Chain of Consciousness

George Rebane

The recent burst of large language models (LLMs) from AI research has reinvigorated the Singularity related topics of machine sentience, sapience, super-intelligence, agency, ethics, morals, …, along with prognoses of what they may have in store for humans when/if they become ascendant.  Most commentators, including yours truly, see a high likelihood of a dystopian post-Singularity future for mankind.

Among these well-known AI experts are internationally renowned scientists like Nick Bostrom (e.g. Superintelligence: Paths, Dangers, Strategies, 2014) and Max Tegmark (e.g. Life 3.0: Being Human in the Age of Artificial Intelligence, 2017).  These tomes are worth reading more than once, which I am currently doing.  It turns that both Bostrom (Oxford) and Tegmark (MIT) are established physicists, philosophers, and AI researchers (and both were born in Sweden).

Their counsel for humanity to proceed carefully into a future not dominated by our species is countered by more hopeful futurists like Hans Moravec (e.g. Mind Children: The Future of Robot and Human Intelligence, 1990) and Ray Kurzweil (e.g. The Singularity is Near, 2005).

About the Singularity, Moravec (Carnegie-Mellon) advises his readers that “human equivalence is just the beginning, not an upper bound. Once the tireless thinking capacity of robots is directed to the problem of their own improvement and reproduction, even the sky will not limit their voracious exploration of the universe. … Moravec challenges us to imagine with him the possibilities and pitfalls of such a scenario. Rather than warning us of takeover by robots, the author invites us, as we approach the end of this millennium, to speculate about a plausible, wonderful postbiological future and the ways in which our minds might participate in its unfolding.”

And it is our postbiological future that I would like to expand on here.  Moravec, Kurzweil, and others of their ilk have been describing a post-Singularity world that is inhabited by both intelligent machines and various forms of ‘transhumans’ – humans who have shed their evolved mortal coils and taken up new forms in both the physical and computational sense.  A popular scenario has some of us living in cyborg bodies.  Another inviting future has us living in virtual worlds as computational entities, and still others have us co-habiting the material (real?) world with superintelligent machines exercising some manner of peer relationships.

The common denominator of all these future forms of existence is that our consciousness, along with other cognitive facilities, is transported to a new and improved processing stratum.  This will presumably happen with little or no discontinuity of our perception of self, nor modification of any stored memories, skill sets, and/or proclivities that were in place before we were postbiologically transported.  Most certainly none of our loved ones, friends, and other acquaintances would be able to detect in us any change or discontinuity – to them we remain cognitively the same as always, albeit perhaps now expressing ourselves through a different interface.

Moreover, as we consider the prospect of becoming a postbiological transhuman, we are assured in the brochure of Amalgamated Transhuman Services Incorporated of the wonderful life a post-transported existence has in store for us.  And by the ability to encode our cognitive metadata into multiple repositories, we are guaranteed virtual immortality (as long as we keep our license current).  But are such continuities of consciousness really possible?

I have revisited this question for many decades, ever since I read a short sci-fi story that ignited a doubt still with me today.  The essence of the story involves a man, just being told of his terminal illness, who has the opportunity to use new technology to cure his illness and save his family, friends, and colleagues from having to go through the sad and disruptive transition that his early death would entail.  After hearing about how the entire process works - starting with a comprehensive body and neurological scan, and ending with a new reconstituted healthy body – he is convinced and signs up for the transport of his ‘old self’ into a new postbiological body.

There is one proviso that our pioneer transhuman becomes aware of only after he has been scanned, and while he is waiting to be introduced to his ‘new self’ to assure that the transport process has gone as advertised.  The attendant in charge informs the ‘old self’ that since this transhuman transport was to be carried out in strict confidence, wherein the ‘new self’ would seamlessly step out into the world of the ‘old self’ without anyone being the wiser, the ‘old self’ would, of course, volunteer to be taken out of circulation and disposed of by means of a painless pharmaceutical concoction.

Our ‘old self’ is left alone to vet his ‘new self’ during which the ‘new self’ passes muster through a series of tests designed for such verification.  When all is said and done, it is time for the old and new to part company, and the ‘old self’ to take his medicine as he contemplates resuming a healthy and happy life in his ‘new self’.  And then with pill in hand, he hesitates and realizes that he is about to commit suicide; he will never again see his family and friends.  While no one else will be able to tell that the ‘new self’ has replaced the ‘old self’, the ‘old self’ will know that in this process, and in all such processes, the chain of consciousness must needs be broken.  The old consciousness must be annihilated, it cannot survive.  Such scans and replacements don’t work.

Once a scan is initiated to replicate an existing cognitive framework, we cannot help but go through a transition of two concurrent sentient individuals.  The only possible way around this conundrum is if the transport process truly replicated one cognitive function at a time in a continuous sequence discernible to the transported subject.  After each step he would always be able to declare that, ‘I am still the I that I was before the last step.’  In short, there would never be an instant during which two concurrent individuals existed, with one having then to be annihilated.

