“AI thus has the potential to be a God of Lies, spamming our already hopelessly contaminated information ecosystem with plausible-sounding falsehoods” 💯🔥 #godoflies
A singular issue facing sci-fi authors since the early days of cyberpunk, is that of how humanity responds to the environments presented by the near futures, whether they be economic, or social or cultural responses to technology, or by the impact of technology directly upon the individual human being. Gibson imagined this the most completely, in his Sprawl and Bridge trilogies, as he could imagine the complete hypothetical technological space and every day peoples responses to the massive forces of corporatism, AI, fragmenting culture and social values. Stephenson also did a wonderful job with Snowcrash and The Diamond Age and Cryptonomicon. All of which were written in the 1980s and 1990s. (I also give a shout out to Vernor Vinge's Realtime series as being especially prescient about the accelerating technological impact on humanity.)
That said, I think we here in 2023, can extrapolate ourselves a bit in the vein of those great masters of futurology, nee sci-fi, into 2073, as is your wont. The physical expiration of Moore's law in silicon will likely be replaced by more computation available due to the rapid application of advanced mathematics utilizing the underlying computational infrastructure. One only needs to read the 2017 paper on transformers published by Google [1] to realize how 20th century graph theory (along with matrix algebras) can be repurposed into novel means to enable LLMs in 2022.
Purporting to see into 2073, with computational biology augmented by ML techniques for protein folding, a NP hard problem, along with operational quantum computing, say by 2030, humanity collectively at this point in time has absolutely no idea where we could potentially be in 2073, 50 years out. It is similar to reading Tom Swift novels of the early 20th century, with its prognostications of the exploration of the solar system as simple solved problems due to the arrival of aviation, as if space flight is solved by machinery, while discounting the massive distances in space and the complexities of human psychology involved in something as simple as going to Mars.
Have you read David Brin's Kil'n People? The core technology in his world is a mechanism for imprinting a person's mind onto a golem, which can then go around doing stuff in the world for about 24 hours, after which it will begin to fail; if the person doesn't re-absorb the copy, the mind will lose coherence (and the golem body might be left stranded somewhere).
What brought this to mind is the fact that the golems don't have to be humanoid. Where in our world a crane operator on a skyscraper construction site is working joysticks and switches, in this world, they have golems that allow people to _become_ heavy machinery. Cop golems are much bigger and tougher than the people that inhabit them. Etc.
I’ll throw out a couple of ideas that may or may not have already been treated:
1) Explore a world where people, companies, and even nations rely on black-box AI that cannot explain its reasons. People grow used to following its sometimes bizarre advice. When they do take a step back, they twist themselves into knots like one who tries to justify a horoscope or theodicy. When an obvious tragedy occurs, people rationalize that it must have prevented an even greater one. Geopolitics devolves into people moving pieces of a game that cannot understand. In the end, this would explore themes of the relation of freedom to flourishing: would you want to make your own decisions when it almost always led to worse outcomes? But what sort of world would that be?
2) Remember how Covid repurposed supply chains and had unintended consequences such as used-car prices exploding because certain cheap microprocessors could not be procured and so leaving nearly complete new cars sitting on lots? What if AI could spot such weaknesses? In the early days, some would make killings in the stock market. Later, the need for protection would lead to massive redundancies and so massive inefficiencies. Perhaps it would inevitably lead to AI undermining itself, all to make a buck.
I’m no writer, so I’m not going to be developing any of these ideas. Then again, I’m no reader of science fiction, so I will not be surprised when people rush to tell me these ideas have already been done to death!
I recently enjoyed Alastair Reynolds' _Blue Remembered Earth_, which has ubiquitous surveillance as a well-done worldbuilding and plot-driving device, complete with an exceptional "Descrutinized Zone" on the Moon allowed to exist as an escape valve, with tight border controls. It did a good job of incorporating augmented reality into the plot, too. I always liked Reynolds' farther-future stuff and it was nice to see him shift so effectively to the nearer term.
