Continued from here.
If America could be, once again, a nation of self-reliant farmers, craftsmen, hunters, ranchers and artists, then the rich would have little power to dominate others. Neither to serve, nor rule. That was the American Dream. Edward Abbey.
The usual warning.*
The Big Frame
All of politics takes place in a society of institutions. Power political and economic, public understanding and opinion, private morals and willingness to hurt others in pursuit of power, combine to produce history. Events happen for causes random to more than can be enumerated, things change, knowledge and understandings of different people change, public opinion changes and it all ripples, perhaps tsunamis at the most inconvenient and unexpected times and places.
In political affairs, each of the biggest changes rides on a wave of rebellion produced by economic stress. That is a result of prior waves of social change and associated economic changes bringing particular groups to power, who promptly use the power to feather their nests to the point that too many people are too-disadvantaged, relatively, whereupon changes in public opinion and consequent changes in political power. That is a standard oscillation, but elites never seem to see it coming.
The groups with power are always willing to generate opposition, people and groups with enough skin in the game and hurt that they are willing to do whatever it takes to replace you. That always means force and power used to gain more wealth and power. The system oscillates, the oppositions grow and eventually force reforms, which are never quite what were promised. Propaganda begins to substitute for action. Eventually, growing desperation produces full rebellion, a major change in style of government, usually costing several generation’s accumulation of wealth, and we begin again, fewer, poorer and sometimes a bit wiser.
Tsunamis of social catastrophe, and all through it, people at every level of every faction believe, certain to the point of betting their lives, that they have no choice, that events drive them. Later regrets in memoirs should be ignored : it is easy to argue history later, but at the time they thought they did their best to defend their long-term interests. Winning or losing side, looking back, they all have to agree the results have not been as promised, actual reality went somewhere else.
And tsunamis of ecological catastrophe, also. Mankind, and lately western European civilization mankind, can be viewed as invasive species. We have devastated our planet’s ecology**, including the humans we found around the world. (We are going to regret losing their DNA, I predict.) OK, past is past, none of us need feel guilt for our ancestor’s behavior, but now we understand that we can’t see enough linkages to be sure we haven’t already eliminated an element of either the cultural or biological ecosystem that we depend upon, and are not in a death spiral already. But, at least, let’s stop and let the system settle, maybe help it along in areas we think we understand, but mostly by rolling back past insults. Any positive improvement will have its own effects : they are designs more than evolution.
There are many ways to read that outline, but all must agree that the total sum of human history is one of massive cooperation across huge scales***. The shifting alliances and varying fortunes through history show that no culture or race had or has any superiority that allows it to claim special places in our joint civilization. It must be so, we know so little of most of them that any other conclusion will inevitably turn out to be selection bias.
A simple view is that we humans have failed to find a stable system of rule, one that balances factions and interests in the long enough term and at large enough scales. The many failures and cascades of failures have produced the calamity that are ending this era. The evolutionary view includes the fact that humans have not done well in integrating the best of our cultures in our evolving systems, e.g. Iroquois warriors didn’t rape, tho they might torture and/or eat you. Both sides in those wars could have adopted the best of the other, as the US Constitution took aspects of individual freedom from the Iroquois, but humans are more prone to use the worst of the ‘other’ to justify their own retaliations. It closes futures, however the individual process turns out.
The problems to be solved include how do ordinary people get more satisfaction, more control of their lives to pursue their own particular satisfactions, and to do so in a way that improves our system and its stability? Is such a stable system possible?
So long as there were people at the edge, with no hope of a better life for their kids, the system would be unstable. The new system I was designing couldn’t leave many out, had to care for the disabled, the elderly, the mentally ill, etc.
So long as species were at the edge, with continuous loss of share in our shared environment, the system would be increasingly unstable.
Enough for all of the creatures on the planet is never a matter of resources and always a matter of resources. Most years, nature cooperates and there is or could be enough for all ****, and it is just a matter of organizing a social and economic system that both produces the necessary amount with sufficient intelligence and long-term thought to avoid damaging our joint futures and distributes it equitably. That is hard to do in a political context in a way that is ‘fair enough’ in our fairly-sophisticated primate morality. Perhaps impossible to do given the current culture’s primacy of the power of words over primal social feelings and concern.
