Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Meh. If we want immorality we'll probably end up converting a brain scan to an LLM.

Medically this could be great for preventing brain damage.



ANNs can't represent everything involved with a functioning brain, never mind a highly specific ANN architecture. Any consciousness or similar would need to arise via an independent mechanism.


> Any consciousness or similar would need to arise via an independent mechanism.

Probably, but even then only due to there being a lot mechanisms and we don't know which of them is related to the subjective experience of existing that is the meaning of conscious I assume you're using here (there's something like 40 definitions of "consciousness").

But because we don't really know where the capacity for an inner experience even is, it's not impossible that even an LLM, which totally isn't designed with the goal of having it, might nevertheless have it.

(I really hope they don't, I statistically suspect they don't, I just can't rule it out).


It's not impossible that rocks have consciousness. Some modern philosophers advocating pantheism think they just might.

Meanwhile, according to Sabine Hossenfelder, a recent paper shows evidence that Penrose was right about quantum effects happening in the brain:

https://www.youtube.com/watch?v=R6G1D2UQ3gg


> Some modern philosophers advocating pantheism think they just might.

This is a terribly low bar. The map of “ Some modern philosophers” “think [something] just might” [something] is an incoherent patchwork.

I think that theorizing about things without hard evidence has value. What-if’s teach us a lot, by raising important questions, and keep our ideas flexible, preparing us for the next crazy insight that nature reveals.

But that isn’t the same as attributing any reliability to the as yet unproven enthusiasms of philosophers.

There is a clear case for a survival benefit for creatures whose awareness and control extends into their own thinking. Thus a reason consciousness would be one of many strategies evolution may produce.

What natural phenomena is explained by a conscious rock hypothesis? What would cause a rock to organize in such a way? Where is this information about itself encoded and experienced in a rock? It’s not a coherent conjecture.


It is incoherent because it lacks evidence, not because the idea is absurd.

My computer is made from rocks.


Natural formation rock consciousness both lacks evidence and is absurd, as there isn’t even a coherent reason to think it might or could.

Your computer processes information, and can be configured to process information about itself. At least the possibility and potential utility exists that conscious computers may one day exist.


Humans are a natural formatiom. The laws of science are our only real limit here so let's be honest about this.


That is a breathtakingly pedantic reading of what I said.

I used “natural” to distinguish minerals created via geologic processes, from minerals and elements intentionally processed into machines.

Similarly, evolution’s long process of accumulated innovations separate humans from our constituent elements as found in the geologic environment.

Both technological and evolutionary iterative and self-reinforcing adaptations provide a mechanism for complex coordinated high level information flows to arise aimed at survival (biological or economic) where consciousness has both a potential purpose and a plausible process for coming into being.

A lump of granite or obsidian presents no evidence of consciousness, no purpose for consciousness, and no plausible process for having accreted or transformed into something with consciousness.

The conjecture that such rocks may be conscious stands on equal footing to the conjecture that they are unicorns wearing magical rock-disguise cloaks.


> The conjecture that such rocks may be conscious stands on equal footing to the conjecture that they are unicorns wearing magical rock-disguise cloaks.

Hah! I would love to prove you wrong by imagining the perfect geological formation (many rocks as humans have many complex organs) that can receive and transmit coherent, dense information,

but I am a mere mortal.

For all practical purposes you're right, and for all philosophical purposes The Thinking Rock is an extreme example.

Let's revisit this thread in 5000 years after humans have gone interstellar and see if we can't review some evidence of unconventional life forms.


> Let's revisit this thread in 5000 years after humans have gone interstellar and see if we can't review some evidence of unconventional life forms.

That would be wonderful!

I have no doubt that different chemical foundations, and different environments will yield extreme differences in morphology.

The uncountable differences between humans and the octopus are a vivid example of how little aliens may resemble us.

It would be most interesting to know if life could evolve in oceans under planetary crusts, sunless rogue planets warmed by radioactive decay, in the extreme conditions around magnetars where chemistry takes on different rules, in thick rings or atmospheres around gas giants, or in methane or other non-water “oceans”.

Surely somewhere there are energy conserving hibernating “rocks” that we might be well surprised to discover are sentient!


> t's not impossible that even an LLM, which totally isn't designed with the goal of having it, might nevertheless have it.

I was referring to the idea that a brain scan of a human could be copied into an LLM.


Oh, in that case I agree. My understanding is LLMs (or at least GPT-architectures) aren't Turing machines, so they can't simulate arbitrary other systems even if you made them very big, and because of this my guess is that it would be extraordinarily unlikely for them to just happen to have the right shape and power for simulating a full human brain.


I think OP means how do we know that the brain tissue that is being used in the experiment is not conscious.

Here is a man that lived a normal life with 90% of his brain damaged, a large portion of that just completely missing.

https://www.sciencealert.com/a-man-who-lives-without-90-of-h...


Sneaky paragraph at the bottom:

> Update 3 Jan 2017: This man has a specific type of hydrocephalus known as chronic non-communicating hydrocephalus, which is where fluid slowly builds up in the brain. Rather than 90 percent of this man's brain being missing, it's more likely that it's simply been compressed into the thin layer you can see in the images above. We've corrected the story to reflect this.


I did not see that actually, thank you.

The fact our brain can be compressed to that level is pretty crazy too, although not as impressive as the original missing 90% :)


You're welcome :)

I very nearly missed it on that page, despite already being aware of the detail from other sources.


Network pruning.

In ANNs, pruning 90% of the weights without substantial loss isn't unheard of. I guess this may be analogous: continuous pruning and fine tuning over a lifetime. Though, is removing 90% of the brain more analogous to 90% of weights, or 90% of the rank?


At this point it feels like we’d required the resources of an entire planet to run a single full-fidelity virtual brain. Which leads to interesting science fiction premises.


Or one entire actual brain - if it can be done once by natural selection, it can be done more efficiently again by a designer.

Time exists because the universe doesn't do calculations - the only way to see the outcome is to do the thing. See: The Three Body Problem



Last I heard (I am not a brain scientist) we don't really know what fidelity we would need.

https://en.wikipedia.org/wiki/Mind_uploading#/media/File:Who...




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: