Arvind Narayanan on gradual diffusion; Andrey Mir on the digital world vs. the real world; Kai Williams on the Tokyo coding tournament; Hollis Robbins on shape rotators
Hollis Robbins says so much with which I fully concur that in an unabashed display of mood affiliation and confirmation bias I will hereby pledge to pre-order her book.
My first thoughts after reading her piece were of tavern scenes from Captain Alatriste novels in which the shape rotator swordsman enjoys the company of his wordcel buddy a subversive poet. Once upon a time, I suppose. I had hoped to expound on the intellectual tradition that she describes and suggest a countervailing tradition rooted in late medieval notions of chivalry and tracing it down through the rise and spread of the printing press to towns and villages across the European countryside serving a growing literate commoner class that would lead to strict licensing of printing presses and then on to the levellers and their mass readings in the English Civil War era and how Bernard Bailyn rooted the American revolution in their tradition and how mass educational events like lyceum lectures, chautauquas, and even polite vaudevillle form a great American tradition of self-directed learning. Great figures from Ralph Waldo Emerson to the great Eubie Blake were largely self-directed in their aspirations and achievements and brought this spirit to their large, mainly commoner audiences. And one wonders if the modern book tour might be considered an extension of this American popular education tradition? Perhaps the young apprentice might not just read at night but take in lectures, talks, and informative variety shows as well in good social company. But alas, there are not enough hours in the day and Frank Furedi’s Authority, the Huizinga, the Bailyn, and the leveller history go back up on the shelf.
that traces political organization via public reading of tracts in taverns, mass petition delivery, apprentice organizational meetings, and church sermons. It's intended for a popular audience and reads easily. I don't think it snoozy but there are others that I like as well
But _which ones_ of the Narayanan tasks AI will be used for depends on the relative value of having AI do them. It’s comparative advantage all the way down.
Arvind Narayanan says some interesting things. But part of what makes them interesting is how badly he says them.
1 The most certain strength of AI is its ability to work in tandem with humans to address exactly the problems he says AI can't be trusted to address on its own.
2 "In affluent societies, 95% of jobs are things we do because we need things for people to do, not things we do for people to survive. I’m saying that in the future, that 95% will go to 99%, maybe even 100%."
What exactly is required to survive? Food, water, and shelter? So ok, maybe most of our shelter isn't required to survive. But is heat and cooling? What about medical care? What of that is "required"? I don't see us at 95% of jobs for things other than what we need to survive now. I don't even see that in the future, much less 99+%.
3 I very much dislike his use of the word value where he means price, cost, or product.
I think 95% isn't unreasonable. Taking an example that has been alluded to in the recent history essays, in 1900 about 40% of the US workforce was engaged in food production. Today, it's about 2%. That hasn't all been automation since some of it has been improvements in agriculture but a lot of it has been. We have huge swaths of employment in leisure pursuits and tourism. Even a lot of medical care is not strictly necessary for survival, though it does definitely improve quality of life. The trend is definitely in the direction he's pointing.
Looked at more closely, your example isn't quite as good. Compared to when 40% were producing food, today's 2% is support by a much larger group of input suppliers, manufacturers, distributors, and food services. These account for an estimated 20% of the workforce.
Another thought: I think that AI economics has more to do with the margins. Things that weren’t worth writing software to do before now have custom software written for them because it’s trivial. People who are clever but who don’t have the time or complete knowledge to learn to write software now can. So it’s not the highly repeatable tasks that necessarily be automated but the idiosyncratic ones. The highly repeatable tasks are the ones where automation by experts through software made sense already.but ai will tackle an entirely different category of stuff. (Some of this will simply be translation from one format to another, which itself used to be very expensive, but now will be relatively trivially expensive). What does all of this mean in terms of what it does to the economy? Who knows!
Numerous highly successful products were widely panned as useless when they first came on the market. Faster and cheaper test marketing has a huge benefit.
