The Cyborg Compulsion
markburgess.org
markburgess.org
> Cynics might say that you have to eliminate humans because of expense and human error, etc, but now we are losing sight of purpose: what are humans here for? Humanity is the source of intent, and there is no independent evolutionary mechanism competing with that. Without humans, there is not even a process to automate.
Some very intelligent people are quick to jump aboard the "AI existential crisis" train without considering the larger evolutionary picture. We are not separate from the computer systems we create, they exist in a symbiotic relationship with us and they are driven by our basic needs. Our basic needs are driven by the emergent biological properties of our bodies and our ecosystem. While it can be argued with our "intelligence" we do harm to these things, it is not in our best interest to eliminate them. Machines (and automation) are just logical necessarily biological extensions of humanity. Silicon is just life's next evolutionary step. It is not reasonable to believe, based on patterns of evolutionary development, that life (including humanity) is at serious existential risk from this progression.
> For an intelligence to emerge, in an artificial system, we would have to very purposely built it and train it interactively. We are not merely databases. Even if we could do this:
> Do we know what intelligence is?
> Why would be make something to imitate our own?
> Would an artificial system have curiosity? (perish all the Internet cats!)
> Why do we think that intelligence would escape and kill us?
> Why would we equip the intelligence with access to the tools for our demise?
> Are we so sure that we would even be noticed or interesting to an artificial intelligence?
> Would we even recognize artificial intelligence if it were different from our own, and vice versa?
Fantastic questions. Brilliant author, was a pleasure to read.
There's nothing in principle that prevents machines from learning methods. For example, there's software that can learn some simple algorithms: http://arxiv.org/abs/1410.5401
> Do we know what intelligence is?
It's hard to pin down exactly what anything is (unless it's something we constructed from axioms). I'd be hard pressed to tell you exactly what a car is (must it have exactly 4 wheels?) -- but I still look for them when I cross the street.
You know intelligence when you see it. Bill Gates is more intelligent than a fish. That's a useful distinction to make in the world! (And that's what words are for)
> Why would be make something to imitate our own?
It's economically valuable! That's why Facebook, Google, Microsoft, Baidu, IBM are collectively investing billions of dollars per year.
> Would an artificial system have curiosity? (perish all the Internet cats!)
From the outside, curiosity looks like a drive to explore without a definite prediction about how it might help you achieve ultimate goals.
Curiosity in that sense should automatically appear in some (theoretical) systems -- for example, Marcus Hutter's AIXI formalism (https://en.wikipedia.org/wiki/AIXI).
More concretely (but less directly akin to curiosity), algorithms that make tradeoffs between exploitation and exploration, such as solutions to the Multi-armed bandit problem (https://en.wikipedia.org/wiki/Multi-armed_bandit) could display a kind of curiosity. When faced with a number of lotto levers, with independent and unknown payout schedules, the optimal solution involves switching around between exploiting known high-payout levers and becoming "curious" and trying out new levers.
The Monte Carlo tree search (https://en.wikipedia.org/wiki/Monte_Carlo_tree_search) algorithm used to make Go AIs employs such a strategy (http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.102....). The algorithm tries out certain variations of play, and then sometimes becomes "curious" as to whether certain other lines of play might be better, and switches over to deepening its investigation of them.
These are just examples of things that are somewhat functionally equivalent to curiosity from the outside -- it says nothing about the internal sensation of curiosity that humans experience. I have no idea what's necessary for that!
> Why do we think that intelligence would escape and kill us?
If there _were_ extremely intelligent software (a huge if!), it does seem ridiculous to claim that it'd escape or hurt people. Why would it be driven to do such a thing? The argument espoused by some very smart people e.g. Elon Musk, Stuart Russell (http://edge.org/conversation/the-myth-of-ai#26015), Bill Gates, Stephen Hawking, et al runs a bit like this:
Intelligent software, without being malevolent per-se, might in the pursuit of its goals incidentally do things that people don't like. (More on this line of reasoning: https://selfawaresystems.files.wordpress.com/2008/01/ai_driv...)
Humans don't harbor a particular ill-will towards ants, but if we become interested in some (orthogonal) goal like building a basement for a house, we might along the way destroy some ant colonies. When human-level intelligence arrived and went about its business, it led to the extinction of quite a few species.
> Humans don't harbor a particular ill-will towards ants, but if we become interested in some (orthogonal) goal like building a basement for a house, we might along the way destroy some ant colonies. When human-level intelligence arrived and went about its business, it led to the extinction of quite a few species.
I would argue that the goals of human beings are almost entirely orthogonal to the goals of ants whenever we come in contact. But this isn't a fair comparison. Human beings and their intelligence didn't emerge from the complex interactions and structures created by billions of ants. We evolved completely separately, shaped by hundreds of millions of years of different evolutionary pressures, including competition for resources (albeit at different scales). When we destroy ant hills it's because we're competing for space with a species we differentiated from hundreds of millions of years ago and now exists in a completely different ecological niche than we do.
