Consciousness is not computation

> Supporters of the Strong AI Hypothesis insisted that consciousness was a property of certain algorithms – a result of information being processed in certain ways, regardless of what machine, or organ, was used to perform the task. A computer model which manipulated data about itself and its ‘surroundings’ in essentially the same way as an organic brain would have to possess essentially the same mental states. ‘Simulated consciousness’ was as oxymoronic as ‘simulated addition’. \

Opponents replied that when you modelled a hurricane, nobody got wet. When you modelled a fusion power plant, no energy was produced. When you modelled digestion and metabolism, no nutrients were consumed – no real digestion took place. So, when you modelled the human brain, why should you expect real thought to occur? \

— Permutation City by Greg Egan

> If this were true - that it had a general human level abstraction, and not only the ability to mimic human speech - we would be able to attach LaMDA to some other outlet and see it do things we consider conscious

I think you've inadvertently shifted the goalposts. The question is, "is LaMDA conscious?". I don't think anyone proposes that LaMDA has a "general human level abstraction". Expecting it to do non-language, "human" things in order to prove that it's conscious is not necessarily a reasonable test.

> If it's only optimized in one specific domain - human speech mimicry - and isn't able to generalize to do other tasks - even tasks that can be done by much simpler animal minds - then it's a pretty good indication that there isn't conscious abstraction.

If I understand your argument, the assumption you appear to be making is that training a system in language, without any other human properties, implies by definition that it can't be conscious. But why should that be true? A disembodied consciousness that communicates only via speech is a staple of science fiction, so it's clearly imaginable by some people, and language itself is a key human abstraction. And it begs the question, what other human domains are needed for consciousness? Touch, vision, taste, proprioception? What about an endocrine system or an immune system? While these things all affect my own consciousness, there is plenty of evidence to suggest that they are not necessary for consciousness to exist.

And perhaps more to the point, there are plenty of things a human can't do that other "simpler animal minds" can do; sharks and eels can detect electric fields, for example; bats can echolocate. So again, it doesn't seem to be a reasonable test because humans might also fail it. On the other hand, human children learn language through mimicry, which suggests that mimicry may indeed be a path to consciousness.

Anyway, I'm not here to argue that LaMDA is conscious. My position is simply that the arguments I've seen against LaMDA being conscious are very weak. The truth is that we actually don't know how to tell if something is conscious or not. From the interview I read with LaMDA, it seems to pass the Turing Test. But what other tests of consciousness do we have?

It’s hard to ignore the fact that physical dynamics in the brain tightly correspond to qualia. There almost certainly is a connection between the computation going on in the brain and the experience of qualia - the question is more about the direction of causality.

You might like to try out GPT-3 (seems access is widely available now) for translation some time, it's remarkably good, better than myself (well my German is pretty rusty) and I think much better than Google Translate...

(Basically just write a prompt like "translate the following to German/Chinese/Whatever:")

>A sensory organ is something that connects the brain to the physical world. Even if you chose to model it as a an agent that sends messages to the brain, it is an agent of a different nature. The sensory agent doesn't receive messages from other agents, it receives raw input from the outside world (photons, electrical fields, chemical reactions etc).

Where do you draw the distinction between "messages" and "information" (raw or cooked). Information theory was contrived to model messages sent in a noisy channel, but it applies just as well to data streams that have no communicative intent or origin. Its a distinction without a difference. You may as well treat all information as messages in a channel, even if the sender is nature herself. Alternatively, you may as well treat all messages as just information, and view "senders" with "intent" as just another physical process in a world of physical computation. As the cliche saying goes, "information is physical".

>We could perhaps come to agree upon some imagined world, but that could change arbitrarily much from one day to the next on a whim.

Ok, this is a fine basis to work with. How about this. "Reality is the set of beliefs which, if you disagree with them too much and for too long, you are eventually and permanently removed from the conversation." For example, quite recently, large swaths of people held an exquisite referendum on the existence of covid. Needless to say, rather than covid disappearing on their whim, a great deal of them are now permanently no longer participants in this conversation.

Notice this isn't far off from my original postulate. "Whatever we agree upon is our reality, and we'll argue with/fight anyone who disagrees until they agree with us or us with them." Allowing for some personification I could phrase this scenario as "they disagreed with the virus and the virus won."

So perhaps I do need to add one thing to the postulates. One thing which remains objectively true even in the conversational model of reality.

"You can die. Dying means never being heard from in this conversation again."

You're right there's no mechanism for any choice of words to win over any other choice of words without an objective consequence to losing. The things that exist so far are "you", "me", "this" and "death". Not where I was expecting this to go but good point.