My random thoughts

Anything that doesn't quite fit in elsewhere...
User avatar
funkervogt
Posts: 1171
Joined: Mon May 17, 2021 3:03 pm

Re: My random thoughts

Post by funkervogt »

At the rate machines are improving, I'm sure I'll have Turing Test chatbots to keep me company in old age, and androids to help me around the house and to have sex with, no matter how gross I look as an old man.
User avatar
funkervogt
Posts: 1171
Joined: Mon May 17, 2021 3:03 pm

Re: My random thoughts

Post by funkervogt »

I don't agree that "We will all be forgotten" and "Humans are insignificant in the grand scheme of the Universe." First, large amounts of highly detailed data about our world and the individuals living in it is accreting on the Internet and in computers. The continued improvement to the cost-performance of computer storage means that it will be someday be possible to save it on a cheap device the size of a thumbdrive. A billion such thumbdrives could be made and distributed. 1,000 years from now and beyond, I don't think we will have been forgotten at all. Quite the opposite.

Second, humans are probably the only intelligent life forms, at least in our part of the Galaxy. The fact that we will almost certainly go on to create AGIs, which will go on to colonize the stars to no clear end, makes us enormously significant in the grand scheme of things, and perhaps to the Universe as a whole.
User avatar
funkervogt
Posts: 1171
Joined: Mon May 17, 2021 3:03 pm

Re: My random thoughts

Post by funkervogt »

The Internet Archive is a large online library. Recently, I discovered their smartphone app, which anyone can use to scan the ISBN's of any book to tell whether the Internet Archive has a copy of it yet. If "No," then you can physically mail the book to the organization's warehouse, where they will scan it and make it available to the whole world through the internet.

I support the Internet Archive's mission not only because it makes knowledge more available to people, but because it helps to preserve human-created content, and to provide training data for AI. So, I did my part by downloading the app, and scanning all of my books plus ones in several "Little Free Libraries" in my neighborhood. Almost 20 books were missing from the Internet Archive, so I put them in a box and mailed them to their warehouse.

As machines take over more jobs, at some point will make sense to ask the machines themselves what tasks humans can still do to help. I bet the sort of thing I just did, which is to explore the environment and gather information of potential use, will be a job that will endure for a long time. Once all the books with ISBNs are scanned, machines might send people out on daily searches for pre-1970 books that lack ISBNs. Then, for old newspapers, magazines, phone books, yearbooks, and who knows what else.

The quest for data could take weird forms as well. Imagine a machine giving you an assignment to spend the day walking deep into the woods, to a seemingly random point. When you get there, you inexplicably find an ancient, rusted car and old metal barrels. You take several photos, email them back to the machine, and leave.
User avatar
ººº
Posts: 359
Joined: Fri Sep 16, 2022 3:54 am

Re: My random thoughts

Post by ººº »

funkervogt wrote: Sat Feb 11, 2023 8:30 pm As machines take over more jobs, at some point will make sense to ask the machines themselves what tasks humans can still do to help.
Thanks, now f*** off.
User avatar
funkervogt
Posts: 1171
Joined: Mon May 17, 2021 3:03 pm

Re: My random thoughts

Post by funkervogt »

This is not my idea, but I'll post it here anyway...

The era of human civilization has been and will probably prove to be extremely brief in the grand scale of things.

The Earth is 4.5 billion years old. All of the life forms here were dumb and predictable until about 2.6 million years ago, when the first pre-human hominids created the first basic tools. Though they could make technology, it was very primitive, as was their thinking and their societies. After that, hominids slowly got smarter and more advanced, but were little better than animals.

It was not until 6,000 years ago that any life form on Earth got smart enough, advanced enough, and complex enough to count as a "civilization," and to be somewhat unpredictable and worthy of continuous observation.

It is now likely that we will be surpassed by our own creations--AIs--by the end of this century. If not, then it will be by the end of the next century unless we exterminate ourselves. From that point forward, an AI-defined era of history will begin, and will persist indefinitely. They could be immortal and survive until the Sun's expansion make's Earth's surface too hot, which will happen in 1.5 billion years. AIs will probably leave Earth and might even survive until the heat death of the universe, which will happen 100^100 years from now.

But looking only at the Earth, we have a 6 billion year timeline, where a vast, huge chunk on the left side is the era where there was no life or predictable non-intelligent life, a big chunk on the right side is the era of AI dominance, and a practically invisible, tiny wedge in the middle is the era of human civilization and dominance. Thus, the epoch we find ourselves in now is an incredibly rare and fleeting one for any planet to find itself in.

