Hacker News Reader

Comment by thomassmith65 | Hacker News Reader

If a device made of human brain cells were designed to function like a traditional, deterministic CPU, that would raise similar questions as LLMs raise today.

View on HN

It would do no such thing. LLMs are neither living nor sentient nor intelligent. The fallacy of computationalism is a failure to read a coarse metaphor as merely analogical. Any correspondence is superficial and limited, and even then, merely correspondence, not identity.

So, the use of human brain cells is different territory. Bioethically speaking, it is illicit to press human body parts (which includes cells) into service for technological ends. They can only licitly enter into service of the integral human whole (such a life-saving transplants). Using them in technology also commoditizes the human body. (The reverse is not the case: it can be licit to use cybernetic devices to restore native function within a human body.)

It will be useful for discussions like this if you state beforehand if you follow any religion eg Christianity.

This seems reductionist with false equivalence. Wouldn't making a brain a "deterministic CPU" negate all the benefits of using organic tissue?

Not if the organic tissue is faster or more energy efficient.

Probably, but the questions I was thinking of are whether it's the design or the substrate that we consider significant.

The question comes up a lot here when there are posts about whether to consider LLMs conscious.

"We're going to condemn conscious entities to slavery"

"They'll always just be machines and it's dehumanizing to consider them anything else"

No, not at all. For one thing, they said "made of brain cells", not brains--those are two very different things. (Yes, brains are made of brain cells, but that's a fallacy of affirmation of the consequent.)

Computers made of transistors were designed to function like computers made of vacuum tubes without negating the benefits of transistors.

> that would raise similar questions as LLMs raise today

I think the questions would be very much different. LLMs are not made of human brain cells.

I get the sense that there is a school of thought I’d call “biological exceptionalism”, a belief that somehow constructs made up of many functionally simple units, is fundamentally different and special compared to a similar construct composed of silicon units. I can’t see how someone could hold this belief and be a materialist (I’m not implying you are btw). Biological exceptionalism surely implies the existence of a “soul” or “spirit” (which raises a whole bunch more questions).

I don't think it'd have to imply a soul - one of these "functionally simple units" is much less well-understood than the other, and you can insert whatever belief you want to fill the gaps.

Maybe a materialist biological exceptionalism position could be that biological substrates are capable of sending kinds of signals that silicon substrates are not, in the sense that brain matter especially might give rise to far greater complex processing in ways that we don't know about or appreciate yet, and which we may discover can't be efficiently recreated in silicon?

I don't know if I buy that argument or even if it makes sense. But yeah this all smacks inextricably of soul talk.

I agree that as per current understanding, the functional biological base units have more degrees of freedom and a greater functional diversity as a result, but that’s merely a quantitative difference not a fundamental distinction. One could imagine equally more complex silicon base units beyond what we have now. Whether they need to follow the same pattern/modes as the biological units remains to be seen (but it’s pretty hard to argue against evolution as an efficiency maximiser).

I think it goes the opposite way, starting with a soul and arriving at biological exceptionalism: You start by saying "I have a soul / spirit". Then you extrapolate "things like me probably have a soul / spirit". But since there is no theory of souls, the extrapolation is not based on principles but rather on vibes and guesswork, ending up somewhere like "there must be something weird going on with neurons that somehow create a soul, connect with an existing one or something like that.. maybe.."

Which is an interesting argument because: do neurons somehow create a soul (as you mentioned), or are do neurons somehow attract souls, or do souls choose what to inhabit. In the first case it posits that a “soul” is purely a function of biology. In the others that there’s some sort of inextricable link between the two (which in the absence of any theory of soul or how these are linked to biology, seems rather far fetched, unless of course souls could choose to inhabit anything with sufficient degrees of freedom: a brain, a mega computer, a planet, a galaxy)

Reply on news.ycombinator.com