As for “The Intellectual Habits of the British Working Class” - I forgot the exact title. I will Google it shortly - I read the couple dozen reviews and responses on the Wikipedia page.
(Itself something of a time capsule of literacy.)
The fellow who suggested the alternative title “what it is we used to do before we watched telly” has it about right.
(Have you noticed that information density seems to be decreasing in the content we consume? I read somewhere the other day that the sentences are getting much shorter … even an episode of M*A*S*H, which was decidedly for *everyone*, contained far more throwaway references than anything now … when they try that sort of thing now, and they do - I’m sure most of our entertainment providers are fantastically literate, in the way that the national lampoon and SNL writers were often just out of the Ivy League - they have to tediously Wikipedia it up in the dialogue. True, the Big Bang Theory was popular, with lots of pop culture references per episode, but I think it was supposed to appeal to a particular Gen X or millennial nerd audience, only turned out we were temporarily demographically all then nerds, at least by that low bar of being familiar with shared American pop culture (sigh, lament), many of us having purchased the Stephen Hawking book, that and the Joseph Campbell book, although not actually read them … I was talking to a sweet Spanish-only-speaking young man, and somehow our fumbling efforts to communicate, had produced that he did not like Star Wars or Harry Potter and had not ever watched them because they were boring to him, too long! Because of the circumstances of where he lives at the moment, and their viewing habits I had a vision of him stuck with his elders watching horrible Turkish soap operas dubbed into Spanish, but he explained that he liked “comedies”. Mexican or Latin American comedies? I asked, doubtfully.
Oh no, he said - American. (Relief.)
I asked him for an example and he said Big Bang Theory, without a moment’s hesitation, his face lighting up.
I said, thinking I had a gotcha: but the characters on that show loved Star Wars and Star Trek. You can’t understand the show without watching those originals.
But he had me: “Penny!” he replied.
OK, that’s true. Penny didn’t like those shows. He was obviously a Penny super fan.)
Anyhoo, reading about the Jonathan Rose book and the supposed uptake of improving “continuing education” made me recall Reader’s Digest condensed books. My husband’s (working class, farm-raised) grandmother had a collection of them, either picked up at the local “paperback” exchange or mailed with her subscription. The colors of the spines were cleverly chosen to coordinate with much mid-century decor. She certainly read each one in its entirety, but she kept them on bedroom shelves for the former reason.
We used to arrive to visit and each select one at bedtime, ready to hate read.
That’s the trouble with definitions like middlebrow.
Compared to the books Americans - “working class” or college-educated, difference unclear - consume now, which they buy at Costco I guess, or hopefully pick up at the library so as to reduce waste - those old readers digest condensed versions were practically Middlemarch.
I’ve observed among the links provided here as elsewhere for some years, a disinclination to examine the existent, the contemporary. People seem much more interested in parsing the nuances of the future as they imagine it.
It would be interesting to have a moratorium on all that. A period where writers - er, intellectuals - must describe all the real physical aspects of the city they live in, and the people therein.
What is the point pf describing the current state of affairs, if not to show what works and what fails? What is the point of knowing what works and what fails, if not to make better plans for the future?
If all that is pointless, then what is left to discuss? If all you want is a log of current events, with no interest in using that information, what do you want?
B. Honesty is relative. The truth, the whole truth, and nothing but the truth ... is impossible with limited resources, and different people want different censoring.
Those first two match my experience and expectations. Like that famous quote by someone smarter than me, stocking manufacturers got rich by making stockings cheap enough for everybody, not just better stockings for the queen. Almost all jobs today are 99% mundane and repetitive. They only require thinking outside the box 1% of the time. All those gardeners hired by businesses and apartment managers are prime examples, just begging for some automation for the routine boring stuff and leaving humans to expand the creative parts.
