Thank you for writing this. This has been on my mind the last year or so as I’ve studied the works of Elinor and Vincent Ostrom. They stood between a mainstream economics and political science that believed people unable to govern themselves. They (especially Elinor) thwarted the popular prisoner’s dilemma conclusion that only radical markets or top-down interventions could solve collective action problems. They believed deeply in the capacity of regular citizens to work together to solve problems.
Vincent’s work is well captured in his book “The Intellectual Crisis in American Public Administration.” For Elinor, read Governing the Commons and her American Economic Review article, Beyond Markets and States.
Even worse: creating something truly novel first requires novel ideas, and most of the new ideas I've seen in the last decade or two seem to be designed towards monopoly and wealth maximization for a few individuals.
Consider how much polarization and disharmony social media has injected into society, yet consider how much *genuine* innovation we see in that space. How many new user-facing features have we seen released by Facebook, Twitter, Reddit, etc in the last decade. Also notice how few people seem to notice this (unless there are conversations going on somewhere I'm not aware of).
I don't know about you, but I smell a rat....or several of them.
>Every human being is capable of something astounding that no AI is capable of yet: creating true novelty.
This isn't true. The vast majority are incapable of original intellectual output, and in fact die having had no intellectual output. ChatGPT is more interesting to talk to than the average person, because it's about as smart but has memorized a lot more.
Wow, that’s cynical. I said we all have the capability, I didn’t say we all use this capability. If you think chatgpt is more interesting then you’re talking to a different crowd than I do. And don’t forget that the conversation with chatgpt doesn’t work without your prompt. You take that part for granted, but that is the actually the creative part. It cannot do anything without intelligent human input. The output isn’t creative. It’s impressive but it’s basically statistics on steroids. It’s fundamentally incapable of novelty-creation and absolutely not as smart as humans. Just the insane amount of sensory fusion, processing and predictive modeling in the “average person” performing an average task is far beyond what any machine can do.
> It cannot do anything without intelligent human input.
Well, no human can do anything interesting (by today's standards) without intelligent human input too.
> it’s basically statistics on steroids.
Be careful with "it's just a ...". "It's just a Turing machine" can run many things, possibly including whole brain emulation. "It's just a statistical model" can encompass all the creative output that can be built upon existing works. Give it a human esthetic module (like they did something similar with ChatGPT and reinforcement learning from human feedback) and allow it to train on its own output and it may even create it's own art style.
I could be wrong but our brain and AI run on very different substrates and follow very different principles. Don’t forget that your brain can do amazing things with just ~200 Watts of power, including “training" (which is orders of magnitude more efficient than GPU-hungry AIs). We are so used to our ways of thinking that we easily take ourselves for granted.
And I think you are capable of creativity even I lock you into a dark room with no inputs! 😅
I've meant the totality of inputs starting from birth. But you can run large language model with empty input buffer too. I doubt that it will come up with something creative, but it's mostly technical limitation. It is trained to produce outputs that are expected in a given context and with no context it doesn't know that it should come up with something creative.
The human brain with 1-2 years of prior experience (a toddler) is more sample-efficient (that is requires less examples to learn a concept) than newly initialized neural network, but trained large language models show quite impressive in-context learning (that is forming a new concept from examples provided in their input buffer). They don't retain that knowledge though.
Yes, modern AIs aren't yet anywhere near human-level diversity of abilities, but it's better to not underappreciate their progress.
Thank you 🙏 this is one of the most hopeful things I have read in a long time...thanks for reaching into my “echo chamber “ and giving me a little shake!
One thing that makes people seem stupid is The Curse of Knowledge. When you know something it's almost impossible to understand what it is like to not know it. For example, I am currently listening to the Bone Valley podcast about a miscarriage of justice in Florida, and all I can think is, "geez these yahoos in Florida are complete idiots, what are they doing putting the wrong man in prison when it's so obvious?" But of course, I know some key facts that the participants didn't know (but should have...)
See also: the curse of ~~knowledge~~ *education* causing delusions of omnsicience. I mean sure, "everyone's doing their best", but resting on one's laurels and "being nice" may not be a recipe for success. If this climate change thing is real and we can't even be bothered to push ourselves....watch out!
There's been a lovely positive vibe on many substacks I'm subscribed to of late. I wonder if it's because this gang are all meta contrarians, and the educated position currently seems to be that people are stupid and everything is terrible. Articles like yours today leave me feeling far more refreshed and optimistic, and it's been a joy to find sources of news and information that give you a lift, rather than always hitting you down. Thank you!
"If a blob disagrees with you, that's because it’s a big dumb sack of gelatinous ooze." This is precisely why we should NOT argue with people we aren't deeply connected with and why social media has convinced us, incorrectly, by sheer inundation with blobs, that the world is full of stupid people.
