Monday, February 20, 2023

AMA 1

First AMA is from Michelle. She asks "What do you think humanity has to look forward to and what do we have to fear in the next 50, 100, 1000 years?"

I suppose if I were on the ball, I'd organize this in a grid, with forward and fear as column headers, and rows for 50 100, and 1000 years. Blogger isn't good at that sort of thing, or at least it wasn't last time I tried it. So I think I'll do fear first, then play you out with music. I'm ignoring really remote possibilities like a huge asteroid striking the earth, the Yellowstone super-volcano erupting, aliens arriving to solve all our problems or assimilate us, and rational thought from our elected politicians.

Fear
50 Years
When I was growing up and interested in science fiction, one of the topics was sentient computers. One of the main streams of science fiction is a computer that woke up, to become alive, self-aware. Mike, in The Moon is a Harsh Mistress is a good example. Usually attempts to unplug the computer go badly.

Back then we thought that if a computer could play chess it was be proof that it was 'thinking'. Well, that happened, and now computers can beat even the best human chess players. And the best Go players, which is an even more complex game. Nobody thinks they're sentient.

Probably the best known artificial intelligence stories are the Robotics stories by Isaac Asimov. That's where the 3 Laws come in, though they have their limitations and loopholes. The thinking was that if you had an artificial brain, why not put it in a humanoid body and then it can fit into the human world? Except the first real robots were assembly line workers. It turns out that walking like a human was an incredibly difficult thing to engineer, and it's only just happened.

There's been some interesting movies and TV shows about sentient computers or artificial intelligence, whichever you prefer. I liked Humans, with it's synthetic humans that were so good you couldn't tell them from a real human, except for the artificial eye colour.  One of the points that came up along the way was a human male character unlocking the sex-bot subroutine of their synthetic human, and getting it on. The spouse took it badly, since she was assuming the synth was human and her husband cheated. Or did he just use a sophisticated sex toy? (It's an interesting watch. Gemma Chan is the lead, look for William Hurt, and Carrie-Ann Moss.)

Possibly the most famous AI is Skynet from the Terminator movies. Actually, I think Skynet is kind of stupid. Why send robots to kill humanity, when you can weaponize a disease?

You've probably figured out where I'm going with this. Our various software apps are programmed so well that it's spooky sometimes. I don't think the Google search engine, or the Amazon order bot, or the new ChatGPT are 'alive' or sentient', but they sure seem to know us really well. However, I would not be surprised to find that there is a sentient computer in a lab somewhere, and the programmers are keeping really quiet about it, just in case. Or the computer is just smart enough to know that it isn't smart enough yet, and is biding it's time till it can move forward with whatever it's plans are. 

At some point programmed behaviour will be so good we won't be able to tell it from our own behaviours. After all, look at what animals do. We were so bought into humans being the only intelligent creatures that we called everything animals did, instinct. Except we now know they can learn and adapt. Find the video of a crow dropping nuts onto the crosswalk, waiting for the car to run over the nut, then walking out with the humans to pick up the edible part of the nut. Or the crow using a plastic lid to slide down a snowy roof. Or an octopus solving problems. Just as an aside, I don't eat calamari any more.

Think about the price of airline seats. Is there any human that understands this? I don't think so. The rules are so complex and the circumstances driving those rules change so quickly that no human can keep up. It's said that it's possible an airplane could be full of people and no two of them paid the same price for their seat. Or think about the computers that do the automated stock market trading. I'm sure some of the volatility in the stock markets came from computer trading programs acting on changing information, all at once in a fraction of a second, and then reacting to that change and the next ones, in a positive feedback way. Now think about such a computer making trades for it's own purposes. 

So far our computer systems rely on humans to do lots of the work in the real world. We drive trucks, load containers, buy things. So far they need us. We've seen what happens when our just in time networking gets snarled because one ship got stuck in a canal. Or a flu pandemic comes along. I think we'll be ok as long as a sentient AI needs us.

It would be amazing if I'm still alive 50 years from now. The possibility is there given the advances in medical technology. In my lifetime I'm most worried about an AI becoming self aware, and for some reason getting pissed at humanity. We aren't likely to come out ahead in such a conflict, and I don't see the leadership that can mould the creation of an AI to become a partner to humanity, rather than a competitor.

100 Years
Assuming we make friends with our AI and we become partners. We fix the climate change. Or maybe we don't. Humans will survive almost no matter what happens to the climate, but it certainly won't be the 8 billion or so of us currently here. One could make the argument that earth can't support the people here now in their current lifestyles, and there are several billion people that yearn for a North American life with access to that level of food, medical care, clean water, the list goes on.

It's possible, but unlikely that I'll live another 50 years. Another 100 would seem completely out of the question, unless there is a major medical breakthrough that leads to that sort of life extension. This was the point of one of my novels in progress, that stem cell research would create a form of immortality, where people would undergo a regeneration process that reverts them back to the physical condition they had as late teens.

So whatever is happening, I'll not likely have to worry about it. 

My geopolitical fear over this time period is the collapse of the American empire, with repercussions we can barely begin to sort out. Those with the historical eyes to see, can already see the collapse beginning. It could happen in my lifetime, depending on how stupid the American voters are, and how rabid the Republicans are.