Grocery stores are an interesting tradeoff -- the cheap big box stores are cheaper partly because they put the shipping boxes on the shelves instead of unpacking them. The seen costs of the stockers are obvious, but people don't see the unseen costs of extra shelving taken up by mostly empty boxes. When robotic stockers bring the price down, the cost differentials will change.
I would argue that very few jobs are anywhere close to only thinking outside the box 1% of the time.
But I would also agree that a lot of workers think outside the box far less than they should and many businesses have procedures that don't benefits from such thinking as much as they could.
Eh, I use 99% and 1% like most people (99%!) do, as synonyms for the equally vague "most" and "few". That's probably why the climate alarmists picked up on 97% consensus.
Per Hollis... 50 years ago, I was the kid in university thinking "this sucks", bailed, got into a trade, then several, and morphed into a few tidy little businesses. At the time, I was written off, a dropout loser, and now... I'm feeling like I saw a better way to engage the world than that promoted via the jerry rigged educational entities pumping out bumwads for the economy.
Re Mir's views, I agree with much of the quoted segment. But I think people will be afraid (and rightly so) that digital minds may not be conscious. That would mean migration into the digital world as he suggests would be death.
“But any work AI can do is infinitely reproducible, and when it’s infinitely reproducible, it doesn’t have a lot of value, because its cost is going to come down to the market—anybody can provide it at very low cost.”
Do people agree with this? It’s not necessarily my experience with AI. rather I’m surprised at how well it conforms to my particular questions and I think I get different answers than people with different questions. I don’t see it as a machine per se, automating repeatable tasks. But I’m still not quite sure what to make of AI.
Hollis Robbins says so much with which I fully concur that in an unabashed display of mood affiliation and confirmation bias I will hereby pledge to pre-order her book.
My first thoughts after reading her piece were of tavern scenes from Captain Alatriste novels in which the shape rotator swordsman enjoys the company of his wordcel buddy a subversive poet. Once upon a time, I suppose. I had hoped to expound on the intellectual tradition that she describes and suggest a countervailing tradition rooted in late medieval notions of chivalry and tracing it down through the rise and spread of the printing press to towns and villages across the European countryside serving a growing literate commoner class that would lead to strict licensing of printing presses and then on to the levellers and their mass readings in the English Civil War era and how Bernard Bailyn rooted the American revolution in their tradition and how mass educational events like lyceum lectures, chautauquas, and even polite vaudevillle form a great American tradition of self-directed learning. Great figures from Ralph Waldo Emerson to the great Eubie Blake were largely self-directed in their aspirations and achievements and brought this spirit to their large, mainly commoner audiences. And one wonders if the modern book tour might be considered an extension of this American popular education tradition? Perhaps the young apprentice might not just read at night but take in lectures, talks, and informative variety shows as well in good social company. But alas, there are not enough hours in the day and Frank Furedi’s Authority, the Huizinga, the Bailyn, and the leveller history go back up on the shelf.
That leveller history if it’s world turned upside down will level the reader and set him slumbering in no time flat.
No, the one that I had to hand was John Rees' The Leveller Revolution (
https://www.friendsjournal.org/book/the-leveller-revolution-john-rees/ )
that traces political organization via public reading of tracts in taverns, mass petition delivery, apprentice organizational meetings, and church sermons. It's intended for a popular audience and reads easily. I don't think it snoozy but there are others that I like as well
But _which ones_ of the Narayanan tasks AI will be used for depends on the relative value of having AI do them. It’s comparative advantage all the way down.
Arvind Narayanan says some interesting things. But part of what makes them interesting is how badly he says them.
1 The most certain strength of AI is its ability to work in tandem with humans to address exactly the problems he says AI can't be trusted to address on its own.
2 "In affluent societies, 95% of jobs are things we do because we need things for people to do, not things we do for people to survive. I’m saying that in the future, that 95% will go to 99%, maybe even 100%."
What exactly is required to survive? Food, water, and shelter? So ok, maybe most of our shelter isn't required to survive. But is heat and cooling? What about medical care? What of that is "required"? I don't see us at 95% of jobs for things other than what we need to survive now. I don't even see that in the future, much less 99+%.
