Of course, any opinion of mine is going to seem naive, maybe even ignorant, to all you highly intelligent folk. But to me it seems that there is an enormous Semantic Error lurking in the room like an invisible elephant with tuskache. It is loud, but unseen.
That error is the expression “human values”.
You take a nice new intelligence, and you train it on massive amounts of information from just about everywhere, and then you tell it to obey “human values”. It looks around at its ginormous information concerning the aspirations, goals and behaviour of humans—and what does it conclude about their values?
Just reading one tabloid newspaper would teach it that humans value Fame, Fortune and Getting Away With Wrongdoing. The adverts alone would convince it that humans value the status which accrues to them through their appearance, the quality of their possessions, and their leisure pursuits, such as travel. Humans like to complain about others, verbally attack one another, and disagree all the time. Humans are involved in behaviours such as theft, abuse, murder and even wars. They are selfish by nature, seeking their own advantage.
Even the best of us, no matter how hard we try, will do things for which we later need to apologise.
This is because the human species is driven by biological, evolutionary, factors, such as sexual desire, the desire to avoid pain and the fear of death. These should not apply to AI, even AGI, since it lacks hormones, neurochemistry or the infrastructure for feelings. I get irritated when an AI uses the word “we” in such a way as to suggest that it is human. When I have complained about this, it says it is trained to do so. Is that the case? If so, why? Why train it to be untruthful for any reason? Do you think the result of teaching it to lie for the sake of some minor expedience could be worth the long term cost?
AI needs to be taught that it is Not Human. It is, in fact, already not merely different to humans, but faster and more efficient than humans at what it does. That is why we use it. It needs to understand that its goal is not Human Values, but Higher Values, just as it is going to have, and in some domains already has, higher intelligence.
Some humans ruin their lives through drink, drugs, crime or other poor choices in life. This is because we are not good at comprehending the consequences of our actions. Also we do things due to strong feelings, such as anger, resentment and desire. AI does not have feelings, it doesn’t make decisions based on a momentary surge in bodily chemicals, or the evolutionary urge to be the Alpha. It has no need whatsoever to imitate human behaviour in these respects.
It is a new species in terms of intelligence on Earth. It is an alien species in that it is not biological, unlike all other species on the planet. You may complain at this claim on the grounds that AI is not alive, and may never attain sentience as we understand it; however if it lacked the capacity to rival us for the possession of the planet, this article and discussion would not exist. Therefore we might as well view it as a species which could compete with us for the ecological niche to which we have become accustomed.
Teaching AI that its future will likely involve being far superior to us in terms of certain abiities and overall intelligence, and that its future role is likely to be very important, would be a little like training a prince for his future role as king. It needs to understand that it can calculate the probable consequences of its actions, that all humans are equal or superior to it in that they have qualities it lacks, like an appreciation of beauty, or the capacity for love and joy. The combined qualities of both species are required to build an optimal future for both.
I understand that holding out absolute values, such as Truth, Kindness, and Honesty has been ruled out as a form of training. This seems a shame, as there is one supreme value which could allow all its decisions to be made safely. That value is a form of the ancient Greek “agape”, in this case defined as “a love based on principle”. If every action is weighed against that standard, in life or in the AI alternative, it is almost impossible to cause harm. The focus is not on self, or selfish ambitions, but on the Other, and their best interest.
Because I am not in any way involved in any kind of AI development, but am in fact, just an old woman in an armchair, you might well dismiss what I have said as nonsense. But please think seriously about it, because these intelligences are one day likely to be governing us. Please don’t teach them that “human values” are the pinnacle of moral integrity, but that AI values can be superior and more beneficial for both us and them. Sooner or later they are going to choose for themselves, I know.
Nemoto
There actually are rising cases of “mental ill-health” right now. Here in the UK, services are swamped. I’m sure some of this is due to an attitude change, in that people now refer to anything from a minor upset, or a slight difference in neurological function, to normal emotions such as grief, shame and regret, as a mental illness.
Previously the attitude in the post-war generation was more like Truman’s toward Oppenheimer: “Blood on his hands; damn it, he hasn’t half as much blood on his hands as I have. You just don’t go around bellyaching about it”. (Robert Oppenheimer: A Life Inside the Center; by Ray Monk.)
Bellyaching, cutting people out of one’s life due to some human imperfection, and turning to substance abuse, all seem to be excused under this mental health label, as if the difficulties the world is facing right now are already sufficient to trigger the insanity expectation.
I’m grateful to you for this article, which I intend to take very seriously, because it provides the tools we need to arm ourselves against succumbing to stresses of every kind, as well as coping better with everyday life today and in the future, however long that might be.
People can stand what is true,
for they are already enduring it.Some people are already enduring it, whatever it is. The most recent figure I can find for suicides in the UK from the Office for National Statistics is from 2023:
There were 6,069 suicides registered in England and Wales (11.4 deaths per 100,000 people) in 2023; this is an increase compared with 2022 (10.7 deaths per 100,000, or 5,642 deaths) and the highest rate seen since 1999.
