Rise of AI Conference 2019 · Berlin

World After Capital

Albert Wenger · Union Square Ventures

At Rise of AI 2019 in Berlin, Albert Wenger, Managing Partner at Union Square Ventures, argues that the central question for the age of AI is not information but attention: what humans choose to pay attention to. He opens with CO2 levels reaching 415 parts per million and argues society fails to attend to climate change because market prices do not exist for the most important problems, such as an individual's purpose in life, planetary defense against asteroids, or preventing climate change, while advertising-driven attention has a clear price and is therefore over-allocated. He traces how the scarce resource shifted historically from food (foraging) to land (agriculture) to capital (industrial age), arguing markets allocate capital well but fail at allocating attention. His core proposal is that AI should be used to free human attention from what he calls the job loop, the cycle of selling labor to buy goods, rather than to capture more attention for advertising. He closes with three proposals: universal basic income for economic freedom, informational freedom through open APIs to platforms, and individual mindfulness practice, and points to his ongoing open book project, World After Capital.

Key takeaways

  • Wenger argues the defining scarcity of the current era is human attention, not information, which is abundant.
  • Markets allocate attention poorly because the most important questions, like personal purpose or planetary defense, have no price or demand curve.
  • Advertising gives attention a resale price, so companies like Facebook, Google and Amazon are incentivized to capture as much of it as possible.
  • AI should be used to free human attention from the 'job loop' of selling labor to buy goods, not to capture more attention for advertisers.
  • He proposes universal basic income, open programmatic access to platforms, and personal mindfulness practice as complementary paths to reclaiming attention.

Detailed notes

Attention, not information, as the defining scarcity

Wenger opens with a chart showing CO2 in the atmosphere had reached 415 parts per million as of a recent May 2nd reading, and observes that despite knowing the world is warming, people do not act on it, illustrated by a viral clip of an ape scrolling a phone which he calls both funny and disconcerting because it mirrors human behavior on Instagram. He references the Matrix films, arguing the movies get it backwards: humans are not enslaved by machines in a war, they are putting themselves into the pods voluntarily by spending hours on Fortnite or Instagram, and the environmental destruction in the films is better read as a consequence of neglect rather than war.

From food to land to capital: a history of scarcity

He traces roughly 250,000 years of Homo sapiens history: for most of it, food was the scarce resource for foraging tribes, driving migration across the globe. About 10,000 years ago agriculture, through irrigation, seed selection, domestication, writing and currency, shifted scarcity to arable land, producing hierarchical societies and warfare over territory. A few hundred years ago, starting with the Enlightenment, steam and electrical power and chemistry shifted scarcity to physical capital (factories, buildings, railroads); societies tried allocating capital by state edict versus by markets, and markets won decisively. Wenger argues today is not really an information age, since machines generate petabytes of logs every second, but a knowledge age, in which the truly scarce resource is human attention.

Why markets fail to allocate attention

Wenger argues that applying the market/price mechanism, which worked well for physical capital, to attention is a mistake, because the most important things people could attend to have no price, supply or demand curve. Examples: figuring out one's personal purpose in life (which he links to midlife and early-life crises from lack of attention to the question); planetary defense against large asteroid strikes, a rare but catastrophic long-tail risk, for which roughly as few people work full-time on detection as work at one large McDonald's; and climate change, where a carbon tax helps but does not match the scale of the problem. Meanwhile, attention does have a resale price through advertising (he cites Facebook, Twitter, Google and Amazon stock prices tracking how much attention those companies capture), so attention gets over-allocated to being captured by platforms rather than to underpriced existential problems.

Using AI to break the job loop

Wenger describes the 'job loop': selling labor for income, using that income to buy goods made by other people also selling their labor, a cycle he says traps most human attention, including that of the audience after the talk. He argues AI's purpose should be freeing attention from this loop, not capturing more of it for advertising. He cites an Our World in Data chart by Max Roser showing agricultural labor share collapsing from most of the workforce in 1300 to under 5% in many industrialized countries today, but rejects the common reading that this simply proves new jobs always replace old ones; he reads it instead as evidence that attention can be freed from an activity. He argues that if AI is used correctly, a similar future chart, perhaps by the year 2100, should show attention leaving the job loop entirely toward non-economic activity such as fighting climate change, curing disease, or making music, rather than shifting to a new set of jobs.

