The Singularity is Near: How Kurzweil's Predictions Are Faring (2017)
antropy.co.uk
antropy.co.uk
Most of Kurzweil's predictions have two parts, implied or explicit. One is the technology itself. The second is how that technology affects the world. Between the two (1.5) is a lot of ambiguity, whenever anyone tries to "score" Ray.
EG Displays will be built into our eyeglasses. Tech that essentially exists but hasn't yet become widespread or impactful yet. Sometimes tech does become widespread & impactful, but we still predict impacts wrong.
PCs have been on every desk for 25 years. But, administration (to the extent that measurement is possible) seems to have gotten less efficient. HR departments, schools, etc need more administrators today than before PCs. Weird considering PCs are paperwork efficiency machines.
Meanwhile, dating and social life have been radically transformed by personal computing.
A lecture hall, library and university are notably inferior (eg you can pause the lecture) when digitised, yet Harvard is still Harvard. Even a local college is still the local college. Universities are persistent despite their most legible reasons raison d'etres becoming nonessentials.
When I was in college, "Keynes' 15 hr workweek prediction for the 80s" really fascinated me. He was right, on paper, in his efficiency/growth assumptions. I felt that the professor's explanation for why the prediction failed (people decided to keep working lots and buy more stuff instead of chillin) totally unsatisfactory.
"Why history didn't work out this way" seemed a much better basis for understanding the past than post fact theorizing of traditional economics. Lets take theory and its predictions and see why it failed. Often, I think the projected premises turn out correct, but their impacts do not. Demonstrates what we do and don't have a grasp on.
Why not?
Essentially because I think Keynes (and most economists of his era) were right in their basic assumptions. If you made the average person in the 1920 10X richer, they'd have behaved much like the "leisure class" of the time.
If you ask the average person "why don't you work 65% less and chill?" most people would look at you funny. It's not a choice. I have rent, etc. A solution that relies exclusively on choice fails the smell test.
I have my own ideas about the solution, but that's besides the point. My point was that Keynes, not knowing the future, gave a more intellectually honest projection of what the theories predict than any post facto economist would have.
note: no reason to downvote a respectful, on topic comment. Asking "why not" is the whole point of discussions.
I see what you're saying, but at the same time, at what annual revenue level do humans actually start "chilling" more and work less? 500 thousand? I'd be curious to see statistics.
Where I live (Canadian city), a lot of parents are working less, through a variety of arrangements (stay-at-home mom, consulting, part-time work, parental leave). Some of them have pretty low incomes. Society doesn't make it exactly easy, but it's certainly possible, it's just not yet the mainstream option. Social safety nets obviously help.
Wikipedia has this nice chart: https://upload.wikimedia.org/wikipedia/commons/3/3b/Heures_t... . Some countries have seen impressive reductions.
Anyway, I definitely agree with that:
A solution that relies exclusively on choice fails the smell test.
Wealth inequality, etc...
For his part, I believe Keynes asked (poor) people at what salary they would work less and also observed the "leisure class" of his day.
For my part, I think the "cost of thriving index" is a good starting point. For most people it's $X plus the amount needed to they buy a house, car, pay tuitions, healthcare, etc.. If you add up all these "basic" expenses, the "cost of thriving" has far outpaced official inflation and even median salary.
I m optimistic about this , with people who work remotely. There are alot of jobs that dont require a lot more than 3 hours of real effort per day.
Where they do acknowledge it, it's usually via frivolous examples. Your friends have cool cars, so you need one too.
The crisis within that explanation is that the number of people employed in the "service" category has not grown proportionately with gains in productivity. In fact, it's barely grown at all. The type of job that has filled the gap is "administrative/clerical."
This observation has led David Graeber to propose that perhaps Keynes was essentially right, and that the only reason we don't have the shorter work week is basically that we refuse to let it happen. We would rather continue "working" whatever number of hours was customary in our culture (40 hours a week in the US, less in Europe, many more in Asia) and if there's not enough to actually do (there isn't) we have a small number of people do the work for 40 hours and the rest are just managers and administrators sitting around in a redundant meeting or staring at a screen waiting for the next redundant meeting.