Well, I'm serializing my near-future SF novel "The Replicant in the Refugee Camp" on my substack The Divergent Panorama, about a reality event ("Who Is the Replicant") where one of the 10 candidates is a 'replicant'--an android whose intelligence is run from the cloud--and who must imitate a human while the audience tries to find out which candidate, indeed, is the replicant. All candidates are followed by a drone, so that the audience can follow the event 24/7 (on the web, in apps, on TV (on demand), everyhwere).
To up the stakes, the event takes place in a refugee camp on a Greek island, where the candidates must find ways to improve the lives of the refugees. They take to that task, starting slowly, but gradually things get out of hand as successful crowdfunding allow for hi-tech implementations. At some point big tech companies intrduce beta equipment,using the refugee camp (now famous through the reality event) as a test ground. Then the refugees begin to take control, and the refugee camp gradually becomes the world in miniature and a blueprint for a better future.
You can read it for free, with an updat each week, and I would greatlywelcome comments. It's free!
I think instead of (or along with?) a future where it will be harder to tell truth from lies, we are likely to see a future where “reading” thoughts and images from brains via computer will be increasingly accurate. It is already in the early stages now.
It will be interesting to see how long various judicial systems hold out against this trend under some kind of “cognitive liberty” idea. I’m sure this has been written about over and over again though.
If you haven't read Linda Nagata's near future science fiction, it's worth a look. Last Good Man is a good one to start with, or Pacific Storm. Her Red Trilogy is also topical...
What sci fi always seems to get right -- particularly David Brin -- is class and power divides, which seem inevitable in sci fi. Tech offers new bread and circuses but still bread and circuses. Shiny cities still have slums too, suggesting tech hasn't solved NIMBY affordable housing. Sad!
How cool you are suggesting these ideas, Noah! So this would be "right now."
>>Universal government surveillance could have a far gentler and more subtle presence in our lives — a form of behavior control and thought control that goes unspoken and usually unnoticed.>>
Without multivariant reinforcement of 19th century norms and this type of government surveillance, it would be a much different world.
It's nice you read Vernor + David. I'm very sorry that what I wrote 20 years ago, is almost literally, completely true today.
This reminded me, how much I loved Ghost in the Shell manga. All of the events are happening after World War 4, Tokyo doesn't exist, Japan has new capital. Nuclear wars started, because of some technologies that ignores radiation. This text completely reinvigorated the excitement about near future sci-fi
Right now, I'm busy with a novel where a starship crewed by humans and AIs (50/50) on its way to Alpha Centauri, where a human/AI pair--Lupita Esperanza and Connoisseur--bide the time (lightspeed remains a limit) by researching the origin of consciousness (both human and machine). On the way, the crew finds out that humans and AIs need each other during the trip (for essential reasons), and when they arrive in Alpha Centauri, Lupita and Connoisseur's findings about the nature of consciousness are instrumental in avoiding a deadly end of First Contact.
Problem is, I find that the market (agents and publishers) are not very open to high concept, groundbreaking SF. Or my novel just sucks...;-)
Anyway, thanks for the interesting thoughts about how scifi might move forwards.
Suzanne Palmer's two Hugo-winning stories "The Secret Life Of Bots" and "Bots of the Lost Ark" probably needed to be mentioned although both are far-future and the first may suffer from terminal cuteness.
Drones as weapons - look no further than William Gibson's The Peripheral. The near-future component of the story uses drones heavily. Admittedly, more for surveillance, but still an active tool in the arena.
The Age of Lies - I can take you all the way back to 1985 for this one. Orson Scott Card, in Ender's Game. The Hegemon is just a kid, though a smart one who manipulates the politics of the day. But on the Internet, no one knows you're a dog.