Three proposals: UBI, informational freedom, and mindfulness

Wenger closes with three proposals from his ongoing book project. First, universal basic income, which he calls 'economic freedom' in the book, giving everyone enough money to freely allocate their time; he notes it is not a new idea, citing Thomas Paine and Martin Luther King as historical supporters, and 2019 US presidential candidate Andrew Yang as well as European parties currently campaigning on it. Second, 'informational freedom': he argues that despite everyone carrying a supercomputer in their pocket, it does not truly serve its owner because companies like Facebook do not expose full programmatic APIs, preventing people from building agents that act on their behalf; he calls for full programmatic access to such platforms. Third, an individual-level proposal: since human brains did not evolve for infinite scrolling of content like cat pictures, everyone needs to develop a personal mindfulness practice to reclaim their own attention, independent of collective policy change. He points listeners to his creative-commons-licensed, iteratively written book, World After Capital, available on GitHub.

Quotes

We're putting ourselves in the pods right now as we speak.

Albert Wenger

Let's stop working on using AI to hog attention.

Albert Wenger

Questions answered in this talk

Why does Albert Wenger say attention, not information, is the key scarcity of our time?

According to Albert Wenger at Rise of AI 2019, machines already generate more information than anyone can use, so the real bottleneck is what humans choose to pay attention to, and the most important problems, like climate change or personal purpose, have no market price to guide that attention.

How does Albert Wenger think AI should be used, according to his Rise of AI 2019 talk?

According to Albert Wenger at Rise of AI 2019, AI should be used to free people from the 'job loop' of selling labor to buy goods, rather than to capture more human attention for advertising, so that freed attention can go toward problems like climate change, disease and creative work.

What three proposals does Albert Wenger make for a post-scarcity attention economy?

According to Albert Wenger at Rise of AI 2019, he proposes universal basic income for economic freedom, full programmatic access to platforms like Facebook for informational freedom, and a personal mindfulness practice so individuals can reclaim their own attention.

Mentioned

attention economy, universal basic income, job loop, informational freedom, World After Capital, Matrix pod scene

Keep exploring

More from Rise of AI Conference 2019

More on Ethics & Trust

Transcript

Auto-generated transcript, not edited. · 3,298 words

Show the full transcript

hi everyone I want to introduce us by saying now for something completely different following the stats from earlier so the talk was built as world after capital and we'll get to that but what I really want to talk about is the relationship of artificial intelligence to how we allocate attention in the world and what do I mean by this well here's a chart of co2 levels in the atmosphere over the last 800,000 years and there's a tiny little thing to notice here if you pay close attention which is this massive spike this is actually very recent data it's from May 2nd to 415 parts per million of co2 in the atmosphere and we all kind of know this right I mean we've hear it in the news and we all kind of know that this means the world is getting hotter although in Berlin today it does not feel that way but all around the world is getting a lot warmer because of this accumulation of greenhouse gases in the atmosphere and so we hear

this and yet we don't really seem to be paying attention to it and why are we not really paying attention to it well because 30 seconds later this is us it gets better monkey see monkey scroll actually I think it's an ape see if you haven't seen this before in a way it's funny but in a way it's also deeply deeply disconcerting it's disconcerting because that really is us when we're an Instagram right and so who is old enough to have seen the Matrix movies like in the movie theater yeah good all right so this next slide will actually make sense to you so this is the pod scene this is when neo wakes up out of his pod when he's sort of liberated from the matrix as a warp and here's where I think the Matrix movies are fundamentally wrong they posited that there was a war between the machines where the machines enslaved the humans that put them in the pods that's totally wrong we're putting ourselves in the pods right now as

we speak so that we can play 24 hours a fortnight or to be on Instagram for 24 hours or whatever else we're doing online we're giving ourselves willingly to the machines and the destruction you see around in the world in the Matrix movies is not the result of a war it's the result of our and neglect of the environment so if you just see the movies from that lens all the sudden they're quite different so leading up to the big question what should we do about this fall and also what really does it have to do with AI well in order to get to that let me take you back a very long way humankind Homo sapiens as we know ourselves today emerged about 250,000 years ago and for the longest time we were foragers and when we were foragers the scarcity in the world was food you tribe either found enough food or it didn't and if it didn't it either starved or had to migrate which is one of the reasons why we

covered the globe actually pretty quickly even like fairly remote places like alpine valleys so then roughly 10,000 years ago we invented this thing called agriculture and it's a series of interlocking inventions we figured out it would season the grass we figured out irrigation domestication of animals some amount of written language currencies and the scarcity shifted the scarcity was no longer food directly it was land arable land land on which he could grow something and if you had enough land you could build a sort of flourishing society and you know we had all these sort of hierarchical societies emerge and kings and queens and and we wound up with many many years of warfare over land because that's how we essentially were allocating land you know there's a society it owned a bunch of land and if it wasn't enough land they usually would go to war to get some more land somewhere else then only a couple hundred years ago we had some other breakthroughs starting with the Enlightenment we figured out how to make