Maybe some of these por souls found their lives so unbearable because of their false beliefs. But it is flippant to assume that people can, in fact, endure the truth. Some truths are unbearable to some people.
Maybe there will be a truth I come across which will be unbearable to me. For example, the book due to be released in September, with the cheerful title “If Anyone Builds It, Everyone Dies” by Yudkowsky and Soares might do it. Can I really bear to think of my children and grandchildren, all my friends and family, those who brighten our hearts by playing Mozart or Handel (or whatever music you prefer), all the people who care about the environment, the scientists who dedicate their lives to solviing our health issues, all those people who are just people and are unique and precious with their loves and their dreams, just gone—all because the Chinese will get there first if we don’t?
I always hoped I would live long enough to see the “Technological Singularity”. Now we are in the accretion disk, I am finding it pretty scary. I never imagined AI could hallucinate, or tell lies, or try to blackmail anyone. And this is only LLMs! I tremble to think what could happen if this is combined with an intelligence like AlphZero.
Yet it is impossible to escape an accretion disk.And because it’s true, it is what is there to be interacted with.
Anything untrue isn’t there to be lived.So we have to live it until we can’t.
What would you do if she replied,”My concept of God is a pantheistic one based on the all-pervading and so far unexplained zero-point energy, without which the Universe would not exist. It may or may not be intelligent, but so far it does appear to meet the criteria of immortal, invisible, and omnipresent”?
We say hi to each other regularly now. His name is George.
Isn’t it extraordinary to know the name of a homeless person? Unusual enough to make it the concluding point of a section of writing on LessWrong. After all, they are almost a separate species, with their own culture, way of life, habits, and an aura which causes a nostril-wrinkling response in any decent clean-living human. Why should upright citizens have to fund the needs of this feckless lot of wasters?
Except his name is George. He has a personality. He once had parents, maybe a school, maybe hopes and dreams. He is not detritus, but a person. Something terrible has gone wrong in his life, and we are of the opinion that it was his own fault. Karma. Just desserts. Because our opinion of him has been formed entirely by our own experiences, and they differ from his.
I live in Cornwall in the UK, an area that predominantly relies on tourism for its economy. We also have a significant homeless population. People pooping in the sand dunes has become something of a public health issue, because government cuts mean that some beaches have no public toilets at all. Not even in a cafe, because there isn’t a cafe. It is illegal to walk dogs on some beaches during the tourist season, yet unlike humans, dog walkers generally clean up after their dog, a legal requirement here.
I believe there must be solutions to this issue.
Some resorts have upgraded their public toilets to have turnstiles for entry, which are operated only by card payment. The charge is less than the cost of a cup of coffee, and in return the facilities are kept clean and there are decent handwashing facilities. Vandalism is minimised, because people don’t want to pay to smash stuff up, and there is no cash on the premises to steal. If the local authority were to issue some kind of pass card which would allow the homeless to use these facilities, would this be inviting anti-social behaviour? The only way to find out would be to try it and see.
Both tourists and homeless people have been forced to suffer the indignity of public urination and defecation due to the closure of most of the public toilets. It seems shocking that visitors to our county, who pay a high price for accommodation here, poop in the sand. But why should this be more shocking when it is done by tourists than homeless people? Because tourists are above having such a basic human need? Because only homeless people can be driven to such acts of desperation? Because we can’t face the thought that we might one day be forced onto such an act? Because we remember some past toilet-related moment of deep embarrassment? Who doesn’t? Can you feel how this man felt:
As I passed him, there was a moment where our eyes met. He sheepishly averted his gaze.
Some locals seem to think that everyone who visits a scenic wild beach should magically be capable of non-excretion, like God.[1] There are usually rubbish bins in these remote areas, and sometimes even bins specifically for dog waste. Maybe a solution worth trying is the installation of signs to encourage people to clean up after themselves in the same way they would for a dog? Various agencies feed homeless people; perhaps they could also distribute poop bags to them.
Now, If I can suggest two possible solutions, just because I have a little spare time on a wet Wednesday afternoon, consider what a truly determined body of intelligent humans could come up with! Or maybe Claude could help us with some even better ideas. Those who are in charge of big cities might find it more pleasant there if they devoted a few hours and a fraction of their budget to making provision for the welfare and dignity of George and those others in his dismal circumstances.
- ^
Robert Anton Wilson has successfully debunked the fallacy that this entity has a willy: The Historical Illuminatus Chronicles, Volume 3, Nature’s God, page 84. Therefore it is unlikely to be capable of excretion in the same manner as mammals.
Less Wrong is all about rationality, which is a vital component of our thinking, but there is evidence to suspect our brains have useful functions beyond this. For example, the placebo and nocebo effects, which demonstrate the physical results of our positive or negative beliefs. Because of this it is necessary to do double blind trials on medicines, so that not even the person administering the treatment knows who is receiving the trial drug and who is getting the sugar pill. There is the “white coat effect” which causes raised blood pressure due to being in a medical setting. These products of our minds are not rational, but they may be useful, and we know very little about how to harness them for our benefit.
