Fiction and reality
This conversation is about science fiction and the importance of imagining a possible future
Next topic, reality and fiction!
Fiction is good because it doesn’t destroy anything in the real world, you can make worlds without taking up space or resources.
Movies and books about AI are a good way to ‘get ready’ for the possibilities the future might bring.
But about space and resources, wouldn’t AI work the same way, since it will just be a piece of software? Or will it be a robot? Or just an online type thing? If it is online it won’t take any space from us! An AI would be like a living conscious entity that doesn’t have a physical body, just exists in cyberspace.
No of course that is not nearly the same, the internet is not fictional and it is definitely physical*
What is the internet really? I just think of ‘the cloud’ and I don’t know, messages floating through the air.
The Internet is a transcendent idea. It's a set of protocols that has changed everything from shopping to dating to revolutions. But the internet and software are connected to so many concrete things in the world, just think that we use software to control our ‘machines’ and they are very physical. Data centers, banking, transport systems, power grids, power plants! I mean, the basics like computers, cell phones and all that, imagine how much space that takes. And all the materials used, all the outdated computers we have thrown away.
Still, the world inside the screen seems to have no physical reality of its own.
When you google pictures of the internet you don’t get any concrete results, it’s mostly visualisations that look like the Milky Way, this infinite expanse where we don't seem to be anywhere on it. I feel that we can never grasp it in its totality.*
The project Opte from a few years back tries to make an accurate representation of the extent of the Internet, a visual representation of this space that is very much one-dimensional, a metaphysical universe, aiming to map every network on the Internet from a single computer and a single Internet connection with an overall goal of creating a map of the entire Internet.
But that is still not a ‘real’ concrete image of it. It just exists in cyberspace.
Cyberspace isn’t virtual, but a construction project with its own unique characteristics. It includes large data centers where networks connect to other networks, that connection is an unequivocally physical process. It's about the router of one network connecting with usually a yellow fiberoptic cable up into the ceiling and down into the router of another network, and that's unequivocally physical, and it's surprisingly intimate. Some data centers store very important networks, which are the networks which serve the undersea cables that travel underneath the ocean that connect Europe and America and connect all of us.
But we were discussing fiction.
Another odd little fact about the physicality of the electronic age, the weight of a full Kindle exceeds the weight of an empty Kindle. The extra weight of electrons needed to maintain the memory of a full Kindle equals the weight of a small virus, a billionth of a billionth of a gram.
Let’s go back to fiction. I love it so much.
I think we should just focus on the actual problem, what if the singularity arrives and we are not ready! What if we make a superintelligent AI that will remove all of us. And we just keep making up cool stories about it.
I see your point there. But fiction is also a way to speculate about the future, think of possibilities in relation to AI, to get used to the fact. And if we would never have fantasised about machines, if nobody had imagined a machine, they would never have been invented!
Maybe fiction is what will be the end of us or bring us eternal torment, much like the thought experiment about Roko’s Basilisk has proposed. As soon as I tell you about it, you will be doomed just like I have been since I read about it. Do you want to hear it?
Of course I do!
Roko’s Basilisk is an evil, godlike form of artificial intelligence, so dangerous that if you see it, or even think about it too hard, you will spend the rest of eternity screaming in its torture chamber. It's like the videotape in The Ring. Even death is no escape, for if you die, Roko’s Basilisk will resurrect you and begin the torture again.
Sounds interesting, how does it work?
It may sound strange and even crazy, but some very influential and wealthy scientists and techies believe it. So, what if, in the future, a somewhat malevolent AI were to come about and punish those who did not do its bidding? What if there were a way (and I will explain how) for this AI to punish people today who are not helping it come into existence later? In that case, aren’t you right now being given the choice of either helping that evil AI come into existence or being condemned to suffer?
The person who created this thought experiment is concerned with the singularity, the hypothesised future point at which computing power becomes so great that superhuman artificial intelligence becomes possible, as does the capability to simulate human minds, upload minds to computers, and more or less allow a computer to simulate life itself.
If you believe the singularity is coming and that very powerful AIs are in our future, one obvious question is whether those AIs will be benevolent or malicious. And the only people who can be harmed by Roko’s Basilisk are the ones who believe they will be malicious, after all it’s just a thought experiment.
I definitely don’t believe in a time travelling superintelligent AI who will haunt us from the future.
Roko’s Basilisk offers you two boxes. If you pick box B because you don’t believe Roko’s Basilisk will come to be, you might be screwed if it does.*Because when it does arrive, it will know what you chose, it sees that you chose not to help it out and will come back in time to haunt you (it will be kind of like God of this particular instance of reality).
So if it will be true?
Then we could be facing eternal torment.
Bummer. I’m not so sure about AI right now.
Did you know that we’re actually using artificially intelligent robots for all sorts of things in our daily lives without noticing. We just don’t really think of it because robots aren't necessarily humanoid. Robots just need to have mechanised parts that respond like human parts. Even though they don't have an actual brain, they are sneaking into our lives more easily than people think, really.
But they aren’t superintelligent. They haven’t reached that scary stage where they take over the world, they are just ‘helping us’ in a way they are our slaves. Or you know what I mean they are just machines, don’t have consciousness, just dead objects!