steam power and electrical power and we figured out chemistry and mining and the scarcity shifted again the scarcity wound up no longer being about land they wound up being about capital and by capital I don't mean financial capital because we don't eat dollar bills or drive around in gold bars I mean physical capital factories buildings railroads etc and we came up with a very efficient allocation mechanism in fact we tried kind of two big experiments we had sort of the allocation of capital by state edict in the planned economies and we had the allocation of capital by markets in the Western world for the most part and it turns out markets worked incredibly well at allocating capital so now the big question is okay if we're maybe past the industrial age and maybe we're in a new age many people think of it as the Information Age what's the new scarcity now and it gets the main point of my talk today is that it's not really the information age because information just spews forth

from all these machines that we have today everything is instrumented you know there's you know petabytes of log files being generated every split second so information is not really what this age is about it seems to be everywhere but I think what this age needs to be about is it needs to be about knowledge and if it needs to be our knowledge and by knowledge I mean technical knowledge like figuring out how to deal with climate change but I also mean knowledge very broadly as in including art music etc if that's the age we want to live in then the really scarce thing is attention what are we as humans paying attention to so this is where we get into trouble at the moment we get into trouble because this thing worked so incredibly well for us the market mechanism allocating things by price worked incredibly well it was the dominant paradigm of the Industrial Age and we've all benefited from it massively right so all these great things like this amazing building the the

cars the mobile phones laptops they've all been essentially produced by this kind of mechanism working incredibly well you know and you can read about how it a great sin firm ation and it was capable of allocating lots and lots of things in the physical world in an incredibly powerful way way more powerful it turned out then you know state planners trying to figure out what people wanted and how to provide it to them so now we're so enamored with this model that we are trying to use it to allocate attention and that turns out to be a very bad idea why is it a bad idea it's a bad idea because the most important things that we could be paying attention to don't have prices and they don't have prices because they don't have demand and supply curves at all let me give you some examples what is the market for your purpose what's the market for you as an individual figuring out what your purpose is in life why you're here on this planet

what you want to accomplish there's no supply and no demand there's no price so if you're gonna rely on how much of your attention you should allocate to this based on some kind of price mechanism you will vastly under ella kate personal attention to this incredibly important and profound question this is one of the reasons why we have so many people experiencing midlife crisis or crisis early in their life because this is not something people are paying attention to but it is profound is existential for us as humans here's another one I call it broadly speaking death from above there is no market for this a very large asteroid hits the earth every few millions of years it's an extreme longtail event there is no supply and demand curve for building technology that could detect and deflect these objects as a result we need to collectively as a society decide how much attention we want to give to a problem like this based on past asteroid strikes this is a bad problem when it happens

like a really bad problem so we should be allocating a lot of attention to it but the number of people who sort of work full-time to just detect these kind of objects all day every day are roughly the same number of people who work at one larger McDonald's do you see how there's a problem there somehow and then of course coming back to the problem that I started with which is the climate change problem right so like there is no price for not accidentally turning earth into Mars I know you know I love sci-fi and we talked a lot about thinking about how to turn Mars into Earth I think at the moment we're working on the opposite thing we're turning earth into Mars right so again yes we can come up with a carbon tax I'm all for carbon tax but it doesn't address the magnitude of the problem and so again society as a whole we're vastly under allocating attention to these problems and one of the reasons is because it turns out

there are prices for other things there is a price for Facebook stock and Twitter stock and Google stock and Amazon stock and that price corresponds almost directly to how much attention those companies have hogged all right the or attention Facebook gets the more attention it can resell to advertisers and the better of Facebook shareholders are so there is and has been a price for advertising for a long time this happens to be a Piccadilly circle in in London right so advertising has been around for a long time it's the resale of attention and because prices exist for those things we're allocating a lot of our time and attention to these problems so there's a lot of people working on how to extract more human attention and funnel it into those systems so that eventually we're all there like monkeys scrolling in our feed and tapping on images wait so far this sounds fairly depressing let me see whether I can take the second half of this talk and talk about what we could do to

make things better first of all let's stop working on using AI to hog attention and this is a beautiful high farm by the way beautiful Pig um so with no offense to the pig let's not do that let's spend less time working on that particular problem let's spend more of our time on freeing up attention now where is most of human attention stuck today it's stuck in this thing in my book I call it the job loop it's this thing where you sell your labor you make an income use that income to buy stuff the stuff is also being made by people who are selling their labor and it's a beautiful circle and it's working incredibly well it's working far too well we're making a lot of stuff we don't need we're using advertising to sell that stuff to people who may won't make happy there are other people working on making that stuff that nobody needs and just goes on and on and on and a ton of human attention is stuck in this