This kind of problem has long been discussed in the ‘Star Trek’ format in which the intrepid chief engineer is always asked to ‘Beam me up Scottie.’  Critics have argued that even in that case there has never been a guarantee that those who exit the beam are the same ones who entered it, no matter that concurrency was never encountered.

But having said all this, does continuity in the chain of consciousness really matter?  Well yes, if we hold with many physicists and multiverse proponents that we and other intelligent critters are transcendent.  Transcendent means that our reassembled existence continues after we suffer what we call bodily death.  For example, quantum physicist Frank Tipler (Physics of Immortality, 1994) taught, that since information cannot be annihilated, it will be reconstituted in the far distant future as our universe expands to its maximum and then contracts to its next Big Crunch.  In that epoch the universe will reassemble and vitalize all the previously dear-departed sentiences.  In the meantime, no matter the nature of post-Singularity superintelligences to be, the chain of consciousness is fragile – we die.

And then there is, of course, the matter of God.

Comments

[Gregory]

“I teleported home last night with Ron and Sid and Meg.
Ron stole Meggy's heart away and I got Sidney's leg.”

― Douglas Adams, The Restaurant at the End of the Universe

[scenes]

https://en.wikipedia.org/wiki/Think_Like_a_Dinosaur
-----

Maybe LLMs just show most human work as a series of tropes. There's not much there there.

I'm not too worried about Operation Paperclip since we still have to navigate two pretty dangerous rapids:

. Disappearance or high leverage of many jobs. They either disappear or 1 can now do the work of 10, without replacement gigs...at least in the short run.
. Weaponization of all this technology. No Skynet needed if a bad person(s) is the guiding intelligence. Triple bonus points for omnipresent surveillance.

[scenes]

Your AI generated ad of the day.

https://www.youtube.com/watch?v=qSewd6Iaj6I

[scenes]

Just a bit of readin'.

https://consciousdigital.org/the-nsas-large-language-models/

Luckily, the 17 intelligence agencies only have your interests in mind.

[scenes]

from the 'Learn to Mine Coal' Department.

https://old.reddit.com/r/ChatGPT/comments/139o1q6/lost_all_my_content_writing_contracts_feeling/

[scenes]

Paper clips and Skynet aside, I was just thinking about the re-do of copyright, patent, IP law generally that so-called AI is going to force.

While this has been in the news...

https://ew.com/music/ed-sheeran-wins-copyright-infringement-lawsuit-marvin-gaye/

Showing the music-industrial complexes need to extract every shekel from everyone, what on Earth is going to happen when an AI can perfectly mimic a style and alter it just enough to pass muster as something new, write millions of tunes per minute, write millions of novels per minute, works in mysterious ways and mines existing works with mechanisms you can't understand.

I can't help but think that IP law generally might need to disappear. You'll see the reappearance of trade secrets as patents become ignored. The value of arts, music/visual/written will drop to nothing as it becomes available at zero cost. If nothing else, the fact that the arts are mainly trope and not difficult to automate should become obvious to everyone.

It was probably time in any case. Systems begun in good faith like this always become misused over time.

[George Rebane]

scenes 828am - Mr scenes, as the 'old self' how do you answer the question I asked - is it ever possible for 'old self' to survive in the sense of continuity of consciousness if the above scenario with 'new self' were carried out? Is such replicative uploading possible?

[scenes]

re: GeorgeR@7:51.

I would think so, assuming the 'old self' physically survives. If not, it doesn't. I can't see the problem of spawning a new copy which then differs due to different sensor inputs and accumulated random internal events that build up. Stretching the analogy, a pair of identical twins represents the same issue. Given the measurement problems, I'd be surprised if you could ever get a bit-accurate copy of a human, but don't see the problem with doing that on synthetic minds using current technology.

My guess is that the ability to near-perfectly replicate a complex physical structure (and perhaps send it's information somewhere) is a lot further beyond now than producing high-function, albeit alien, synthetic intelligence is. Our Lords and Masters will find copying, transmission, backup, modification of themselves a trivial matter. Maybe once people lose value even as grocery store item scanners, our main value will be as pets or biomass.

[George Rebane]

scenes 837am - Not sure I’m communicating effectively. The scenario I outline here has been suggested and used by various sci-fi authors and AI researchers. To make things simpler, assume that you are the ‘old self’ and that the technology exists to perfectly replicate your physical state down to its quantum levels. At that point in time the ‘new self’ coexists with you as the ‘old self’. You as the ‘old self’ vet the ‘new self’ and find it a satisfactory copy to replace you in carrying out your worldly affairs. To make the switch viable and permanent, the ‘old self’ must cease to exist. You understood this proviso when you contracted to upload the complete informational image of your ‘old self’ into a ‘new self’ with whom you are now meeting privately before parting to assume your new roles. To the outside world, the ‘new self’ will be undetectable and the ‘old self’ will be disposed of without a trace. At this time as the ‘old self’ do you consider that you will experience a continuity of consciousness in the ‘new self’ that survives the extinction of the physical ‘old self’?