“AI thus has the potential to be a God of Lies, spamming our already hopelessly contaminated information ecosystem with plausible-sounding falsehoods” 💯🔥 #godoflies
A singular issue facing sci-fi authors since the early days of cyberpunk, is that of how humanity responds to the environments presented by the near futures, whether they be economic, or social or cultural responses to technology, or by the impact of technology directly upon the individual human being. Gibson imagined this the most completely, in his Sprawl and Bridge trilogies, as he could imagine the complete hypothetical technological space and every day peoples responses to the massive forces of corporatism, AI, fragmenting culture and social values. Stephenson also did a wonderful job with Snowcrash and The Diamond Age and Cryptonomicon. All of which were written in the 1980s and 1990s. (I also give a shout out to Vernor Vinge's Realtime series as being especially prescient about the accelerating technological impact on humanity.)
That said, I think we here in 2023, can extrapolate ourselves a bit in the vein of those great masters of futurology, nee sci-fi, into 2073, as is your wont. The physical expiration of Moore's law in silicon will likely be replaced by more computation available due to the rapid application of advanced mathematics utilizing the underlying computational infrastructure. One only needs to read the 2017 paper on transformers published by Google [1] to realize how 20th century graph theory (along with matrix algebras) can be repurposed into novel means to enable LLMs in 2022.
Purporting to see into 2073, with computational biology augmented by ML techniques for protein folding, a NP hard problem, along with operational quantum computing, say by 2030, humanity collectively at this point in time has absolutely no idea where we could potentially be in 2073, 50 years out. It is similar to reading Tom Swift novels of the early 20th century, with its prognostications of the exploration of the solar system as simple solved problems due to the arrival of aviation, as if space flight is solved by machinery, while discounting the massive distances in space and the complexities of human psychology involved in something as simple as going to Mars.
See [1] https://arxiv.org/abs/1706.03762
Have you read David Brin's Kil'n People? The core technology in his world is a mechanism for imprinting a person's mind onto a golem, which can then go around doing stuff in the world for about 24 hours, after which it will begin to fail; if the person doesn't re-absorb the copy, the mind will lose coherence (and the golem body might be left stranded somewhere).
What brought this to mind is the fact that the golems don't have to be humanoid. Where in our world a crane operator on a skyscraper construction site is working joysticks and switches, in this world, they have golems that allow people to _become_ heavy machinery. Cop golems are much bigger and tougher than the people that inhabit them. Etc.
I’ll throw out a couple of ideas that may or may not have already been treated:
1) Explore a world where people, companies, and even nations rely on black-box AI that cannot explain its reasons. People grow used to following its sometimes bizarre advice. When they do take a step back, they twist themselves into knots like one who tries to justify a horoscope or theodicy. When an obvious tragedy occurs, people rationalize that it must have prevented an even greater one. Geopolitics devolves into people moving pieces of a game that cannot understand. In the end, this would explore themes of the relation of freedom to flourishing: would you want to make your own decisions when it almost always led to worse outcomes? But what sort of world would that be?
2) Remember how Covid repurposed supply chains and had unintended consequences such as used-car prices exploding because certain cheap microprocessors could not be procured and so leaving nearly complete new cars sitting on lots? What if AI could spot such weaknesses? In the early days, some would make killings in the stock market. Later, the need for protection would lead to massive redundancies and so massive inefficiencies. Perhaps it would inevitably lead to AI undermining itself, all to make a buck.
I’m no writer, so I’m not going to be developing any of these ideas. Then again, I’m no reader of science fiction, so I will not be surprised when people rush to tell me these ideas have already been done to death!
I recently enjoyed Alastair Reynolds' _Blue Remembered Earth_, which has ubiquitous surveillance as a well-done worldbuilding and plot-driving device, complete with an exceptional "Descrutinized Zone" on the Moon allowed to exist as an escape valve, with tight border controls. It did a good job of incorporating augmented reality into the plot, too. I always liked Reynolds' farther-future stuff and it was nice to see him shift so effectively to the nearer term.