Stuff that was simply "out of theory" and irrelevant to it despite seeming logically contradictory to theory. The observations about efficiency in administration in the digital age is particularly striking.
I do think Graeber tended to overstate conclusions. I actually think we understand inflation wrong, and that this becomes relevant over long time scales. Some productivity trends just overestimate reality. We haven't really become as wealthy and efficient as the measures suggest.
Another problem is non fungibility. Not owning a smartphone in 1970 is different from not owning one in 2020 in ways that we can't be quantified in dollars. You can't buy a 1960 education or 1990 healthcare. It isn't sold, and it wouldn't have the same value even if sold. It's way too complicated to dismiss with a "our expectations have gone up."
I disagree with this conclusion. I, for one, never decided to work more, I simply don't have a real choice. No one will hire a 15hr/wk person except in positions that don't pay enough per hour to make a living wage anyway. If I could cut my current salary to 37.5% to only work 15hr/wk I'd do it in a heartbeat.
It’s up to the individual if doing animation work in 6 month bursts for individual movies is enough or if they want near full time employment. 6 month rotations are common in many tiny industries like super yacht crews. Major industries like education, farming, construction, and tourism are quite seasonal and it’s up to individuals if they take that time off or work an off season job.
In a nutshell, Keynes made assumptions about economic growth (correct). He then riffed on assumptions for income effects (if I pay you more, will you work more?) and substitution effects (if you have more money, will you work less?). Those are not very hard to populate. Ask poor people, observe higher income people's behaviours.
The standard retort my professor gave (people decided to keep working lots and buy more stuff instead of chillin)is, IMO, the favoured explanation because it's parsimonious with theory. It fails when it meets reality, in the form of people. Ask a median person why they don't chill more and work less and you'll get strange looks.
If interested, I think concepts in the "cost of thriving index^" are probably key to understanding why economists are (I agree) mostly wrong on this.
In any case, I think the keynes prediction, is an invaluable starting point... being from the past. A modern economist might defend the theory, with the "chill more" defense. I doubt they would bring it up though. No one is going to make a false prediction about the past to test a theory they support.
^ https://media4.manhattan-institute.org/sites/default/files/t...
Microsoft may be making this more available... https://www.theverge.com/2020/10/9/21508964/microsoft-remote...
> Flexible working hours will also be available without manager approval, and employees can also request part-time work hours through their managers.
this raises another problem that keynes may not have forseen (not sure when he made this prediction). people don't necessarily realize how much it costs to employ them beyond just their salary. stuff like health insurance, software licenses, and physical equipment/space doesn't scale with the hours worked by an employee (at least, not linearly). 37.5% of pay for 37.5% of the normal work week might sound good to you, but sounds pretty bad to your employer. they get 37.5% of the working hours for you, but pay well over 37.5% of the original cost. even if given the choice, I highly doubt most people would accept a superlinear pay cut in exchange for decreased hours.
You will see those one time/fee costs do not scale, but to the benefit of the employer.
just get your sales in order and fill a niche
It sounds like you're in a great position.
"Getting one's sales in order" is by far the hardest part, though. I have no idea how to do that.
after a year or so, new jobs will begin to just flow in as you will have built creds. then you essentially can pick and choose.
also learn to think like your buyers. help them to win. dont sell, cooperate. be a partner.
This comment really demonstrates why certain kinds of wrong predictions are very valuable.
Extremely relevant as we move into our second attempt at making WFH work.
This is an interesting "problem". After all we now have more ways to manage data, so we generate more data. And usually this means more administration. The question is, are we getting more out of it?
Also, we see what reduced administration looks like. Every webshop, every online platform where you just click something and you can buy/sell/report/ban.
Aaand, there's the incumbent lobby doing regulatory capture to resist de-administration, eg. filing taxes.
> Harvard is still Harvard.