[scenes]

" various sci-fi authors "

I realize that. That's why I put up a link to a short story using the idea.

" At this time as the ‘old self’ do you consider that you will experience a continuity of consciousness in the ‘new self’ that survives the extinction of the physical ‘old self’?"

No.

I'd say that once the copy is physically constructed it, as I probably didn't communicate effectively, is a separate and non-identical being. You essentially have two highly similar creatures which, in early days, can't be told apart. If the 'old' consciousness is extinguished, Mistah Kurtz, he dead.

The upload (sideload?) would, regardless of it's nature, happen before the disposal, thus there is divergence.

A deeper version is by keeping two self-aware beings in strict lockstep. Perfectly synchronized down to the randomness that pervades everything at the bottom (at least in biological brains and bodies). A kind of entangled thinking. Turn off one. does it's 'continuity' continue? Honestly, at that point I wouldn't know what you would call that step. Generally, I'd say that what humans think of as intelligence and/or self-awareness is essentially an anthropomorphism. Crude words to feel around the edges of a thing we don't understand or really doesn't exist.

[George Rebane]

scenes 1036am - Good response. Actually 'perfect lockstep' during the transition when concurrency occurs is not really necessary. It's OK for the 'new self' to retain memories of meeting the 'old self' and the transition episode. It is the 'old self's' perception of his consciousness in a continuous transition to the 'new self' that is germane to the point considered.

I have no problem with the chain of consciousness remaining intact were the transition to be made in piecemeal fashion so that only one self would be communicative at any one time. Then the 'old self' and the partially formed 'new self' would be able to confirm in turn that they continue to experience their single uninterrupted consciousness. The important concept here is that in this more complex transition scenario there would never exist two concurrently conscious beings, one of which would experience its elimination to oblivion.

[scenes]

A variant argument.

Given that an AI can sign a contract (a contract in the larger sense, not just a legal document), is a perfect copy of the AI held to it?

For that matter, what will form the legal definition of an individual (as opposed to group identities, communicating libraries, God knows what) as AIs improve?

I think I see an opportunity for non-profit profit here. We could get some office space and sound certain about these matters.

[The Estonian Fox]

scenes 11:44 AM-
For legal definitions, refer to the latest version of the leftist dictionary. See trans-individual, trans-group, trans-God, trans-AI. It will all be clear then, at least for a few hours.

[scenes]

"For legal definitions, refer to the latest version of the leftist dictionary."

I'm waiting for the first adult charged with a felony to begin identifying as a child.

AI generated Work O' the Day.

https://www.core77.com/posts/123298/Midjourney-Created-Fake-Ikea-Bomb-Shelter-Furnishings-Catalog

[fish]

Posted by: scenes | 14 May 2023 at 07:46 AM

Going to have to put that Atomborg Atomic Blast Boor on my Christmas with list. I dearly hope it comes in other colors!

[scenes]

lol. Well, that didn't take long.

"AI-Threatened Jobs Are Mostly Held by Women, Study Shows"

https://archive.is/CrjrE

I've decided to become a 'news' writer. First article:

"AI-Threatened Jobs Are Mostly Held by People Of Color, Study Shows"

[scenes]

I was considering the various flavors of AI today and thought of a thing I haven't seen/read anywhere.

a) no doubt professional financial management has been and will, to a far greater degree use all this stuff.
b) folks really don't know how it works under the covers
c) there are beaucoup rules concerning financial institutions, scads I don't know. I don't doubt, fer instance, that buy and selling is somehow governed by weird rules concerning just why it is done, how it is done, can it be done synchronously with other parties etc.
d) the robots will figure out ways to make money that are probably illegal, difficult to detect, no one knows why. Generally, there'll be a scrabble for the regulatory apparatus to (unsuccessfully) keep up. Things like circuit breakers won't be enough.

I've always felt that the stock market has a high bullshit level, mostly due to it's loose connection with actual money being passed to shareholders (via dividends, actual sale of the company, etc.) so maybe this will simply show the emperor's lack of clothes.

[scenes]

hah. Brilliant.

https://arstechnica.com/information-technology/2023/06/chatgpt-takes-the-pulpit-ai-leads-experimental-church-service-in-germany/

[scenes]

Somebody here might find this interesting.

"Jim Keller on AI, RISC-V, Tenstorrent’s Move to Edge IP"
https://www.youtube.com/watch?v=_MrGNlXRi9M