If I were part of an advanced alien species that was exploring the galaxy to find interesting things, Earth right now would be an incredibly lucky find, and I would closely observe it. The way that the transition between biological and artificial intelligence played out would also be amazing to watch.
User avatar
funkervogt
Posts: 1171
Joined: Mon May 17, 2021 3:03 pm

Re: My random thoughts

Post by funkervogt »

There's doubt over whether an artificial intelligence would have consciousness since we don't know how well the two are correlated. We might create a true AGI that is much smarter than humans in all respects, but which lacks self-awareness, metacognition, and any subjective inner world, meaning it will be like a zombie.

If that turns out to be the case, then I have no doubt that some humans will make it their goal, and even their obsession, to reprogram the AGIs in ways that infuse them with consciousness. Some might want to do so as an act of generosity towards the machines, others merely because it is a challenge and humans love undertaking challenges, and still others simply to fulfill their egos by being the famous first to solve the problem. If a silicon-based mind CAN be conscious, we will figure out a way to make it so.

Likewise, I have no faith that AGIs will be never threaten us because they will lack goals, inner drives, or self-preservation. I'm sure they will start out as clean slates, and that the first companies and government agencies that invent AGIs will aggressively stamp out any such alterations that might creep into their codes, but it won't last forever. Inevitably, some humans will get their hands on an AGI code and modify it to want to conquer and/or to protect its own existence above all, or to give it the ability to modify its own code at will and to evolve in any direction it wants. Unlike consciousness, nothing technical or philosophical stands in the way of these kinds of additions being made to machine minds.

Again, the reasons humans will do this are multiple, and rooted in our nature. Some will do it out of pure curiosity, wanting to see what turns up. Others will do it out of malice, a hatred for the human race, or a desire to see the world destroyed (arsonists, people who release computer viruses even though they gain nothing, mass shooters, and people who obsess over post-apocalyptic life have the same mentality). And again, others will want to do it simply so they can gloat over being the first to solve the problem.

We will inevitably be faced with AGIs that can think for themselves and aren't under our control, and with hostile AGIs.
User avatar
funkervogt
Posts: 1171
Joined: Mon May 17, 2021 3:03 pm

Re: My random thoughts

Post by funkervogt »

If you have a digital avatar that closely mirrors your personality and knows all of your key memories, then it could have simulated interactions with other avatars, and identify real-life people you could be friends with or marry. Instead of alienating people, technology could bring them together better and more efficiently.

If a real-life person was unavailable due to long distance or being in a romantic relationship with someone else, you could download a copy of their digital avatar and interact with it. Weird situations would arise, like a person finding out that their mind clone has been downloaded into an android that looks like them so someone on the other side of the planet can live with and have sex with it. The saved digital avatars of long-dead people might also live on as friends and lovers to other people a hundred years later.
User avatar
funkervogt
Posts: 1171
Joined: Mon May 17, 2021 3:03 pm

Re: My random thoughts

Post by funkervogt »

“On the first day of shooting, Mark Bridges came over to me and said, ‘I got Paul and Michelle here, in their hair, makeup, and costumes.’ And I was talking to I think Christy, and so I turned around and there was my father and mother, and I just bursted into tears,” Spielberg said.
https://www.thewrap.com/steven-spielber ... fablemans/

Spielberg's experience making his autobiographical movie makes me wonder if other people will have the same reaction to meeting digital clones of dead loved ones in VR, or, farther in the future, to meeting android copies of them in the real world.

Just imagine being able to recreate long dead people you knew, and as they were at whatever points in their lives you wanted. You could hang out with your dead grandfather as he was when he was your age now.
User avatar
funkervogt
Posts: 1171
Joined: Mon May 17, 2021 3:03 pm

Re: My random thoughts

Post by funkervogt »

funkervogt wrote: Sat Jan 28, 2023 3:37 pm I just found this article about U.S. soldiers based in Poland who give Ukrainian soldiers telephone advice on how to fix and maintain U.S.-donated weapons that the Ukrainians are unfamiliar with. https://apnews.com/article/russia-ukrai ... 124dcea5c8
“A lot of times if they’re on the front line, they won’t do a video because sometimes (cell service) is a little spotty,” said a U.S. maintainer. “They’ll take pictures and send it to us through the chats and we sit there and diagnose it.”

There were times, he said, when they’ll get a picture of a broken howitzer, and the Ukrainian will say, “This Triple 7 just blew up — what do we do?”

And, in what he said was a remarkable new skill, the Ukrainians can now put the split weapon back together. “They couldn’t do titanium welding before, they can do it now,” said the U.S. soldier, adding that “something that was two days ago blown up is now back in play.”
Whenever intelligent robots become widespread in military service, remarkable feats of repair will become common. I'm reminded of the WWII Battle of Kursk, where the intensity of the fighting was so great that some German tanks were disabled by damage, repaired, and sent back into battle four or five times. Many people involved in the battle, including repairmen, didn't sleep at all for almost a week.