Maybe we are only wired to learn or teach meaningfully to people that we do have a connection with? It seems that whenever we try and separate human things into their component pieces, to rebuild them according to our vision, we just wind up with broken pieces that don't do anything anymore.
It’s true if you are the Yanomami of Brazil...they have little need of the modern world, bureaucracy and are comfortable with lower life expectancy too. If everything in human societies is interrelated, though, creating thousands of specialists in this or that bit and NO generalists is a problem. My next book aims to be part of the solution.
I think you are absolutely right. The modern world is kind of all arms and legs and no head isn't it? Really good at getting stuff done but no overarching vision?
I get asked from time to time, what is the use of having a lot of mathematical training when a computer can do all of that better than you ever will, and my answer is that it is the engineer's job to know which problems the computer will get wrong. In too much of our world it is very difficult to have both the breadth and the depth of knowledge needed for meaningful oversight, so we try and delegate to a system and procedures. I think we have gone far enough down that road now to say that it is not working out and is not likely to in the future.
So many of the changes that have been made to our society were made without adequate cost-benefit analysis or rather with the assumption that there is no cost and we have lost much of what traditional peoples have preserved. Can it be recovered, and recovered without losing all of the gains of modernity will be the question of the next 100 years.
"my answer is that it is the engineer's job to know which problems the computer will get wrong"
I've been saying for a while now that the purpose of much modern education is to get people to the point where they can tell when the computer is giving them the wrong answer.
There's a tribe in I think the Amazon? They routinely kill any outsiders that approach. That would be an interesting study of anyone got close enough, although I suspect we'd just discover that they live like everyone else in the world.
Your post reminded me of something Chesterton said(I'll be paraphrasing due to laziness), 'There are two types of things in life, things that a person should only do if exceptional like composing an opera and things that everyone should do for themselves whether they are good at them or bad, like composing a love letter. Democracy is simply the belief that the really important things(presumably including running the country) fall into the latter category.'
Perhaps only tangentially related but that is what your post made me think of.
Your link for "naive realism" links to the thesis, in the philosophy of perception, that we are directly aware of the things we see (eg trees), something that's surely true (esp if the alternative, as the article says, is the—crazy!— thesis that we're directly aware of our mental representations of trees instead). Probably you meant to link to the page on naive realism in social psychology (https://en.wikipedia.org/wiki/Naïve_realism_(psychology))?
I've been inhaling cognitive distortions through Freakonomics and associated reading of behavioral economics. I love the topics, but it has changed how I view other humans (this has also changed through years of working in psychiatry). I do view other humans as stupid, and this is disheartening if not terrifying. Thanks for challenging my paradigm! I've loved your newsletter! I would be appreciative if you would check mine out. I'm looking for feedback on how to improve it. Thanks Adam!
Brilliant piece! Thank you Adam. You tickled my brain today. I gave up thinking people were stupid when I decided to adopt a live and let live philosophy. Before that, I attacked others for their idiotic ideas and pummeled them with the shortcoming of their cognitive bias. But alas, I had to keep looking up which C.B. fit their deficiencies and too often found myself within the problem. Humans are like that.
seeing others as merely stupid is a first step toward a dehumanisation of them; of seeing them as something _less_ than human. your piece is a lovely corrective to that rather sinister idea. cheers
The human brain has what, 90 billion neurons? Yep the top selling book, perhaps of all time, is Prince Harry's. As many people have pointed out in the past, most humans live because a small fraction of others are at the controls, making discoveries in science, medicine, and keeping most of the others from killing themselves.
Doubtful. When the people or systems at the controls change, when their philosophy of operation is entirely replaced, there are hiccups but no mass suicides, no loss of ability to function. I think that the small fraction are largely LARPing as saviors.
Yeah, and we still got explosive population growth never before seen, thanks to experts. I'm not saying experts are perfect, but they can definitely do much good, and I don't think they're ever going away. The experts that killed all those people were basically unaligned AI.
Thank you for writing this. This has been on my mind the last year or so as I’ve studied the works of Elinor and Vincent Ostrom. They stood between a mainstream economics and political science that believed people unable to govern themselves. They (especially Elinor) thwarted the popular prisoner’s dilemma conclusion that only radical markets or top-down interventions could solve collective action problems. They believed deeply in the capacity of regular citizens to work together to solve problems.
They sound interesting! Have you read/written anything good on them?
Vincent’s work is well captured in his book “The Intellectual Crisis in American Public Administration.” For Elinor, read Governing the Commons and her American Economic Review article, Beyond Markets and States.
Beautiful post.
Every human being is capable of something astounding that no AI is capable of yet: creating true novelty.
The notion of intelligence, especially when measured as IQ, so profoundly misses the point of how amazing human minds are.
*The degree to which people actually use* this capability on the other hand is a very different matter.
