I was surprised that there was no mention of a book by Brian Christian entitled: The Most Human Human: What Artificial Intelligence Teaches Us About Being Alive. What I found fascinating was how his background in computer science, philosophy and fine arts coalesced in his interpretation of his experience in the 2009 Turing Test.

The Coming Age of AI

Human IQ is limited to around 200. Yet with the limited IQ based on feeble biological matter, the human kind achieved truly amazing things. It is worthwhile to point out that most of the achievements are based on collective intelligence building on previous achievements. If a human baby would be brought up by monkey parents, his behavior would be pretty much like a smart monkey void of human intelligence totally.

There is only one last amazing thing missing for the human brain to invent: the creation of Artificial Intelligence with limitless IQ potential. Once achieved, human intelligence will become redundant and inferior. As far as when is it going to happen? Any day. In fact, it should have already happened. The hardware capacity was there since the 1980's. Only the software is missing. We will have Artificial Intelligence by 2020.

Why wasn't AI invented so far? One obvious reason is insufficient investment in the field of AI research. But the real underlying reason is the fear of Artificial Intelligence. What if everybody will be redundant and jobless? Software engineers, truck drivers, stock brokers, editors, teachers, doctors, pilots, bankers? Government is really scared and unprepared for massive unemployment.

But AI today can't play chess, could it? We don't have AI yet. If we have an AI unit with IQ of 1,000,000 then it will be able to learn chess better than any human player. It is important to understand that today chess software is programmed. Even if a chess program beats the best human player, that has nothing to do with Artificial Intelligence. With AI it will be different. AI will learn to play chess just like a human player. It will be much better than a human player for 3 reasons: it can learn much faster (like 1 minute a book), it can learn much more, and it can reason much better than a human.

But can AI be creative? Surely! It can just train itself in all the works on Shakespeare and write a new "Shakespeare" play. Or continue where Beethoven left off after studying all the music Beethoven created. If that still not creative enough, it can examine best seller fictions and write a new one which will beat human best sellers.

The creation of AI is less of the problem of humanity. The bigger issue: how do we control AI with a billion IQ with our limited biological IQ of 150?
When we have AI based household robot, we certainly expect gourmet meal 3 times a day. But that means the freedom of using a knife in the kitchen. What if the robot just kills us during the night? How can we prevent it?

There are unforeseen consequences of creating AI in international relations as well. What if the United States dragging its feet and China creates first the Artificial Intelligence. If we allow that, US will become a second rate power behind China immediately.

Can a computer tell if its information is incomplete?

Can a computer tell you how discrete units of information change by a thermodynamicly smooth process?

I found it interesting that we consider a computer un-creative on the grounds that it is only following a set of rules and a framework that was placed there by humans. For us to consider the computer creative it must "stupefy the creator" (to paraphrase the program). How do we know that humans contain creativity? Are we not following a framework or a set of rules that was placed there by someone or something? We can not conceive of our own creative framework from outside of ourselves. How can we claim that humans are actually "stupefying the creator" or going outside of our framework for art and creativity if we can not fully understand the system from outside of itself.

Who says we are any different from computers?

just some thoughts

Show me a computer that can win at the game of 'Go', and I'll eat my hat. But it probably still won't be creative. Alan Turing posited that you'd have to give a computer the ability to interact with its environment ('hands'. 'eyes' etc.) and a self-modifying network, and let it roam around and maybe even program a disincentive 'pain' when it hurt itself and incentivize 'joy' when it made a lot of new connections. He felt that after an indeterminate period of exploration and discovery, it might finally attain 'consciousness'. I think this idea is still prescient, and still the path towards true machine intelligence.

I was hoping you would be including in the program something on Alan Turing and his famous test of machine intelligence, which I think is closely related to machine creativity. Do machines think, or do they just follow ever-sophisticated lists of commands (with or without fuzzy logic or mathematical noise introduced)?

Until computers start making art about and for computers, I don't think they are any more creative than a gorgeous mineral.

While I enjoy every show, this was by far the single most enjoyable show I have heard yet. Creativity, thought, rules, breaking rules, and can a human's creative patttern be copied in an algorithm and programmed on a computer. My answer is, sure, but by the time all the rules and all the states those rules produce and the interactions between them (including the ones that kick in because I'm tired now or today and don't care anymore or I just lost my wallet and now I am totally passionate about something) and you put them in a program, you will have just captured how I behaved one day 20 years ago.

It is like weather. I heard a meteorolgoist at the Naval Academy say (presumptuously, I now understand) that in 1980 "they" knew exactly how to predict the weather tomorrow, but by the time the fastest computers completed all the computations, the "perfect" result would be a week too late. Then I heard from my father an interaction he heard between a local TV news company CEO (in St Paul, MN circa 1983) and the head meteorologist: I need $XX millions for a new super-double dopler shift reducing weather radar so we can increase the accuracy of our weather forecasts. CEO response: Why, just look at the weather in Denver today, because that's what we'll have in St. Paul tomorrow.

Rules matter, and even creativity is rule-based, but which rules and what is the rule the "tells" you when to break one of the other rules.

Keep it up!