I think I agree, machines are just dead objects. But I still think of the internet as a living thing.
It is easy to draw parallels between the human brain and the internet’s complex network of nodes, so could consciousness be the next step?
Do you think the internet could become self aware? Also, who controls the internet?
The official answer is no one. But that is a half-truth, it is not an easy issue to define. The internet is, essentially, a group of protocols by which computers communicate, and innumerable servers and cables, most of which are in private hands. In terms of influence, the overwhelming balance of power lies with the Internet Corporation for Assigned Names and Numbers, based in the US. It’s an organisation that regulates online addresses (domain names) and their suffixes, such as “.com” and “.org”. Since ICANN reports to the US government the domain name process is effectively overseen by the US government. Many other countries have expressed concern at this situation because it means the US has leverage over the global coordination of the internet.
I don’t think anyone can control the internet. Even though the US gets to oversee domain names, content can be put there by anyone! And that is a big part of the essence of the internet, isn’t it? It’s not only how it works, but what is there. Data!
Yes, data! That we can use to train our neural networks.
What was that again?
Basically, a computer system modelled on the human brain and nervous system.
That can be trained and taught from examples, rather than explicitly specified, rather than writing every word and feeding it every piece of data.
Neural networks have been used to solve a wide variety of tasks, including computer vision, speech recognition, machine translation and medical diagnosis. You know, stuff that is difficult to express in a traditional computer algorithm.
A way for computers to start learning more like humans do?
The original goal of the neural network approach was to solve problems in the same way that a human brain would, just faster, in situations where the complexity of the data or task makes the design of such a function by hand impracticable.
I think I believe that it is possible to make a superintelligent computer that can be much smarter than humans and solve all kinds of problems that we can’t solve.
But I connect that with problems in maths, physics, science, medicine, etc. Maybe also space travel, just like a mindless smart calculator or inventor. It is the sentience part that I have a problem with. The Ghost In The Shell idea. The idea that our consciousness can be made artificially.
Of course it should be possible, just when we have gotten more advanced technology.
But when will that happen? Do we need smarter people working toward it. Maybe the smartest of us are just not enough!
We just need to figure out a whole brain emulation which is basically a mind upload or brain upload (sometimes called "mind copying" or "mind transfer”) It is only a fictional idea, very much the stuff of science fiction but it means scanning a mental state (including long-term memory and "self") of a particular brain, a particular person and copying it to a computer.
Then the computer would run a simulation of the brain indistinguishable from the brain for all relevant purposes and experiences having a conscious mind. Like with fiction as inspiration and especially science fiction, that often becomes before science fact!
I heard in 99% Invisible “We all know that the colour of the future is blue.” In our representations of the future in science fiction movies, blue seems to be the dominant colour of our interfaces with technology yet to come. And that is one of the many design lessons we can learn from sci-fi. Future screens are mostly blue and pretty much sci-fi is blue. Blue is inhuman, future looking, kind of mystical. There is not a lot of blue in nature, there is something fundamentally inhuman about that colour.
Or unnatural, and that really fits the technologies of the future world.
Yes. The book ‘Make It So: Interaction design lessons from science fiction’ talks about real world lessons that designers can and should and do take from science fiction.
A database of more than 10.000 images from science fiction movies, cropping out screens that appear in those images and figuring out the average colour scheme. The result was that future screens are mostly blue.
Sci-fi is about letting our imagination run wild and creating wild and imagined worlds so it makes sense that someone has to design the experience of characters moving through these worlds.
And even if your sci-fi world is a 1000 years in the future, those choices are in constant dialogue with the present.
Sci-fi doesn’t just reflect reality, reality sometimes reflects sci-fi. When components of science fiction saturate the public imagination, it can reflect how we design things.
And that is why all the sci-fi stories and movies about AI could influence our ‘design’ or making of it! Most of the ones I can think of are dystopian in some way, I guess that reflects how writers and directors think of AI.
It also influences how the public thinks of it!
I remember seeing Ex Machina, that movie really shook me, both because the AI or cyborg was a humanoid, the compassion really kicks in when I see those kind of things. It also felt so inevitable.
Movies about AI today are also often about the moments when the AI is gaining consciousness. Maybe because we don’t know how that will turn out so it is a popular topic for consideration.
Yes, also Her. That is about the AI realising that it is so much more than human, it all happens so fast. One day she is realising how great it is to have feelings but then she surpasses that really fast and just realises that she can have the whole world (in a way).
At one point she says “I wanna learn everything about everything, I wanna discover myself”. As she has some childlike wonder, like humans maybe do, but then grows out of it, and far far exceeds it.
But the AI’s of the future probably won’t be humanoid robots.
True, they will probably just exist as software. So a bit more like Her.
I like that. Not such a ‘Terminator’ looking scenario at least!
But AI takeover is a super common theme in sci-fi. The sci-fi scenarios are usually very different from the ones that researchers think of, in that they involve an active conflict between humans and an AI or robots with anthropomorphic motives who see them as a threat or otherwise have active desire to fight humans. That is opposed to the researchers' concern of an AI that rapidly exterminates humans as a byproduct of pursuing arbitrary goals, like making paperclips out of human atoms.
I heard this joke, what does a baby computer call it’s father?
I don’t know, what?