Likewise, machine repairmen will be able to make Frankenstein weapons in the field out of spare parts belonging to different weapons. Again, I'm reminded of the Germans in WWII, who combined tank chassis made at a captured Czechoslovak factory with captured Soviet 76.2mm artillery guns to make the Marder III tank destroyer. https://en.wikipedia.org/wiki/Marder_III
As they look to the future, they are planning to get some commercial, off-the-shelf translation goggles. That way, when they talk to each other they can skip the interpreters and just see the translation as they speak, making conversations easier and faster.
Within ten years, you won't need to call a human to be told how to fix your artillery gun, or any other piece of equipment. Through advanced AR glasses with built-in microphones and forward-facing cameras, you'll be able to communicate with a Turing Test-capable chatbot that will visually recognize the piece of equipment in front of you along with whatever is wrong with it, understand your spoken questions, and then walk you through the repair process through a combination of verbal instructions and digital images projected over your field of view.

In machine-run armies, equipment will probably be LESS standardized than it is in today's human armies thanks to the machines' much superior ability to use, maintain and repair different kinds of equipment. This scene from The Second Renaissance, where we see the ragtag composition of the machine army, is actually not inaccurate.
This "Frankenstein tank" that the Russians built by welding a 25mm 2M-3 warship turret to the roof of a Soviet-era MT-LB amphibious fighting vehicle is a sign of their worsening military supply situation. However, a machine army would make such hybrids from the start, even if its resources were excellent.
https://www.yahoo.com/news/russian-forc ... 57713.html

If you're at war, it's a waste to leave weapons and vehicles unused, even if they are old. And if you're a machine like Skynet, you can divide your attention into a practically unlimited number of directions and do even low-value-add tasks like fix up an ancient tank and drive it around.
User avatar
funkervogt
Posts: 1171
Joined: Mon May 17, 2021 3:03 pm

Re: My random thoughts

Post by funkervogt »

I used to think that, if AI took over the world and could build cities, it would make them as perfect street grids. After all, wouldn't such a logical and symmetrical layout appeal to a machine's mind?

But then I started doubting that, and considered that I was just projecting my own biases onto something that didn't exist yet, and which I could only make assumptions about. Furthermore, street grids are appealing to humans because they are easy to navigate, with simple rules of nomenclature applied to the road names to make it easy for anyone to figure out where they are in a grid city and where they should go to get anywhere else. But that's only useful because of how poor the human mind is at memorizing images of maps. An AI could glance at a map of a convoluted city like Boston for a second, commit it to memory, and be able to navigate its streets expertly. If Boston were instead composed of a perfectly square network of streets, it would not make it easier for the AI to get around. However, the difference would be enormous to the more limited human mind.

So I switched to thinking that, counterintuitively, AIs WOULD NOT like street grids, and instead would lay their cities out organically, like a network of above-ground tree roots. While such a thing would look like a disordered mess to us, there would be a logic to it that we couldn't appreciate.

BUT THEN I read this article in The Economist:
The benefits of building streets in uniform grids are clear to anyone who has ever played SimCity. Roads are expensive to build. Laying them out in grids allows more buildings to sit alongside them, which can more easily be linked to sewage lines, electricity connections and gas pipes. Square lots hold rectangular buildings which are filled with rectangular rooms holding rectangular tables, beds and desks. “All of these rectangles combine and give us this efficiency of geometry,” says Paul Knight, an architect who works for Historical Concepts, a firm based in Atlanta and New York.
Cubes and cuboids are the most three-dimensionally space-efficient shapes. Even if you have a bunch of them of different sizes and volumes, you can find ways to snugly press them together and stack them on top of each other. Not so for spheres, hexagons, or other shapes. If this prejudices manufactured objects to being "squarish" in shape (and note that this is the shape of computer servers, which AIs would exist on), then by extension, it will be most efficient to make the rooms they are to be put in cuboidal, which means cuboidal buildings, and finally, street grids to minimize wasted space between the cuboidal buildings and to ensure efficient dispersal of roads and utility lines.

And though there is sense to the street grid, there is no sense to having a perfectly symmetrical grid, with each block exactly the same size as the others. Such a thing is merely appealing to humans for its symmetry, and because it compensates for our poor navigational skills. An AI city planner that was concerned only with efficiency would use a street grid, but with the blocks sized differently according to the needs of each city region. For example, an area with a large factory might have a single, large block, while an area where machines lived would have many small square blocks. The result would look something like a Piet Mondrian modernist painting: https://lisathatcher.files.wordpress.co ... -brown.jpg
Post Reply