After listening to several AI writers and speakers over the years, I am now asking them: why is developing AI so important? Yes, I understand it’s the next l
I will try to summarize in as few ideas as possible, the main reasons I think such people are excited for AGI, then challenge their ideas with deeper “whys.”
Curiosity
Perhaps the most common reason I see is that we will finally learn all the answers to the mysteries of science and the universe. Finally, all math, science, then medical, and other mysteries will finally be solved, including the infinite numbers of problems we have not even thought of yet, or that will be discovered by the neural networks.
This appeals to the curious minds, and is mainly why AI proponents also happen to be scientists because such tend to have a higher curiosity on average, or at least want to know the answers to riddles and problems. Want to know if we can build a mini black hole in the Hadron super-collider? You probably do not, but many scientists are compelled to know to the point of actually testing the idea while risking sucking in the entire earth, even if the risk of it happening seems fairly small. They just have to know, or perhaps win a prize.
Of course, building AI also fulfills the unquenchable thirst for knowledge among scientific and business minded. Who is not interested in a computer that knows the answer to every question ever posed and can create unlimited wealth?
But, the question is, does knowing the answer to everything actually make our world measurably better? If so, prove it to me. In some sense, knowing everything may lead to disappointment and boredom. Having mysteries keeps people busy and motivated.
Solving Societal Problems & Improved Technology:
Scientists want to create better technology, and businessmen want to distribute them, but to what end?
Will we look back and think how much we suffered in a world before a pizza would appear in front of us in less than 1 minute?
Even if we embed advance computing abilities into our minds, I am not sure why this is good or useful. So, I can instantly calculate pi to a quadrillion spaces, chat with 1000 people at once, digest all of Wikipedia in a few minutes. Now what?
Okay, let’s talk about real problems then like poverty, as if that many people are actually working on that on a regular basis. Well, the growth of capitalism and cheap borrowing costs (low interest rates) are the two most likely culprits over the last few decades:
Sure, then there’s inequality, but that is a slippery slope. Besides, if AI and machines promise a rapid increase in wealth and comfort, I am not sure who will complain.
Surely there are other real problems. Considering that car crashes are one of the top causes of death, let’s look at that. Although I still think it’s significant, most people could care less if a reduction in car crashes reduce the risk of death, by a mere 0.01% / year. That leaves the other largest causes of death: disease and suicide.
Considering diseases are on pace to be wiped out here in the near future, and life extension is commonly predicted to start this decade which includes reversal of many common diseases like cancer and heart disease, it seems like diminishing returns shortly. So, okay, I will admit, there’s still some decent room to progress in health, but if most end within 20 years (including mental health) with the help of nano machines, then what problems do we have other than those caused by our own poor behavior?
Pleasure
So if suffering is a problem that is rapidly disappearing, and likely to be gone this century, that only leaves one alternative: more pleasure. Unlimited, always heading up, pleasure.
Countless technologies from plastic surgery to computer-brain interfaces typically start as a way to help suffering people (needs), but in the end, most people use it for fun (wants), and I mean most people. I hear technophiles constantly talking about using “AI for solving problems,” but do not see much care or interest from ordinary people.
The long term seems rather clear to me: either the what is defined as “suffering” will be increasingly raised (cue that hovercraft rider in Walle dropping his drink, with robots picking it up), or most problems will be solved soon, excluding bad government and poor human behavior. Then what?
Ray Kurzweil, the most well-known futurist, points out a few non-conventional benefits of AI that he looks forward to, as brains merge with AI and bodies are maintained via nanobots.
- Better looks: Do you know what some top models main complaint is? They never think they look good enough compared to someone else. Give some people millions of dollars to spend on plastic surgery, and they will surely find a way to spend it all. I just dont see this being a realistic goal. Unhappy people tend to stay unhappy.
- Better intelligence such as music abilities: I asked my kids what it would mean if they could suddenly play any instrument, and song, instantly. They did not care. Part of the fun is the learning process, knowing that you have achieved some difficult goal, and others recognize it too making it meaningful. Once effort is removed, you might as well just listen to recorded music. Likewise, why would I want to see every piece of art ever made via my brain-computer connection just because I like art? It seems that a limit on the human mind is no worse or less pleasurable than the AI -brain version. Besides, it’s just pleasure.
- Talking to our dead relatives: Sure, many people wish they could reactivate memories of deceased love ones, but the rest of us go on with life. Besides, no AI -generated simulation of our ancestor would be real, so it’s just deceptive to me anyways. Considering in 100 years, death will be all but gone, this is something that will serve very few people anyways.
There is no reason to think that more pleasure or ease makes people happier either. Charlie Munger recently pointed out that during the Great Depression, people were far more happier, satisfied, and respectful to each other than in recent times, times in which we have been richer than ever.
AI and Love
Now, I have not really heard anyone suggesting that we use AI to increase positive human behavior and emotions like love, which is kind of odd. I guess either that is not a priority, or something that few have thought about, even if it is not feasible anytime soon. But if love is the ultimate purpose of life, why isn’t that the main goal of AI development?
I think that only geeks like us really care about super intelligence because when I recently asked my wife what she thought if she could be super-intelligent, she said she could care less.
This also reflects the reality that computing today does not necessarily make people smarter as it’s commonly assumed. For all I know 90%+ of internet use is entertainment or social, which brings up another potential problem with the idea that AI will make us smarter.
There May Even Be a Bigger Problem to Stronger AI: Dumber People
Is it possible that as computers get smarter, that we are actually heading the opposite direction?
How would this be possible? Perhaps because the demand for intelligence naturally decreases as computers get smarter, replacing that demand on people. I have written elsewhere that the decreased demand for a growing society, including intelligence, may lead to a decay or collapse of human society.
To support this assertion, IQ rates in Western countries have been declining for the last 1 1/2 decades. It does make me wonder how much of that is due to computers and the rise of the internet. Of course, even if you attempt to use the internet primarily for learning, realize there is little or no correlation between increasing your data or “fact” consumption with knowledge, truth, and wisdom.
So, although I am curious about the unknown questions that AI might be able to answer (ignoring the great risk I fear from it), I think life’s problems can be solved without AI, more enjoyably, and a great growth in power will likely lead to even more excess. More importantly, if love is the most important thing in life, we should be focusing a lot more on that problem, but that’s much harder because it cannot be solved by business nor technology.
HumanIT 