The popular (or, unconsidered) notion that humans are running somewhere near optimality seems rather dangerous to me.
Sadly, very little, yes. Creating something truly novel takes very hard work.
Even worse: creating something truly novel first requires novel ideas, and most of the new ideas I've seen in the last decade or two seem to be designed towards monopoly and wealth maximization for a few individuals.
Consider how much polarization and disharmony social media has injected into society, yet consider how much *genuine* innovation we see in that space. How many new user-facing features have we seen released by Facebook, Twitter, Reddit, etc in the last decade. Also notice how few people seem to notice this (unless there are conversations going on somewhere I'm not aware of).
I don't know about you, but I smell a rat....or several of them.
>Every human being is capable of something astounding that no AI is capable of yet: creating true novelty.
This isn't true. The vast majority are incapable of original intellectual output, and in fact die having had no intellectual output. ChatGPT is more interesting to talk to than the average person, because it's about as smart but has memorized a lot more.
Wow, that’s cynical. I said we all have the capability, I didn’t say we all use this capability. If you think chatgpt is more interesting then you’re talking to a different crowd than I do. And don’t forget that the conversation with chatgpt doesn’t work without your prompt. You take that part for granted, but that is the actually the creative part. It cannot do anything without intelligent human input. The output isn’t creative. It’s impressive but it’s basically statistics on steroids. It’s fundamentally incapable of novelty-creation and absolutely not as smart as humans. Just the insane amount of sensory fusion, processing and predictive modeling in the “average person” performing an average task is far beyond what any machine can do.
> It cannot do anything without intelligent human input.
Well, no human can do anything interesting (by today's standards) without intelligent human input too.
> it’s basically statistics on steroids.
Be careful with "it's just a ...". "It's just a Turing machine" can run many things, possibly including whole brain emulation. "It's just a statistical model" can encompass all the creative output that can be built upon existing works. Give it a human esthetic module (like they did something similar with ChatGPT and reinforcement learning from human feedback) and allow it to train on its own output and it may even create it's own art style.
I could be wrong but our brain and AI run on very different substrates and follow very different principles. Don’t forget that your brain can do amazing things with just ~200 Watts of power, including “training" (which is orders of magnitude more efficient than GPU-hungry AIs). We are so used to our ways of thinking that we easily take ourselves for granted.
And I think you are capable of creativity even I lock you into a dark room with no inputs! 😅
I've meant the totality of inputs starting from birth. But you can run large language model with empty input buffer too. I doubt that it will come up with something creative, but it's mostly technical limitation. It is trained to produce outputs that are expected in a given context and with no context it doesn't know that it should come up with something creative.
The human brain with 1-2 years of prior experience (a toddler) is more sample-efficient (that is requires less examples to learn a concept) than newly initialized neural network, but trained large language models show quite impressive in-context learning (that is forming a new concept from examples provided in their input buffer). They don't retain that knowledge though.
Yes, modern AIs aren't yet anywhere near human-level diversity of abilities, but it's better to not underappreciate their progress.
Most people have better things to do than be needlessly intellectual.
well said!
Thank you 🙏 this is one of the most hopeful things I have read in a long time...thanks for reaching into my “echo chamber “ and giving me a little shake!
Renée
Indeed. It is hopeful to take a breath and allow people to get on with life—their life.
One thing that makes people seem stupid is The Curse of Knowledge. When you know something it's almost impossible to understand what it is like to not know it. For example, I am currently listening to the Bone Valley podcast about a miscarriage of justice in Florida, and all I can think is, "geez these yahoos in Florida are complete idiots, what are they doing putting the wrong man in prison when it's so obvious?" But of course, I know some key facts that the participants didn't know (but should have...)
See also: the curse of ~~knowledge~~ *education* causing delusions of omnsicience. I mean sure, "everyone's doing their best", but resting on one's laurels and "being nice" may not be a recipe for success. If this climate change thing is real and we can't even be bothered to push ourselves....watch out!
It's not real. https://maajidnawaz.substack.com/p/the-cumbria-coal-mine-and-180-climate?utm_medium=email
Educate yourself from real scientists, not biased grifters: https://skepticalscience.com/argument.php
There's been a lovely positive vibe on many substacks I'm subscribed to of late. I wonder if it's because this gang are all meta contrarians, and the educated position currently seems to be that people are stupid and everything is terrible. Articles like yours today leave me feeling far more refreshed and optimistic, and it's been a joy to find sources of news and information that give you a lift, rather than always hitting you down. Thank you!
"If a blob disagrees with you, that's because it’s a big dumb sack of gelatinous ooze." This is precisely why we should NOT argue with people we aren't deeply connected with and why social media has convinced us, incorrectly, by sheer inundation with blobs, that the world is full of stupid people.
Maybe we are only wired to learn or teach meaningfully to people that we do have a connection with? It seems that whenever we try and separate human things into their component pieces, to rebuild them according to our vision, we just wind up with broken pieces that don't do anything anymore.