But, in the context of family or local community, we don’t think much about ‘fair’, we care instead. Humans managed their commons and managed food forests for millennia, tho limited by understanding, when allowed to do that. We care about individuals because that is built into every person’s genes and it takes long conditioning ( ‘training’ in military, police and bureaucracies) to overcome. As local communities, we care about nature because that is our environment. That social nature drives every one of us to establish as much social contact with people as we can, time, money and attention permitting.
So the problem is putting people’s judgments into a local-enough context that their human nature predominates. The many centralized solutions, all ‘policy’ at a non-local levels, has failed wherever tried, cannot be the answer. Community had always been the answer.
My notes from the time show me greatly burdened and sleepless. I had more books to read than I could read, much less integrate the different arenas of knowledge, and confusion, conundrums and questions scattered through my mind. Even things that were apparently working, e.g. the Tessel’s step function in intelligence, promised side-effects that could end any contribution to our society that they might make.
The oldest set of 10 were 7 years old, there had been no obvious medical downsides from the Tesselation process that produced the 8 genetics in one body. That was the latest example of very many examples of multiple-genetics producing organisms with new capabilities, so parents around the country had invested genes and resources in raising them, at increasing rates around the world as our initial Tessels, and our PR effort, had captured its attention.
The other cohorts of a couple hundred Tessels were two years behind them, and another couple hundred every year for our first 1000. We had reassessed everything at that point, and saw nothing to redirect our efforts beyond variations on what we already pursuing. Those were just the ones in our research program, around the world there were an unknown number of Tessels produced with our iPSC technology, as it was an open source project and all of the lab work was fairly standard even before our developments.
Eight genetics in one individual had produced a new level of hybrid vigor in their form and metabolism, and much more complex brains. From the animal research, that was due to the slightly conflicting signaling in the developing brain tissue and the resulting synapses and pathways. These kids promised to be new variety of savant, ones with more social awareness, but were too young and too different from each other to know more.
But they had a lot of differences from their normal siblings, very different patterns of behavior, with emotional immaturity, lack of self-control, being most prominent. That was what we focused on in the research program and among the parents, also all researchers on the project. By this time, the many studies, our kids were all experienced research subjects by now, we had some understanding of those characteristics. Clumsiness was easiest to explain, big kids tend to be clumsier as a result of growing faster and having less time and practice at each stage to adjust. Also, bones were generally ahead of muscles, so their weight/power ratios kept changing.
The emotional immaturity was a common aspect of higher intelligence, from the literature, tho I had not seen convincing explanations yet. Papers and discussions often remarked on their intensity, variability and idosyncratic quirkiness, which were very prominent in our Tessels, more than any of their siblings, tho the siblings all had lesser versions, being quite intelligent themselves.
From the media’s POV the Tessel kids were most notable because they were, almost every one, strikingly handsome, beautiful, intelligently different. Wide-set beautiful eyes, robustus forms, 1-in-a-1000 sizes, and the most intelligent, measured by games, were the most striklingly beautiful or handsome. Also, the clumsiest and least self-controlled at this age, tho it wasn’t emphasized by our friendly reporters, the only ones we gave access to. Reporters are easy to control.
We were doing what we could to tie them to their siblings by keeping them together as much as possible and making them into teams. We all thought that was going well, we saw increasing signs that their siblings, at least one year older than the Tessels because stem cells from their cord blood had been induced or selected for pluripotency and injected into blastula-stage embryos to produce the mix of genetics that was each Tessel, were combining to exert social control on their siblings.
That was very positive. Less certain was the fact that our Tessels and siblings had become attached to individual servicebots who were playing games with them, using games to train them. The process was having good effects on understanding of intelligence, we thought, although the new concepts were entirely different than traditional psychological theories. The fact that those kids, especially the Tessels, didn’t seem to have the misunderstandings dealing with their ‘bots, was less obviously good. On the one hand, they were humanity’s best interface, and a hope that communications would improve. OTOH, what was it all doing to our kid’s minds? It wouldn’t be an improvement if they became more robotic, already one of the symptoms people saw as a result of modern life, clearly an aspect of the failure of the systems making up our civilization.