Well, I'm serializing my near-future SF novel "The Replicant in the Refugee Camp" on my substack The Divergent Panorama, about a reality event ("Who Is the Replicant") where one of the 10 candidates is a 'replicant'--an android whose intelligence is run from the cloud--and who must imitate a human while the audience tries to find out which candidate, indeed, is the replicant. All candidates are followed by a drone, so that the audience can follow the event 24/7 (on the web, in apps, on TV (on demand), everyhwere).
To up the stakes, the event takes place in a refugee camp on a Greek island, where the candidates must find ways to improve the lives of the refugees. They take to that task, starting slowly, but gradually things get out of hand as successful crowdfunding allow for hi-tech implementations. At some point big tech companies intrduce beta equipment,using the refugee camp (now famous through the reality event) as a test ground. Then the refugees begin to take control, and the refugee camp gradually becomes the world in miniature and a blueprint for a better future.
You can read it for free, with an updat each week, and I would greatlywelcome comments. It's free!
Limited compute is part of the background to Permutation City, although it's mostly exploring the nature of simulation.
Cool ideas.
I think instead of (or along with?) a future where it will be harder to tell truth from lies, we are likely to see a future where “reading” thoughts and images from brains via computer will be increasingly accurate. It is already in the early stages now.
It will be interesting to see how long various judicial systems hold out against this trend under some kind of “cognitive liberty” idea. I’m sure this has been written about over and over again though.
If you haven't read Linda Nagata's near future science fiction, it's worth a look. Last Good Man is a good one to start with, or Pacific Storm. Her Red Trilogy is also topical...
What sci fi always seems to get right -- particularly David Brin -- is class and power divides, which seem inevitable in sci fi. Tech offers new bread and circuses but still bread and circuses. Shiny cities still have slums too, suggesting tech hasn't solved NIMBY affordable housing. Sad!
Very scary:
https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/
How cool you are suggesting these ideas, Noah! So this would be "right now."
>>Universal government surveillance could have a far gentler and more subtle presence in our lives — a form of behavior control and thought control that goes unspoken and usually unnoticed.>>
Without multivariant reinforcement of 19th century norms and this type of government surveillance, it would be a much different world.
It's nice you read Vernor + David. I'm very sorry that what I wrote 20 years ago, is almost literally, completely true today.
This reminded me, how much I loved Ghost in the Shell manga. All of the events are happening after World War 4, Tokyo doesn't exist, Japan has new capital. Nuclear wars started, because of some technologies that ignores radiation. This text completely reinvigorated the excitement about near future sci-fi
Right now, I'm busy with a novel where a starship crewed by humans and AIs (50/50) on its way to Alpha Centauri, where a human/AI pair--Lupita Esperanza and Connoisseur--bide the time (lightspeed remains a limit) by researching the origin of consciousness (both human and machine). On the way, the crew finds out that humans and AIs need each other during the trip (for essential reasons), and when they arrive in Alpha Centauri, Lupita and Connoisseur's findings about the nature of consciousness are instrumental in avoiding a deadly end of First Contact.
Problem is, I find that the market (agents and publishers) are not very open to high concept, groundbreaking SF. Or my novel just sucks...;-)
Anyway, thanks for the interesting thoughts about how scifi might move forwards.
Suzanne Palmer's two Hugo-winning stories "The Secret Life Of Bots" and "Bots of the Lost Ark" probably needed to be mentioned although both are far-future and the first may suffer from terminal cuteness.
A few other examples:
Drones as weapons - look no further than William Gibson's The Peripheral. The near-future component of the story uses drones heavily. Admittedly, more for surveillance, but still an active tool in the arena.
The Age of Lies - I can take you all the way back to 1985 for this one. Orson Scott Card, in Ender's Game. The Hegemon is just a kid, though a smart one who manipulates the politics of the day. But on the Internet, no one knows you're a dog.