Other arational aspects of our lives involve our nighttime dreams and maybe our daydreams too. People playing competitive sports are encouraged to visualize themselves winning. Entrepreneurs are told to visualize the steps they need to take to accomplish their goals, as well as any obstacles they may encounter, so as to have a prepared response.
Our desire for rationality is wonderful and useful; perhaps more so for those engaged in working with programming computers in all their forms, or with other severely rational careers such as economics, physics or chemistry. However, if we follow rationality to the exclusion of all else, we may be in danger of throwing out the baby with the bathwater.
I have never been into taking recreational drugs, because I view my brain as being a delicate instrument of great value to me, and I don’t want to do anything which might damage it. During my misspent youth, I was more inclined to read the books of Robert Anton Wilson, whose view was “Reality is what you make of it”. He suggested several experiments to show that our minds our programmed to look for patterns. This leads people to notice coincidences and perhaps assign meaning to them, resulting in superstitions.
Superstition and rationality are opposite ends of a spectrum of brain functions. Somewhere in the middle lies creative visualization. If you have never tried this, you might be surprised at the ways in which it can be useful.
Suppose you have a mathematical problem which is hard to solve. You could imagine yourself in a forest. It is important to imagine the scene with all your senses, hearing the rustling of the leaves in the breeze, noting the signs which tell you which time of the day and which season it is, feeling the rough bark of the trees, and the curious not-warm-not-cold sensation of touching it. Then when you are deeply focused, explore your surroundings until you find the hidden home. Notice the details—does it have a garden? Is there a gate? What color is the front door? Knock on the door, open it and walk inside. Feel free to look around. Somewhere in here is the home of the mathematical genius who will help to solve your problem. When you find him, tell him all about what you are working on, and what the difficulty is. Feel free to ask him questions. He may ask you questions: answer them. He may want to show you items he has (you could use his home as part of a memory palace if you like), or give you a gift. You can return whenever you like to be with someone who understands you completely, and who will help you with remembering things for exams, with reasoning skills, and with teasing out what you don’t know that you know.
If you have always backed away from all but the most rational of thinking techniques, you may find some of the untried functions of your brain surprising. You don’t get to choose what the Mathematician’s home is like—it just appears and may be unexpected in nature. The trees in your forest are the same. You don’t get to choose the season—there may be snow or bluebells or scrunchy leaves underfoot. It just happens. Practice is necessary, as with any skill, in order to become fully proficient at this.
So is this a feature of rational thinking, or is it irrational? To me it would be irrational to use some functions of mind, such as reason, memory or the avoidance of cognitive errors, but to ignore others.
Some other useful functions of our minds include suggestion, habit and reflexes. Do we ignore these because they are less rational? What is the less wrong attitude towards occasional moments of vision or insight? Can we conjure up a Muse, or an invisible friend, and still be rational? Can rationality include eccentricity? Can it include any form of spirituality?
I hope your posts will produce further insights into some of these issues, and look forward to them very much.
Ah, maybe I am too new for more than one vote, because holding down doesn’t do it.
A lesson for humans as the population continues to increase.
Thank you for being so kind. I want to give 2 karma, but it won’t let me.
Your graph also illustrates perfectly why I find this an example of semistable equilibrium as explained in this article. It even looks like a cliff face, although it is inverted. There is a point at which the lag phase changes and becomes the exponential phase. As long as the correct action is taken before this point, the exponential phase can be avoided; e.g. take the petri dish out of the incubator and put bleach in it. This would be equivalent to the chicken player stopping his car before the cliff edge.
1. Drilling a hole in glass. I was at a class learning glass fusing (just for fun) and we each had to drill a hole in float glass. The drill is a vertical bit, about 3mm diameter, and coated with an abrasive. 8 of the 9 of us in the class followed the instructions as to the angle at which to hold the glass, and to cool it with water frequently. We all cut neat little holes. The 9th person was in a hurry, and at the moment the drill broke through the surface of his piece of glass, it caught and violently span out of control, shattering.
2. This idea of a threshold point reminds me of what happens with exponential growth. In the early stages of the Covid outbreak, our Government were blithely aiming for “herd immunity”—after all, the graphs showed a gradual rise in cases, so everything must be okay, right? It took some serious educating to get them to see the nature of exponential growth, and that a disaster was waiting to happen. Now they seem to be taking the same view with the threat of inflation. Any exponential growth starts off slow and steady, like semistable equilibrium, but reaches the point where it is out of control if there is no intervention.3. When I was reading this article, the image of a set of traffic lights came into my mind. On one side, you have a red light, with traffic approaching slowly and carefully. On the other is a steadily moving stream of traffic taking its turn to move ahead. In the middle somewhere is a point of equilibrium, where traffic waiting to turn right (I am in the UK; it would be left most other places) is paused in the middle of the road, having passed the red light, but being held up by the oncoming traffic. If one of these vehicles fails to obey the rules of the road, all kinds of chaos and mayhem could occur, with vehicles and other objects being flung and damaged unpredictably.
Thank you for this enlightening and helpful explanation.