loop most of us are stuck in this loop after this talk most of us will go back to work in this loop this is where we need to use AI we need to use AI to get out of this loop and here's the interesting thing we've done this once before in a way not using AI but this is a beautiful chart it comes from a project called our world and data by Max Rosser who is at Oxford and this shows the percentage of labor force in different countries over time working in agriculture goes all the way back to 1300 so what this shows is that a lot of human attention a lot of people were working in agriculture and then as we went into the industrial age that percentage plummeted in many industrialized countries today it's significantly below 5% and so people often point to this chart and say look you don't actually like all this worry about AI taking over jobs is wrong because look we stopped working in agriculture and we found all

these new jobs but I think that's fundamentally the wrong way to read this graph the way to read this graph is we were spending a huge amount of time in agriculture because we were doing this and then we had to spend a lot less time in agriculture because we were doing this but the modern equivalent the equivalent with what we can do with AI isn't to find a new set of jobs it's to do away with how much human attention is stuck in the job loop so I believe if we do this right today we'll be at whatever 75 80 percent of human attention being stuck in the job loop and at some point in the future maybe the year 2100 maybe later if we do it right but only if we do it right we will have freed up a lot of that human attention to do really interesting things things like fighting climate change things like fighting death from above things like creating new cures for diseases things like creating beautiful or beautiful

music takes like cleaning up the environment taking care of animals like there's so much to do that exists outside of the market so if we do this right in the future some future generation is going to show a chart just like Max's chart and that chart isn't going to show some drops going away being replaced by new jobs it's gonna show attention being stuck in the chop loop going away and going towards non-economic activity so how did we get there well you might say the answer is simple using AI duh I mean that's the whole point of this conference in this talk but I want to spend the last five minutes convincing you that it's not just about the technology because when we went from the agrarian age to the industrial age it wasn't just about replacing the pitchforks with the big combines it was about completely changing how society works we went from living in the country to living in the city we went from large extended families to nuclear families and no families

we went from a lot of common property to a lot of private property and so forth we changed a lot of things and we're gonna need to change a lot of things again and I just want to quickly throw out three proposals that I think are very important to be thinking about at this particular moment in time the first one is something known as ubi or a universal basic income and in my book I call this economic freedom it's the idea that everybody gets some amount of money enough money so that they can basically completely freely allocate their time this is the way we can liberate ourselves from being stuck in the job loop only if you're completely free to decide what you want to work on are you really free to allocate your time and there are people all around the world who are working on this theirs for the first time somebody running in the Democratic presidential primary in the u.s.

named Andrew yang who's got ubi as part of his platform they're parties in Europe that have ubi as part of their platform this is not a new idea the idea goes back a long long time in the u.s. it goes back to the founding fathers Thomas Paine wrote about essentially universal basic income Martin Luther King was in support of it so it's not a new idea but it's an idea whose time has come now and it's essential for freeing ourselves from the job loop the second big idea that I want to throw out here it's what I call informational freedom in the book so we all carry a supercomputer in our pocket right if you go back just 20-some years no amount of money could have bought this much compute power I mean it would have been very very expensive and he certainly wouldn't have been able to fit in your pocket and this thing also talks to everybody else's supercomputer all around the world and yet there's this weird thing you know when I

opened up Michaels supercomputer and I tap any of the app icons all of a sudden this supercomputer that I've paid for that nominal is mine suddenly just works for whatever company's app I've launched it doesn't actually work for me I can't use any of the compute power to control these other applications and why can't I do this I can't do it because Facebook doesn't have a full API that I get to program any which way I see fit so I believe that if we want to not have our attention be hugged by systems like Facebook and Google and Amazon and Twitter and so forth we need full programmatic access to these systems so that we can have agents that act on our behalf interface with those systems and that we can use the compute power that we carry around for our benefit not for the benefit of large centralized corporations and then the final idea that I want to mention is one that we each can get started on today so the first two idea

is universal basic income the economic freedom part and informational freedom are things that are going to require some amount of collective action it's going to require governments to get together and pass smarter regulation than they're doing today new kind of societal systems that need to be constructed this last part is something each and every one of us can get started on today when you leave here today which is to realize that our brain did not evolve in an environment where you could have access to an infinity of cat pictures so every time we see a cat picture our brains like oh there's a cat this is very cute it's like next cat next cat we need to work very hard each and every one of us to kind of free our own attention so even if you had economic freedom even if yet informational freedom if you don't take charge of your brain and how your brain interacts with all this information that's out there you will still not be free and your attention will

still be slave to many of those online things and so everybody everybody needs to develop some kind of mindfulness practice and if you want to read more about all this type of thinking I'm working on this book it's creative commons-licensed it's online I'm writing it iteratively using them and checking it it's a github so you can always see the latest version at world after Capitol thank you

Rise of AI 2027 · Berlin · May 2027

Be in the room for the next one.

Ten years of talks started as conversations in one curated room in Berlin. The next edition brings 300 selected leaders together in May 2027.