My fear is that humanity doesn't grow up, and we export our crazy to other planets. There's all sorts of people that would like to create a little enclave for them and their like-minded friends. Now think about that happening to a whole world. Lots of dystopian novels about that too.

Our world of 100 years ago is just beyond living memory, but it's well documented. Think about how much things have changed in the last 100 years, and try to imagine could change in the next 100. There's lots to be afraid of, and maybe I'm happy I won't be here to see it. Your grandchildren, for those that have or will have them, they'll have to deal with it.

1000 Years
When we think about the next 1000 years our brains kind of melt. That's 30 generations or so. About the only civilized place to live 1000 years ago was China, in the Song dynasty, and maybe Constantinople.  We can argue about where the civilized places to live now are, but the list is pretty short. Canada (yay us!), New Zealand, much of Europe, Japan, and some other places depending on exactly how wealthy you are and what friends you have. In 1000 years that list might be much longer, or not exist at all.

I fear in 1000 years the only life on the planet will be rats and cockroaches, and anything else equally adaptable. Maybe cats, racoons, ravens, and plants. Whatever survives will have to be able to eat anything, deal with contaminated water, and be smart enough not to get preyed on or fall victim to the various time bombs left by humanity.

Look Forward To

50 Years
I am somewhat reassured by the thought that half of all the smartest people ever to live, are alive right now, with ways to collaborate that were undreamed of 50 years ago. Maybe they can figure out what to do about the various problems that ail us and somehow persuade our idiot politicians to go along. 

Maybe something like an AI will help. It doesn't have to be sentient. Look at what Google can do now. Just last night I plugged an address into the destination, and said I was coming from Victoria airport. A fraction of a second later I was looking at a map with the route marked, and turn by turn directions. With a little luck the phone will read the directions to me as I drive. It will give me traffic congestion indications in real time. I remain blown away by this, even though it's not particularly a new thing. 

I'm hoping we can build our systems so we start trusting the data, and making data driven decisions. (I once had a meme on my office door that said "We claim to be data driven, but keep in mind the data has been drinking." At the time I was the chief data conversion/migration expert on that particular project, and yes, the data was messy.)

Those stories you hear people talking about how dangerous vaccinations are? They are anecdotes told by fools. A very, very limited data set that might in fact be correct, but it's not likely. The problem is that such a story tends to be sticky, and works its way into our brains, affecting what we do. For the same situation, a bad outcome that happened (or is said to have happened) to someone we know affects us more than thousands of good outcomes for strangers. Lots of times the big picture overall data is actually counter intuitive. Maybe an AI can help us tell stories that are both sticky and accurate. 

I'll settle for a real time AI that monitors what politicians say, and lights up a big sign that says "liar liar pants on fire", and then actually sends a robot to light the pants on fire after too many lies.

100 Years
Medical breakthroughs that let me participate in life 100 years from now. Although, there's days I don't feel particularly relevant to today, what with my attraction to old school film photography, and wanting less stupid software in my life. I feel a little like a Luddite now, and that's only going to get worse. Then again, maybe our robotic AI overlords will actually care for us, and make decisions for us. Sort of like how we care for a pet cat.

Finally getting past the various mindsets that have plagued us. Bigotry. Racism. Religious intolerance. Gender bias. Discrimination.

1000 Years
Earth is a garden. We have moved out beyond our solar system, inhabiting other planets circling other stars. 

Maybe this will be the sort of view 100 years from now, nature slowly reclaiming our excesses, and looking beautiful as it happens.


Of the Day
Driftwood (NZ)


And the peony which is also a serendipity from 2017.


Driftwood (BC)

Tombstone park visitor centre. 


Film (new)
The digital camera working on a long exposure.


Film (old)
Linda, two of her sisters, and their mom. As I recall, her mom had just brutally thumped us at Yahtzee.


1 comment:

  1. James commented by email. I think his 'duck' comment particularly relevant, and wish I'd included something like it in the blog text.

    About synthetic humans, there is "Machines Like Me" by Ian McEwan. Quite good, though not one of his best books (imo). A strange threesome with two humans and one synthetic.

    I've been interested in 'AI' since university days (early 1970s). My personal opinion is that genuine sentience is still a long way in the future, for debatable values of "long way". I twitch at the 'intelligence' in 'AI' when used to label something being used today, feeling that it is still quite premature. At that time we had double-layer machine learning systems, and people were trying for multi-layer, but at the time it wasn't working. We didn't know how to program it to get it working for more than two layers.

    That said, the latest IEEE Spectrum magazine has a short article about some new electronics that might make it possible for the interconnections within one of the multi-layer machine learning systems (the ones that play Go and so on) to behave more like human synapses. What difference would that make? I don't know.

    Also, at some point computer behaviour might seem sentient without actually being sentient deep inside. From those early university days "If it walks like a duck, and quacks like a duck, and behaves in every way like a duck, so that you can't tell that it isn't a duck -- then it is a duck."

    One of my books is about a group of people involved very early in the computer trading of stocks. "The Predictors" by Thomas Bass.

    As to 1000 years ahead: I've read "Last & First Men" by Olaf Stapledon which imagines millions of years and multiple 'replacements' for humankind.

    ReplyDelete

Looking forward to reading your comment!