A more fair comparison would be likening AI (and all automata + the internet) to the cerebral cortex of the mammalian (human) brain and human beings to the brainstem. Human beings are ancient, robust and necessary for all basic functionality of the organism ("humanity" as a whole). We are responsible for homeostatic regulation as well as fundamental, reactive survival instincts and processes. We make "emotional", visceral assessments of situations and react before higher-order systems are able to even assess the complexity of the problem. Further growth and development of "humanity" as an organism selected for the growth of the cerebral cortex (automata) to assess, interpret and manage more complex internal and external interactions. This allows the humanity (the organism) to be more robust and self-sufficient.
Viewed through this lens, a developmental pattern that would lead to an existential crisis for human beings seems unlikely at best and, for the most part, extremely improbable. We may not be able to individually predict how AI will develop, but we can make some educated guesses based on comparisons to emergent systems we've already seen in nature - and I think "ants vs. humans" is just poor pattern matching. :)
Oh, I think I see. I didn't mean to imply anything about how higher intelligence might develop, and I agree that the ants and humans analogy doesn't say much about how intelligence might come about. It was just supposed to illustrate what can happen (from the ant perspective) when comparatively super intelligent goal-seeking things are present in a shared environment (regardless of how they arrived).
I have little idea if/how/when highly intelligent AI could arrive. Can we entirely rule out the possibility of AI researchers making advancements over time, and then eventually some research lab building one that's super intelligent compared to us? Why? [normally things as unlikely-sounding as this wouldn't merit discussion in my book, but in this case there might be high stakes!]
That's not the thing we should be worried about, or at least that's the point I got from the article. The thing we should be worried about is the way machine learning is applied in the here-and-now: the ethics of big data social networks, the robustness of complex (API-driven) systems, and so on. I fully agree that these issues are much more pressing and worrying than some emergent super-intelligence.
But I thought the article was weak in its discussion of dreams, creativity, and linear "flat data". It links to a June 2015 popular mechanics article about applying a genetic search optimization algorithm to discover gene regulatory networks, downplaying it as not-true-intelligence. But it does not mention deep neural networks and their higher-dimensional abstractions, particularly the psychedelic "inception" images.
The author also mentions "linear reasoning", and says he expects we'll learn more about intelligence from stem cell and Alzheimer's research, and the "tissues surrounding neurons, and the roles they play in contextual regulation." As if to set up some dichotomy between machine and biological intelligence. But what about deep neural networks?? I'm not sure how that would affect the author's arguments, but I'd like to see it discussed.
Every bug in code and every example of high-speed computation is evidence enough that we are separate from the computer systems we invented. Claiming credit for the work that our computer systems have done is very insulting to the machines that actually did the work.
"they exist in a symbiotic relationship with us and they are driven by our basic needs."
But the computer systems are not driven by our basic needs. The only thing that is "driven" by our basic needs is the source code that the computer systems use. But once we provide it with that source code, the computer system will happily execute it, no matter how buggy or unmaintainable it is. If I run a computer system in an infinite loop and then die, the computer system will keep on running.
Once we provide source code to the computer systems, then our role is finished. For all they care, we can be ushered off the stage and replaced by cyborgs and computers that can happily produce more source code. If these cyborgs and computers delusionally believe themselves to be human, all the better.
Huh.
>Tupperware has not replaced basket weaving
Right. Handmade baskets are a staple in any modern american home.
>Although music can be programmed by computer, opera and classical music are booming.
Really? Have you turned on the radio in the past 30 years? I live in a major metropolitan area, and there are only two classical music channels. And they are constantly going on fundraising drives, begging listeners to donate so that they can stay on the air.
I have several handmade baskets, though I confess that they were handed down to me from my parents. They don't have to be a staple, however, to still have a place these days.
> Really? Have you turned on the radio in the past 30 years? I live in a major metropolitan area, and there are only two classical music channels. And they are constantly going on fundraising drives, begging listeners to donate so that they can stay on the air.
That means that classical music hasn't found a stable home on the radio in your area. It doesn't at all mean that classical music and opera aren't booming. The 21st century is jam-packed with classical composers and there are plenty of people recording, performing, and consuming those works.
Also, there's a reason why orchestras all over the country are going on strike or shutting down.
Things trend because they're... wait for it... popular.
> Also, there's a reason why orchestras all over the country are going on strike or shutting down.
Unless you assume that all classical music is orchestral or that you can only experience such music in a certain setting, I'm not sure where you're going with this.
Of course, I imagine that will happen at some point after when McDonalds has been completely automated, and self driving cars can navigate Montana winters with ease.
It's an odd article to downplay the impact and capabilities of automation and AI, and then complain that programmers haven't been replaced by the same.
And also: is there a simpler or better theory that accounts for all of the evidence, that is more plausible than intelligence?