It took this pandemic to show the world how fragile that status quo is. Plus the recent admission discrimination lawsuits and scandals show that it isn't really about education.
This really is the question, quite literally, in economics. That's basically what economics studies.
The outputs of a car factory are easy to measure. How many cars came out? We can quibble over quality, but quantity is most of the game and that is measurable. Administration is harder to measure.
David Graeber, in his provocateur style, concluded that they are more efficient in the sense that they produce a lot more paperwork... which is their output.
Economists would consider this nonsensical as paperwork itself doesn't benefit anyone and can't be considered an output. He wasn't an economist though.
Or they would observe that wages are paid to produce it and estimate it's worth based on the total wages and total pages of paperwork. It's a very blind method, but it makes a certain sort of sense.
When theorizing though, economists don't consider the paperwork itself an output. If that administrator works for a car company, it's considered to be a contribution towards outputting cars.
Most people just try to escape bureaucracy, so good empirical measurement is about all I currently hope to get from economists.
And both seem to be very much true. After all the economy is not just made of cars, but there are paintings, and festivals, and there are people who pay to travel to other places just to see that place with their own eyes, and it's also charity, and it's also insurance, finance, and other meta-services, and whatever humans prefer more than keeping their money in their wallet (or in ETFs). Add to all this the good old agent-principal problem, and we can quickly see why there are bullshit jobs. Because lot of decision makers (especially when the responsibility is diffused) opt to organize work in a way that helps themselves by helping those who are close to them. (Eg. nepotism, cronyism, but also the watered down versions of them too, when someone just gets a bullshit job.)
Furthermore, it's a testament to how much economic surplus is produced nowadays that so many relatively low-output jobs can be sustained.
Aaand, there's also one other important point considering these seemingly do-nothing paper-pusher jobs. They are (knowledge and know-how) buffers for organizations against fluctuation. They provide stable (hand off) points for more productive more fast moving folks. (Literally, if someone changes positions inside an or between organization(s) they need someone to tell them about the boring details. And since reading all the thousands of documents produced each year is very useless, it's economical to have people keep stuff in their head.)
Could it be done better? Sure. That's what startups try to do, to "disrupt" the old inefficient network of work. But as they grow they too face the same problems (though of course usually on a different level).
To be clear, the vast majority of these people are lying, either to us or themselves. What’s important is not what they actually work, but the fact that claiming to objectively overwork is something that they feel is socially necessary.
I'd argue the culture glorifies being busy in part because the folks benefitting the most from the current arrangement want it to.
And also because of the whole America/Protestant work ethic thing, which isn't all wrong but ...
If we accepted Keynes' era's quality and standard of care for healthcare, entertainment options, transportation options, clothing options (almost wear the same clothes every day), appliance options, food options, and real estate constraints (for most young people who do not already have pledge-able assets today, that means rent or cash-over-the-barrel-no-mortgages), I can see a 15 hour work week. There are very few people who willingly accept those limits and eschew modern improvements. Keynes to my knowledge never said 15 hours with improved technological benefits in all areas of life [1], only mostly the technological improvements that brought about economic productivity enhancements.
Also, there’s a hidden assumption in there that child care doesn’t count as “work”. A child needs 16 hours of care and attention every day and no economic system can reduce that.
That's not why people don't work 15 hour work weeks, it's because we're forced to due to artificially induced scarcity of essentials like housing, along with a competitive labor market where employers generally demand 40+ hour work weeks and can get away with it because they have the leverage.
Actually, this part is regressed back these days. Invention of the smartphone makes many people not to have a computer with efficient input devices: mouse and keyboard. Although smartphone is handy, it makes people dumb. Smartphone makes people a mere consumer of the art, not producing it. Because smartphone is inefficient at input devices but people don't much care about it.
Yes. Majority of the people are consumers but if not for the smartphone, we could have potentially more producers of the art. It's so unfortunate future I didn't expect 25 years ago.
Also people are competitive - maybe 15 hours can pay for food and shelter but then the Joneses next door work 60 and buy a Porsche.