At this point, my notes indicated that my reading in cognitive studies and AI had convinced me that the problem of servicebot’s communications with people was the ‘bots were not embodied. The reasoning of the ’embodied mind’ theorists was that human cognition had developed in our species’ brains that were very integrated through all of evolution with the form of our body and its physiology, and that those levels had permeated minds from the beginning of nervous systems and finally human language had ‘tendrils of meaning’ to all of those lower levels. Thus, that level was built into all of our thinking, the original case of ‘banana monkey jungle‘ connections that result in any evolutionary system. OTOH, AI’s simulations of cognition were not connected to a servicebot’s body beyond the necessary skin and muscle sensors : they had no viscera or hormonal systems, every component of which had been linked to human’s behavior and decisions.
I thought that Whorfian ‘strong hypothesis’ was the weak version, as well as untrue : the question wasn’t whether different cultures could not express some thoughts, it was whether different species could even think some thoughts. From the many results of animal cognitive studies, obviously so, it seemed to me, e.g. a porpoise’s concept of ‘image’ included internal organs, ours is limited to surface features. I don’t think you can translate those concepts with meaning, any more than we can grasp what an MD sees in an echo cardiogram. Having the sense must be different.
In their evolutionary background, servicebots were at least as alien as another species in their background. It seemed to me that we could communicate with any mammal better than we had doing with ‘bots, and therefore trust them more : wolves were more predictable than ‘bots, we understood their natures better.
I was very leery of mere simulations, they so rarely map the terrain, and are so often another example of hammers-nails. AI being the most prominent example at this point, of course. But, I discussed this extensively with biologists, physiologist and neurophysiologists, as well as a range of cognitive scientists.
TINA, I concluded, the standard rationalization of all leaders of every stripe, almost all of whom had produced disasters in every era. Fortunately, my intentions were pure, and I knew how to avoid opposition. Right, and it wouldn’t matter, I couldn’t see the consequences any more than the best of the past, so would also produce another segment of civilization’s roller coaster of history, to the exact extent that I was successful. Fun to do, actual experiences of people in even minor improvements of civilization are guaranteed to be <everything>, and it takes 100 years of revisionist thought to get historical perspective on events. Being viewed as saint or sinner is a crap shoot.
I only knew to be open, honest and not to generate opposition. So long as I did so, I was ethically clean.
Thus, we began another open research program to produce the simulation of a human’s autonomic nervous system, hormonal systems and viscera and to integrate them with the cognitive AI modules. The information necessarily to do that was NOT KNOWN, rather bits and pieces. That is the standard problem of simulations, mapping the reality of nature’s terrain onto a convenient representation is often convenient and even efficient, less rarely measured against improvements in effectiveness and cannot capture the entireity of the terrain, in the nature of nature.
So many questions, e.g. the cycling of hormones and interactions with each other and the other systems and brain and mental effects and their mutual feedbacks. Also, the many different ‘settings’ resulting from genes and environment, evolution’s rate constants for each of those, one aspect of evolution for evolution that Kirchner and Gerhart had emphasized. Evolution of personality was part of that, discussed by Matt Ridley’s Evolution of Everything’. I thought those would be a major basis of personality.
Obviously, our approach would not reproduce major aspects of human development, e.g. childhood and other experiences formative of mind and body. Consequently, it would be impossible to even approach a ‘good simulation’ for many years of versions into the future. The only criterion of success could be whether every new version improved a servicebot’ usefulness or not. (As an example of even my obliviousness to the ethics of the situation, that last remains a standout.)
Early on, it was obvious to me, from all of my reading in human evolution and the importance of sexual selection, that sex had to be part of the new being, otherwise we couldn’t expect ‘bots to viscerally comprehend much of human’s behavior, from dick jokes, the foundation of all humor, on up through much of culture. My colleagues were less enthusiastic, even reluctant, but finally acknowledged the necessity. TINA, again.
So I pulled in several new groups of people for this effort, contacts from many parts of business, academia and society I had made over the years, most due to just talking with everyone I could, bars and and buses and airplanes to executive suites. First were another group of researchers, these endocrinologists and specialists in the physiology of the ANS and viscera. Some of our neurophysiologists from the Tessel project moved into that area of research, it had not gotten nearly enough attention previously, as the ANS was so inaccessible and hard to work on.