3 I very much dislike his use of the word value where he means price, cost, or product.
I think 95% isn't unreasonable. Taking an example that has been alluded to in the recent history essays, in 1900 about 40% of the US workforce was engaged in food production. Today, it's about 2%. That hasn't all been automation since some of it has been improvements in agriculture but a lot of it has been. We have huge swaths of employment in leisure pursuits and tourism. Even a lot of medical care is not strictly necessary for survival, though it does definitely improve quality of life. The trend is definitely in the direction he's pointing.
Looked at more closely, your example isn't quite as good. Compared to when 40% were producing food, today's 2% is support by a much larger group of input suppliers, manufacturers, distributors, and food services. These account for an estimated 20% of the workforce.
So can we stop having think pieces about how affluent people should have more babies, or should we do AI on odd days and babies on even days?
And no days on more immigration.
Another thought: I think that AI economics has more to do with the margins. Things that weren’t worth writing software to do before now have custom software written for them because it’s trivial. People who are clever but who don’t have the time or complete knowledge to learn to write software now can. So it’s not the highly repeatable tasks that necessarily be automated but the idiosyncratic ones. The highly repeatable tasks are the ones where automation by experts through software made sense already.but ai will tackle an entirely different category of stuff. (Some of this will simply be translation from one format to another, which itself used to be very expensive, but now will be relatively trivially expensive). What does all of this mean in terms of what it does to the economy? Who knows!
Numerous highly successful products were widely panned as useless when they first came on the market. Faster and cheaper test marketing has a huge benefit.
As for “The Intellectual Habits of the British Working Class” - I forgot the exact title. I will Google it shortly - I read the couple dozen reviews and responses on the Wikipedia page.
(Itself something of a time capsule of literacy.)
The fellow who suggested the alternative title “what it is we used to do before we watched telly” has it about right.
(Have you noticed that information density seems to be decreasing in the content we consume? I read somewhere the other day that the sentences are getting much shorter … even an episode of M*A*S*H, which was decidedly for *everyone*, contained far more throwaway references than anything now … when they try that sort of thing now, and they do - I’m sure most of our entertainment providers are fantastically literate, in the way that the national lampoon and SNL writers were often just out of the Ivy League - they have to tediously Wikipedia it up in the dialogue. True, the Big Bang Theory was popular, with lots of pop culture references per episode, but I think it was supposed to appeal to a particular Gen X or millennial nerd audience, only turned out we were temporarily demographically all then nerds, at least by that low bar of being familiar with shared American pop culture (sigh, lament), many of us having purchased the Stephen Hawking book, that and the Joseph Campbell book, although not actually read them … I was talking to a sweet Spanish-only-speaking young man, and somehow our fumbling efforts to communicate, had produced that he did not like Star Wars or Harry Potter and had not ever watched them because they were boring to him, too long! Because of the circumstances of where he lives at the moment, and their viewing habits I had a vision of him stuck with his elders watching horrible Turkish soap operas dubbed into Spanish, but he explained that he liked “comedies”. Mexican or Latin American comedies? I asked, doubtfully.
Oh no, he said - American. (Relief.)
I asked him for an example and he said Big Bang Theory, without a moment’s hesitation, his face lighting up.
I said, thinking I had a gotcha: but the characters on that show loved Star Wars and Star Trek. You can’t understand the show without watching those originals.
But he had me: “Penny!” he replied.
OK, that’s true. Penny didn’t like those shows. He was obviously a Penny super fan.)
Anyhoo, reading about the Jonathan Rose book and the supposed uptake of improving “continuing education” made me recall Reader’s Digest condensed books. My husband’s (working class, farm-raised) grandmother had a collection of them, either picked up at the local “paperback” exchange or mailed with her subscription. The colors of the spines were cleverly chosen to coordinate with much mid-century decor. She certainly read each one in its entirety, but she kept them on bedroom shelves for the former reason.
