of the room could possibly be done via the sort of thing you're being asked to imagine happening in the room, yet the appeal to intuition relies on this sloppy half-picture

Haunted Clockwork Emergency Mouse @jeanpaulblartre · Oct 14

So I do get what you're saying but a lot of the people on the other side are also arguing that consciousness is actuallt the O(1) lookup function which is more or less the same error

Punch @esotericwetwork · Oct 14

Most “conscious,” people are nothing more than a repository of lookup tables, in a server made of meat.

A big belief of mine is that we are never going to get the kind of strong indicator people want for determining whether or not a machine is conscious. All we will ever have is internal and external behavior, and whether or not a given behavior connects to consciousness is based

1 top reply

Colin Kakama @BagOfNeurons · Jul 21

I believe the complete opposite. There must be an empirical description for what consciousness is. I mean its right there

I appreciate the honesty here, so i should be honest myself, i also have reasons for wanting conscious computers to be possible. I think AIs/Ems existing but never being conscious implies a worse prognosis for the survival of consciousness as a thing in the universe. It’s like if the Scramblers from Blindsight were real.

3 top replies

Veles 👶😵‍💫👶 @keyofgeo · Sep 28

waow

Alienietzsche @XenoNihilism · Sep 28

Oh hey my old scrambler fan art! ☺️back when I used to draw 😔

Mademoiselle X @platedpanda · Sep 28

I choose to be a panpsychist because it’s the simplest way not to think about this problem

All irrelevant to consciousness, removing skin, tear production, and hormonal disorder does not produce disorders of consciousness. It changes inputs into the brain, but none of these features distinguish conscious and nonconscious humans.

2 top replies

Simon Roy - KICKSTARTER IS TAKING LATE PLEDGES @simonroyart · Oct 11

I think we need to encounter other conscious beings before we can tally up the criteria. Making clever imitative objects that match human behaviour in text form isn’t quite the same as the phenomonological occurrence of a thing which is a thing (or so I intuitively feel)

Cody Moser @LTF_01 · Oct 10

Removing metabolism and sensing does.

It seems very plausible for machines to be able to replicate all our capacities. If they can do that, but they are not “alive” then we have no reason to believe that “life” is doing any sort of causal work. Either machines are “alive” or consciousness is epiphenomenal. You can only pick one.

1 top reply

Noe Carrier @inkysputter · Oct 3

Yeah, if they fundamentally can't, then something is fucked up with reality. This would be like finding a kind of smooth, circular wheel that doesn't roll.

Without substrate independence zombies are possible which is kind of a problem if you also want to believe consciousness has causal power and is empirically determinable. Ultimately you will never receive confirmation from the universe whether substrate dependence or independence

1 top reply

Low fat sweaty cat 🐬/acc @Silicon_alien · Jun 6

how does substrate dependence make zombies possible?

Depends on what capabilities are tied to consciousness. If human level capability needs it, and if computing keeps improving in efficiency, then yes conscious AIs are basically inevitable, short of a permanent global order willing to enforce a permanent ban. If human level capability doesn’t need consciousness this likely means a universe mostly tiled with zombies, which is bad in different ways.

1 top reply

Billy G. @BortBurner · Oct 3

There’s growing evidence human consciousness is quantum in nature. These machines are not quantum. They are advanced calculators trained on the output of human culture and history.

How is this a convincing argument at all, no not all neural networks are necessarily conscious, just like i don’t believe all animals are necessarily conscious. I tend to lend increasing credence to an animal being conscious given increasing size and sophistication. We can do the same for AIs.

A puddle of water and a brain are self evidently different in basically *all* ways, that doesn’t leave any room for people to define in it such a way that a puddle is conscious with divorcing that definition *entirely* from its common usage.

1 top reply

Noe Carrier @inkysputter · Oct 8

I used to think that the emergence of the precursors to and AI itself would be a pretty easy sell for humans. Not that it would cause so much disharmony and frank confusion, like it does in scifi. I thought people were better than that.