In Isaac Asimov’s classic novel series (“Foundation”), the long-dead Hari Selden had set up a “simulacrum” of himself-whether this was a robot or holographic image is not clear (Asimov never gave any details).
At any rate the simulacrum would appear when the Second Foundation faced a “Selden Crisis” (an epoch in which the Foundation was under assault).
My question: at the present level of technology, could a believable human simulacrum be built? Suppose you could store millions og megabytes of personal data-could you have a reasonable facsimile of a human brain? Could such a program answer questions?
No, no, and no. And to answer the question in the title: probably never, but it doesn’t hurt to dream.
Meaning such that it mimics correctly the thoughts of the person it is a copy of?
At present level of technology we can’t really put together that level of processing power.
Also, the storage of mass amounts of data does not equal intelligence or ability to think. If it did then Google would be smart, which it isn’t.
Something like that is probably pretty far off into the future, somewhere between 50 and 500 years.
No problem at all. Just program it into the holodeck.
If you could recreate humans, you’d become like another Og.
I am a human simulacrum, so yes. I’m 25 years old too, so the tech has been around for quite a while.
Its impossible to say. We really don’t know a whole lot about how the brain works. We can more or less model how a neuron works computationally, so maybe when we have the processing power to create a network equally sophisicated to a human brain, we might believe it would have similar behavior. But there’s tons of problems there. Is our ability to model a neuron even accurate? Is there more to how the brain works than just connecting a bunch of neurons in a similar pattern, particularly with relation to chemical or possibly quantum effects? If our model is accurate how does the simulation error compound? It very well could compound quickly enough, which it likely would because of the high degree of interdependency, then even an extremely accurate simulation would still degrade into utter garbage very quickly.
I think we’re far enough off of understanding how it works and that the eventuality is complex enough that we simply cannot make it happen without a fundamental technological shift, like to biological or quantum computing and so I can’t possibly see it happening until such a breakthrough happens, and then there’s still a matter of developing it. I think any estimate that’s within this century is extremely optimistic.
I think we’ll see better and better fakes as time goes by. The most modern versions of “Eliza” type conversational programs are eerily convincing, and, of course, we all saw “Watson” kick butt on Jeopardy!
Purists will insist that we can never truly obtain machine intelligence. I don’t agree, but I also say it doesn’t matter: the approximations will be more than good enough.
(A Cal-Tech computer scientist once said to me that modern pseudo-random numbers are so close to truly random that “God himself couldn’t tell the difference.”)
I’m betting that we have functional AI at a level good enough to serve as aides, butlers, shopping assistants, etc. inside fifty years.
I should have been clearer in my definitions.
What I imagine a “simulacrum” to be: a program that would pull up my dead uncle’s image-and I could talk to him-meaning that the program would mimic his manner, behaviour, etc.
In short, I would gte the definate impression that I was talking to a human being.
Is such a thing possible?
I have step by step instructions to do this that I don’t mind sharing with all you fine folks on the internet. Follow these directions and I guarantee you’ll attain immortality and wealth beyond your wildest dreams! All you need to do is work out a few of the minor details I can’t be bothered mentioning and profit!
Step 1: Realize that “you” mostly reside in your Neocortex in your brain. The Newcortex recieves electrical signals from nerves connected to your body’s senses, processes the information (your thought process) and then returns electrical signals on nerves back to the body where the signal is translated to mechanical action.
Step 2: Intercept the input signals and replace with whatever information you want the brain to sense. Spinoff applications: Cure blindness, deafness, etc.
Step 3: Intercept the output signals from the Neocortex and use those to perform whatever tasks you want. Applications: cure paralysis; eliminate the need for remote controls; reasonably approximate telekinesis!
Step 4: Create a virtual computer world. Build a device that a person can use to override the inputs and outputs from the brain so they can truly feel that they exist in the virtual world you create. If any of you have seen Caprica (anyone? anyone?) the holoband is exactly what I’m describing. From here on out, I’ll refer to this as V-World. Applications: unlimited! The first application that will be made? Porn.