I'm no statistician, but my intuition tells me that if you drop 1/3 accuracy every 10 years, you're shooting well under 10% by 2029, so it seems crazy to say that we're "on track" for that date. This reminds me of the insanity that some project managers use to estimate their completion dates. They say something will take 1 year. Then after 2 months they have done 1 month of work and then update their total estimate to be 13 months instead of 24. And I'd argue that its even worse for Kurzweil's prediction because in that case its an exponential growth factor thats off - not a constant factor.
Want a prediction to sell a book? Singularity by 2029. Want the truth? The singularity won't happen by 2029.
Note: I'm using 10 years here because the predictions from 2020 were not graded, and the last timeframe for graded predictions was ~2015
That’s quite a bit more than TT.
> This is actually quite impressive! Although 9 out of 25 is only a 36% accuracy rate, I still remember when reading The Singularity is Near for the first time that almost all of the predictions seemed wildly optimistic and sort of crazy. It seemed a bit unlikely that it would be possible to have a high speed internet connection from a touchscreen super-computer everyone has in their pocket that can also act as a personal assistant that you can speak to in natural language and it will usually understand and respond appropriately, albeit unable to have a full conversation at this point.
So his best record is in making predictions about the internet and ubiquity of portable computers (smartphones), but he's been less successful in his optimistic predictions for full-immersion virtual reality, and the jury is still out on the singularity.
It's useful to consider obvious misses - for example, he failed to predict fake news or the dystopian surveillance and behavioural influencing culture being created by ad tech.
This has obvious implications for practical AI. If the black-hat possibilities aren't considered and mitigated, experience to date suggests they'll dominate.
In former days, "fake news" was simply named "propaganda" or "bad/manipulative journalism".
In fact, fake news started WWI.
We assumed that only happened with state-controlled media, that one would first take control of the State, before taking over the media to sediment it. We weren't prepared to deal with the Murdochs taking control of the State by controlling part of the media.
The only thing that changed was the distribution method (i.e. over the internet/social media).
> American newspapers fanned the flames of interest in the war by fabricating atrocities which justified intervention in a number of Spanish colonies worldwide.
[1] https://en.m.wikipedia.org/wiki/Propaganda_of_the_Spanish–Am...
To give a particular example, the exponential growth of computing power has been a close-enough match to exponential since Kurzweil became famous: It was about 20 years ago that Apple was boasting that their latest PowerMac was banned from export to Cuba because a 1 gigaflops computer counted as a weapon, now they boast their phones do 11 teraflops — compatible with 1.5 year doubling time.
(That said: while I have met someone who expected the exponential to continue past single atom transistors, I do expect it to stop soon-ish; superconducting/3D/quantum computing may replace classical transistors, but I never expect to see multiple transistors on a single atom).
It works until it doesn't.
Extrapolation outside your dataset is only a problem if you take it too far, and how far is “too far” depends on how well you modelled reality with your dataset in the first place.
But this energy is not infinite. Despite all innovations in solar/wind, and assuming efficient storage that does not exist yet, EU won't be able to absorb a 3% decrease in oil+gas importations by 2030. Rystad (https://www.rystadenergy.com/), pre-covid, was forecasting a 2 to 8% decrease in oil and gas import from the EU according to the think tank "the shift project" [0][1].
I won't talk about manpower much because there is still a chance that nothing more heavy than Covid will hit us in the near future, but meals will be hard to get this year in eastern Africa and while this year India did not suffer as much as in 2019, another drought like 2019 next year would be even more disastrous. For information, this winter could have been worse for Africa than 2011, luckily the wildfires and drought in russia were localized further north than a decade ago.
But well, we still have two silver bullets: fusion and renewable hydrolyse.
[0] https://theshiftproject.org/wp-content/uploads/2020/06/Study... complete data analysis
[1] https://theshiftproject.org/wp-content/uploads/2020/06/fig40... short graph.
A certain type of person is basically wired for it.