That sums up the pre-literate tribal critique of social scientists like me very well. There is no easy rebuttal on my end!
Granted that there is no easy rebuttal, is the critique true, in your opinion?
It’s true if you are the Yanomami of Brazil...they have little need of the modern world, bureaucracy and are comfortable with lower life expectancy too. If everything in human societies is interrelated, though, creating thousands of specialists in this or that bit and NO generalists is a problem. My next book aims to be part of the solution.
I think you are absolutely right. The modern world is kind of all arms and legs and no head isn't it? Really good at getting stuff done but no overarching vision?
I get asked from time to time, what is the use of having a lot of mathematical training when a computer can do all of that better than you ever will, and my answer is that it is the engineer's job to know which problems the computer will get wrong. In too much of our world it is very difficult to have both the breadth and the depth of knowledge needed for meaningful oversight, so we try and delegate to a system and procedures. I think we have gone far enough down that road now to say that it is not working out and is not likely to in the future.
So many of the changes that have been made to our society were made without adequate cost-benefit analysis or rather with the assumption that there is no cost and we have lost much of what traditional peoples have preserved. Can it be recovered, and recovered without losing all of the gains of modernity will be the question of the next 100 years.
"my answer is that it is the engineer's job to know which problems the computer will get wrong"
I've been saying for a while now that the purpose of much modern education is to get people to the point where they can tell when the computer is giving them the wrong answer.
There's a tribe in I think the Amazon? They routinely kill any outsiders that approach. That would be an interesting study of anyone got close enough, although I suspect we'd just discover that they live like everyone else in the world.
yep!
yep!
Your post reminded me of something Chesterton said(I'll be paraphrasing due to laziness), 'There are two types of things in life, things that a person should only do if exceptional like composing an opera and things that everyone should do for themselves whether they are good at them or bad, like composing a love letter. Democracy is simply the belief that the really important things(presumably including running the country) fall into the latter category.'
Perhaps only tangentially related but that is what your post made me think of.
The know it all never learns. The know it all thinks people are stupid.
I wholeheartedly agree, proving that you are NOT stupid!
well said
#metoo
I would replace “people are not stupid” with “people are doing the best they can at the time.”
Your link for "naive realism" links to the thesis, in the philosophy of perception, that we are directly aware of the things we see (eg trees), something that's surely true (esp if the alternative, as the article says, is the—crazy!— thesis that we're directly aware of our mental representations of trees instead). Probably you meant to link to the page on naive realism in social psychology (https://en.wikipedia.org/wiki/Naïve_realism_(psychology))?
Thanks! People aren't stupid, but I am.
I've been inhaling cognitive distortions through Freakonomics and associated reading of behavioral economics. I love the topics, but it has changed how I view other humans (this has also changed through years of working in psychiatry). I do view other humans as stupid, and this is disheartening if not terrifying. Thanks for challenging my paradigm! I've loved your newsletter! I would be appreciative if you would check mine out. I'm looking for feedback on how to improve it. Thanks Adam!
Brilliant piece! Thank you Adam. You tickled my brain today. I gave up thinking people were stupid when I decided to adopt a live and let live philosophy. Before that, I attacked others for their idiotic ideas and pummeled them with the shortcoming of their cognitive bias. But alas, I had to keep looking up which C.B. fit their deficiencies and too often found myself within the problem. Humans are like that.
Lovely text - and if ever come to Brazil, please let us know!
seeing others as merely stupid is a first step toward a dehumanisation of them; of seeing them as something _less_ than human. your piece is a lovely corrective to that rather sinister idea. cheers
The human brain has what, 90 billion neurons? Yep the top selling book, perhaps of all time, is Prince Harry's. As many people have pointed out in the past, most humans live because a small fraction of others are at the controls, making discoveries in science, medicine, and keeping most of the others from killing themselves.
Doubtful. When the people or systems at the controls change, when their philosophy of operation is entirely replaced, there are hiccups but no mass suicides, no loss of ability to function. I think that the small fraction are largely LARPing as saviors.
Experts have killed more than they have saved.
Harry's book sales numbers are lies.
Mark my word
Not really no, I think we wouldn't have had such explosive population growth if experts failed in that particular way
200 million killed in the 20th century by experts.
Yeah, and we still got explosive population growth never before seen, thanks to experts. I'm not saying experts are perfect, but they can definitely do much good, and I don't think they're ever going away. The experts that killed all those people were basically unaligned AI.
Clean water didn't require experts. Unless your counting experts for creating capitalism.
Implementing democracy required experts. Technology required experts. Are you an anarcho-primitivist or something?
Good point. But whose Prince Harry?
Ironically, it is discoveries in science that lead to climate change that might kill us all!! 😂😂