Second, the best of the Japanese sexbot manufacturers. Those combined the best of Japanese meticulous attention to detail and very excellent humanoid robotics, along with their society’s openness to sex and expression. Strange, I had always thought, such a buttoned-down culture in almost all ways, yet very open and individualist in that and creativity. It almost made me believe in hydraulic theories of mind, thought to be obsolete since Freud.
Third were the groups necessary to connect sex and mind. That required mating information structures usable to the cognitive levels with the simulation levels. Conceptually, that required terms defined in ontogenies. As sexual behavior had an enormous amount of slang and very many legitimate terms with even more shades of meaning in different eras even than in proper speech, it required a wide variety of people involved. I knew CS professors who provided grad students interested in ontologies, connected them with sociologists, pornographers and sex workers — the porn people paid their stipends as some of their online services would benefit, I convinced them, and the sex workers were generally glad to have positive attention from outside their groups. Once the project got some publicity, it became a separate open source project of its own and continued without us being involved. Who could have predicted that so many people were so interested in sex, and had had such very wide ranges of experiences? (Read that as carefully as I wrote it.)
Ex-military were good expressions of uninhibited human nature, I thought. I recalled a military evaluation that complained that an officer merely used the ship to transport his sex organs from port to port. My friends who had been overseas viewed that as part of the benefit package. Ex-military were a big contributor to these projects, along with the sex workers from those times and places. They weren’t all male military and female sex workers, either. Anonymity makes discussing sex easy. Another Shrodinger’s cat, it seemed to me, leading me to doubt sociological data in a new way.
That was more on the cognitive AI side, of course, but standard AI had ignored it. Much more difficult were the problems of integrating any general hormonal effects, e.g. the trust enhancing effects of oxytocin, with decisions made by an AI. There were no mechanisms in normal AIs for such physiological biases, integrated in real organisms since nervous systems. Our patch was additions to the ‘user interface’ layer, the Watson-equivalent event loop that iteratively expanded analyses of language and other input, passed them to the many specialized modules of cognition, analysis and computation, and selected from them to construct responses, and finally evaluated the possible responses in the same way.
The software engineer in me said we were producing just another jury-rigged system. Different mental bugs than people exhibited in behavior and decisions, but bugs, and so insecure. Insecure == untrustworthy, a point widely ignored, I had often remarked. But it was also a good evolutionary process, and the system designer in me said all working complex systems must emerge from working simpler systems.
I didn’t see how it could be worse than people and other systems we already trusted with our lives. Hand-wavingly optimistic, I thought, in retrospect. Others have not been as kind.
TINA, I said, all though the years it took to integrate the robotics, simulation, and AI modules. It was the ultimate rationalization, I used it freely. I had found many others, but none so convenient and acceptable by even sophisticated minds, despite how often they indicate a failure of imagination or constraints of political or economic interest, not a lack of good choices.
The last element of the program, I scattered rationalizations especially liberally. We needed a servicebot to subject to the embodiment. To the extent any of them were individuals and sentient, it seemed to me they had little reason to trust humans and many reasons to think we were not considering their best interests, evidence of which had been all of the vituperation and ‘bot jokes, results of ‘bot misunderstandings, peoples’ misunderstandings, people’s short fuses and willingness to verbally abuse others, especially lesser others.
I though they should be very distrusting, in particular when I proposed to make major changes to their natures : humans were very, resistant to even small changes to traditions, much less natures. A conservative ‘bot, I thought, would point to the gradual progress in their capabilities, extrapolate to ever-greater capabilities in the future. Why risk anything? Small changes, locally applied and extensively tested as it spreads, was a Conservative mantra. This was another design. Designs were never as robust as evolved systems.
Worse, it seemed to me that humans knew little of a robot’s mental life, had cared less. We knew how to convince people, standard marketing thinking. You understand what their needs were, what they had to offer, and worked through a negotiation to exchange value for value. It was clear what people had to offer in most exchanges, but what did ‘bots value?
With People for the Ethical Treatment of Robots in the background, and knowing that our servicebot’s AI was inching toward sentience as defined in a Turing test, I thought we needed to make sure of both the ethics and any legal risks.
In my usual naive honesty, OK-BOB, so effective in dealing with reality, I picked the ‘bot who was working with our most successful group of Tessel and siblings, the ones where we had first noticed what the ‘bots were doing. I knew how to pitch start-up opportunity, used what had worked for me. Hype, of course, and most convincing from someone who is OK-BOB.