I think that the reality is that worldviews are necessarily reductionist at some point. Think about just literally looking down a road far into the distance. If you can see far enough the stuff at the end of that road just looks like a tiny dot which you can't make out.
No one can actually predict the future. Some people like me think that technology will probably shortly change things dramatically.
I would be interested to hear your worldview and predictions for 2029 and the 2030s. But just a warning, I am going to classify those predictions as beliefs.
I think the extreme negative or positive outcomes are both unlikely long-tail events but having a "singularity" that benefits the wellbeing of humanity is definitely worth striving for.
If it even happens it will not be shared equally. A regular person will have to login with a Facebook account with the right amount of social credit to vicariously experience a part of the ascended's experience, with ads. And no, trickle down isn't real.
In general, people do not adhere to the egalitarian message board nerd ethos that all opinions are equally important. The opinions on technological progress of someone who has invented as much technology as Kurzweil carry extra weight.
(I agree that he's wrong about, you know, a ton of stuff...)
For 2019: https://www.lesswrong.com/posts/NcGBmDEe5qXB7dFBF/assessing-...
My own takeaway: Kurzweil's self-assessments are misleading and borderline dishonest; Kurzweil is overoptimistic in general, and you could improve by adjusting by about a decade; there seems to be a split in accuracy to my eyes about hardware vs software - when Kurzweil talks about software such as natural language translation, he is often right or delayed by a relatively short time, but when he talks about hardware (and biology especially) he is strikingly wrong and often the prediction looks like it will never come to pass; and a lot of the software wins were driven by the deep learning revolution: even in 2015, you would have reasonably dismissed some of the 2019 predictions working out, but by the end of 2019...
It is fascinating though, to see what predictions still seem as likely (though less optimistic) when we go from a "don't be evil" perspective to the 2020 one.
Only 2 of them rely on advances in machine learning (real-time translation of foreign languages, personal assistant).
Most of these are consumer tech or interaction with computer advances. Two predictions about education are nice.
All predictions about advances in biotech, medicine, nanotech, etc. have failed.
At the same time we have a small experiment that successfully made 9 people younger using growth hormone therapy, so we know that there are targets that work in human, we just need to decrease the side effects significantly (which can sadly take 20 years realistically, as experiments are extremely slow and expensive compared to running an A/B test on a web site).
There are some other ways in which we accumulate damage, such as scar tissue, and increasing amounts of embedded foreign bodies. But not nearly as important as dna damage.
For this reason, I see crispr as the only path that could lead to immortality, and frozen stem cells as the most likely way to stall until that happens.
That's why is such a big thing that reversing methylation already happened in humans, as it means that the human body already knows how to regenerate itself (just like other animals), we just don't yet know how to turn on the regeneration pathways safely.
> The Singularity is always one Kurzweil in the future.
(That is a pun on Kurzweil's name. It consists of two german words that can be roughly translated to "a short amount of time").
Another quote that this may be based on: "The semantic web is the future of the web, and always will be." (Peter Norvig) Source: https://news.ycombinator.com/item?id=8510953
Neither did I, until I started studying aculeates in detail. Only a few polistid species are even described in Wikipedia, and for the most part their pages share identical content that appears to have been copy-pasted and which, while a reasonable general description of polistid social behavior and colony cycle overall, totally fails to describe heterospecific variations which can be fairly easily found in papers which treat individual species in detail.
For a general lay overview such as an encyclopedia is meant to provide, sure, it's accurate enough. But - given that even an interested amateur like me can not only soon exhaust Wikipedia's treatment of the subject, but almost as quickly also become aware of that treatment's shortcomings - I no longer have any difficulty understanding why it's so negatively regarded as an academic reference.
On reflection, I don't understand why this should be controversial. When is any encyclopedia expected to be able to serve in this role? The genre is, as I mentioned, intended to provide a lay overview of many subjects; its treatments are thus broad and shallow by deliberate design.
https://web.archive.org/web/20201016100027/https://www.antro...
https://www.youtube.com/watch?v=_luhhBkmVQs&feature=emb_logo