Step 5: Now that you’ve fixed the inputs and outputs, it’s time to improve the Neocortex itself. Create a device that mimics how your brain processes information and connect it to your Neocortex so that new thoughts and thought patterns can be stored inside the implant in addition to your meat brain. As the meat brain becomes less efficient with age it will naturally tend to store more and more of “you” on the implant. The implant can also pull existing patterns from the meat brain to improve memory retention and retrieval. Applications: Cure for old age senility; cure for Alzheimers; super intelligence!
Step 6: Die
Now that your meaty self is dead, remove the implant which contains your thoughts (ask for help on this step, you may have difficulty doing this yourself) and connect it to V-World. This is NOT a copy of you, it IS you. You can simulate a 25 year old body and as far as your brain is concerned, you’re still completely human!
Step 7: Give this to everybody you know, death is no longer the end. When people die they still live on in your virtual world. When a person needs to ask his great great grandfather what it was like living in the past he can just ask him over the beers they’re sharing after playing a round of golf in V-World.
So there ya go - 7 extremely easy steps spelled out in detail. Now get to work! I want to see prototypes by next year!
A true emulator of how the human brain works would be difficult and will require lots of processing power. That isn’t something I want to say is impossible, but when it happens it will be exceedingly expensive and will require a lot of infrastructure, this will be a massive array of supercomputers with thousands of processors working in tangent. You’ll need a whole support staff to keep everything running not to mention significant cooling and etc.
Quantum computing is essentially in an embryonic stage right now, but if that ever takes off it could result in serious miniaturization of all of this.
But anyway, actually emulating the real human brain isn’t necessary. You always hear projections of how many FLOPS the human brain operates at, but keep in mind the human brain has to process continuous input of massive amounts of information (through the senses), and has to govern various automated processes that are necessary for keeping the meat-sack of our bodies alive. If you just wanted to emulate a person’s personality, mannerism, voice, etc in some hologram you could do it with considerably less processing power than a human being has.
You could create a hologram projection of this person, that would be capable of answering voice queries and perhaps doing so with something akin to the mannerisms of the human that inspired the simulacrum. That’s not nearly as big a feat.
So far, quantum computing is a niche that is well suited to just some specific types of problems (e.g. factoring large numbers), it’s not a more powerful general purpose computing platform.
I agree with you and have indeed thought about this pathway to immortality myself, however, because there are no sustained processes that link you to the “rebooted” you every time you start up, there isn’t really a difference between you and a copy of you! Let me explain.
– Let’s say that all brain activity were to completely stop in a person, then re-start. It would be equivalent to your scenario of a portable electronic you.
– However, let’s say that all brain activity were to stop in a person for around 20 years. In that time, all of the molecules in the brain would have been replaced by other molecules. When their brain starts up, there is no way to differentiate between that and a different copy, because none of the molecules are the same.
– Let’s say brain activity stops, but starts again short of complete replacement. There isn’t any one point at which you can say “beyond this level of replacement, it’s a copy that’s restarting rather than the original.” Ad infinitum.
– Via this logic, there is no difference between stopping and restarting the brain without replacing anything, versus stopping and restarting the brain with a full replacement.
I am forced to conclude that “you” reside in the processes of your brain rather than the contents. So as long as your machine does not stop I agree that it will still contain you and not a copy. But if it stops, I may as well cease to exist.
Note that I also have the same misgivings about surgery, if it stops brain activity. I seriously would resist undergoing full-anaesthetic surgery unless it was life-threatening. (Note that at night when you sleep, brain activity does not stop. I think. :eek:)
I say…yes. Not today, but it is within a meaningful technological horizon. Like landing astronauts on Ganymede or sending an unmanned probe to Alpha Centauri. Frazzled pointed out how it could be done.
(In practice, it would actually be easier than it should be, since we, as humans, are instinctively programmed to react to images as if they were persons. Little children playing with dolls…or even adults interacting with muppets. We’re a species that sees faces in clouds!)