We used to arrive to visit and each select one at bedtime, ready to hate read.
That’s the trouble with definitions like middlebrow.
Compared to the books Americans - “working class” or college-educated, difference unclear - consume now, which they buy at Costco I guess, or hopefully pick up at the library so as to reduce waste - those old readers digest condensed versions were practically Middlemarch.
I’ve observed among the links provided here as elsewhere for some years, a disinclination to examine the existent, the contemporary. People seem much more interested in parsing the nuances of the future as they imagine it.
It would be interesting to have a moratorium on all that. A period where writers - er, intellectuals - must describe all the real physical aspects of the city they live in, and the people therein.
What is the point pf describing the current state of affairs, if not to show what works and what fails? What is the point of knowing what works and what fails, if not to make better plans for the future?
If all that is pointless, then what is left to discuss? If all you want is a log of current events, with no interest in using that information, what do you want?
I would amend your first sentence to “What is the point of describing the current state of affairs *honestly*.”
A. That's a given.
B. Honesty is relative. The truth, the whole truth, and nothing but the truth ... is impossible with limited resources, and different people want different censoring.
Evidently so. I’ve noticed that the powers that be cannot even use the word “Americans” as an identifier truthfully.
What is your truthful definition of that identifier?
Oh, I’d be satisfied with just a gesture at the original dictionary definition, but the long standing colloquial one would be even better.
I would love a log like that. The step was skipped, as if our intellectual class threw the pebble on that hopscotch square.
Maybe it is an artifact of even the reporters, mostly reporting from their screen.
ETA: I find the world very interesting.
Then surely you aren't the only one. You should create one for the others.
I know, and I’m also supposed to pick up all the trash in my city of 4 million people. I have received the wisdom of the internet often enough.
Those first two match my experience and expectations. Like that famous quote by someone smarter than me, stocking manufacturers got rich by making stockings cheap enough for everybody, not just better stockings for the queen. Almost all jobs today are 99% mundane and repetitive. They only require thinking outside the box 1% of the time. All those gardeners hired by businesses and apartment managers are prime examples, just begging for some automation for the routine boring stuff and leaving humans to expand the creative parts.
Grocery stores are an interesting tradeoff -- the cheap big box stores are cheaper partly because they put the shipping boxes on the shelves instead of unpacking them. The seen costs of the stockers are obvious, but people don't see the unseen costs of extra shelving taken up by mostly empty boxes. When robotic stockers bring the price down, the cost differentials will change.
I would argue that very few jobs are anywhere close to only thinking outside the box 1% of the time.
But I would also agree that a lot of workers think outside the box far less than they should and many businesses have procedures that don't benefits from such thinking as much as they could.
Eh, I use 99% and 1% like most people (99%!) do, as synonyms for the equally vague "most" and "few". That's probably why the climate alarmists picked up on 97% consensus.
Per Hollis... 50 years ago, I was the kid in university thinking "this sucks", bailed, got into a trade, then several, and morphed into a few tidy little businesses. At the time, I was written off, a dropout loser, and now... I'm feeling like I saw a better way to engage the world than that promoted via the jerry rigged educational entities pumping out bumwads for the economy.
Re Mir's views, I agree with much of the quoted segment. But I think people will be afraid (and rightly so) that digital minds may not be conscious. That would mean migration into the digital world as he suggests would be death.
"Don't beam me up, Scotty!"
“But any work AI can do is infinitely reproducible, and when it’s infinitely reproducible, it doesn’t have a lot of value, because its cost is going to come down to the market—anybody can provide it at very low cost.”
Do people agree with this? It’s not necessarily my experience with AI. rather I’m surprised at how well it conforms to my particular questions and I think I get different answers than people with different questions. I don’t see it as a machine per se, automating repeatable tasks. But I’m still not quite sure what to make of AI.