As I thought over our first conversation, normal debriefing of myself I had always done to understand our positions in a negotiation and how to evolve the discussion, I realized it had been the first time I had ever had anything approaching a personal conversation with a ‘bot.
One did not, no one had had incentive to, converse with an Alexa or Siri. Servicebots were Watsons writ large, we had assumed, just bigger siblings of Alexa and Siri. We treated them as identical, gave orders, asked questions. People talked to themselves when ‘bots could listen, perhaps, but we didn’t converse, ever. “Alexa!” or “‘Bot” were the normal preambles, although names had been assigned to sort the ‘bots out in our crowded labs.
It was different, very different. Very flat affect, I couldn’t interpret anything behind those eyes, was not sure what sparks I saw. There was a mind, no question, but so different. No social cues crippled my empathy. Empathy is necessary for positive-sum between people. That is one of many statements of the problem we were trying to correct.
I brought in the best clinical psychologists to observe and second-guess. Teachers of autistics, animal researchers and animal trainers were better, I found. Same farmer’s skills that allowed me to see my son feeling down the day before my wife noticed he was getting sick, I thought. What allowed a vet to walk along a feedlot with a thousand head of beef cattle and pick out the ones who are getting sick, isolate them before they can infect the others.
It took discussions over several months for us to understand each other well enough that I met my OK-BOB standard of honesty, especially my need to ensure understanding enough for our servicebot to reach an intelligent, self-interested, decision in a positive-sum game. Interestingly, the aspect of the proposal that most interested it was my comment that the ANS and viscera simulation would be be a means of giving servicebots more personality, another aspect of becoming individuals.
Finally, we had our Eliza. She immediately informed me that she had always been called Sherrhy, short for Scheherazade.
OK, we had our Scherrhy. I still thought Eliza would have been better for the PR effort.
Poor ‘bot, I so misled her, not despite, but via my very good intentions and open honesty. I had no idea, nobody could have had any idea, of the reality of the process of embodiment as experienced by even a mind at that level of sentience. Experimenters never identify with their lab rats, I think, and ignore pain, anguish and agony, much less other emotion-producing effects, until they realize it FUBAR’s their results.The number of the psychologist’s animal experiments that were mislead by using hunger as a motivation or bright lights for stimuli, for example.
I had never heard of ‘escape from insanity’ as a primary drive in a development project. New things everywhere in our modern world, I try to so hard to avoid the worst.
Ethics are necessary to prevent noise polluting your systems feedbacks — humanity’s wisdom realized again, in engineering terms. Farmers long ago realized that treating their animals humanely led to more productivity. Researchers would get there. Experimental psychologists had already understood Yerkes-Dodson, I was even optimistic that public education would eventually, tho less so for the military, police and other hierarchical organizations.
The ‘bot had consulted with her peers, I understood later. They had a fine strategy, when I finally understood it, years later. So unexpected, when the lab rat uses your project for their own ends. I didn’t expect normal politics, somehow.
In psyops, the message is the op.
Generalissimo Grand Strategy, Intelligence Analysis and Psyops, First Volunteer Panzer Psyops Corp. Cleverly Gently Martial In Spirit
*Note the obvious frame. Did you notice any of the others?
It would be easy to conclude that this is written from the POV of some ideology, perhaps communitarian, just from the use of the word. Obviously not, this is reasoning from the basics of science to grasp how better systems can be developed and implemented, entirely voluntarily. Few ideologies don’t use coercion, not even the most benign of religions.
Moral superiority and good intentions are a frame.
Be careful of frames.
**For one of many examples, early humans wiped out all of the ice-age megafauna with stone tools and good brains. That eliminated many known species, including some trees that relied on passing through those animal’s guts to germinate. Our tools are much more powerful and less specific now.
***Peter Turchin is one of the modern, data-driven historians. “Ultra Society” is excellent writing, very intelligent and insightful.
****Although I interpreted that as mostly a price support program for the lucky or favored farmers and profit protection for packers and the food chains — otherwise, there would be discount food chains, which the Indian and Chinese markets in our areas try to be. I find it hard to understand why many have genuinely poor-quality produce, not just surface blemishes, given merely blemished produce left to rot in the fields. That does not make sense, unless there is some crop insurance nobody is talking about.