« Could it Be? Have We Gotten Through to the Federal Reserve? | Main | ATOM Award of the Month, November 2019 »



The technological singularity is defined as the point where we don't know "what" with any certainty. We'll never know the what till the what happens.

Your prediction method seems reasonable, and it is interesting that as time passes, your predictions remain consistent. But man, I got to live another 40 years to see this? bummer. That would make me 96 at the singularity - I just might squeak it out, if I'm lucky.

One problem I see is we always assume a smooth the flow of technology. History suggest it advances in lurching surges. Much like the punctuated equilibrium of evolution. It only looks smooth over centuries or decades. And I suspect, the flow will get even more turbulent as we get closer to the singularity.

One item - Moore's law is ending shortly. It will be interesting to see the rate of acceleration after this has happened. My guess is a period of stagnation, not lasting more than 10 years. Then the next surge upward. It'll be okay, because it will focus people on cleaning up software to make it run better and faster - something I think has lagged as the software developers chased the rapidly improving architecture.

Kartik Gada


One problem I see is we always assume a smooth the flow of technology. History suggest it advances in lurching surges.

True, as each graph is a trendline, not the actual progress.

That said, it is smoother than that, since different sectors advance at different times. In terms of dollar size, Photovoltaics and EVs are the two biggest disruptions in the global economy at present. Someone working in 3D Printing or VR at the moment might feel as nothing is advancing, when sectors they are not exposed to are advancing. There is also a long list of technologies that went nowhere in relation to initial hype.

One item - Moore's law is ending shortly. It will be interesting to see the rate of acceleration after this has happened.

Note that Moore's Law is just for transistor shrinkage. The commoditization of parallel computing (through NVidia's GPU) allowed new forms of computing to happen on the same transistor sizes. AI, for one, runs 20 to 50 times faster on parallel computing vs. serial computing. That is why AI, dormant for ages, came from nowhere to something everyone talks about in just the last 3 years.

So the advancement of Computing can decouple from Moore's Law for a few years.


So do you believe you will reach actuarial escape velocity on your own life? That is to say, life extension technologies advance faster than you age?

That is another area where an immensely disruptive surge might just happen. Imagine life extension to 150 or 200 - but only for certain people. It could tear society apart.

Moore's law is funny - we advanced through a very simple mechanism that was easily understood. Even Moore's laws is being disrupted! No we will advance via a variety of pathways - optical computing, quantum computing, massive parallelism, more specialized chips, better programming....even the easily understandably rate of increase becomes complex.

Kartik Gada


So do you believe you will reach actuarial escape velocity on your own life? That is to say, life extension technologies advance faster than you age?

I do not think that is happening anytime soon, in fact. Human life extension is insufficiently correlated to the furtherance of overall technological progress, so this is unlikely to happen.

Medical technologies (mostly of a preventative and early-diagnostic nature) may extend human life to 100, but that is it. Even this may already be too unattainable to people who are already over 40.

I will change my tune when I see a distinct and sudden rise in the percentage of people over 100 who manage to reach 110 (currently very low; under 0.1% of people who cross 100 reach 110).

What may happen, however, is the mapping of a person's entire memory and personality (or a better version of their personality) into a VR avatar, for permanent preservation. This could be possible as soon as 2040. I do think a person's entire personality can be simulated from data collected from tests/questionaires that take under a day to complete, combined with consent to use data about their web search history, etc.


I’m ecstatic that the diffusion rate has already accelerated this much. I can sit on my couch and read great thinking on prediction walls in almost real time, directly from great minds like Kartik’s, and I don’t mean that as a joke.

Perhaps just a technicality but singularity technically means a point of infinity. Kartik seems to use a different definition whereby singularity is the point where “the rate of technological change becomes human-surpassing”? The two definitions may be close in time but are they the same thing?

I understand the bias of singularity optimists hoping the singularity arrives in their own lifetime. I also wonder though if there’s a bias in pushing the singularity past one’s lifetime out of fear of being proven wrong during your lifetime in case the singularity happens much later.

Quantifying and following the length of the “Prediction Wall” is a brilliant way of timing the singularity. One of the challenges in this approach is estimating more exactly the instantaneous length of the prediction wall, as well as what percentage of the time to the singularity the prediction wall represents. Seems to me that the singularity timing estimate should be quite sensitive to changes in the Prediction Wall’s time placement.

For example, I’m wondering... what percentage increase in total world knowledge (or perhaps capability) does the prediction wall represent? If eg. we assume it represents a doubling then the length of the prediction wall keeps halving. In that case the time to the singularity at any point in time is double the time to the current Perdiction Wall, as the series 1/2+1/4+1/8... converges to 1.

Another singularity related general concept that has been in my mind is whether we will reach “Actuarial Escape Velocity” before the singularity. I’m thinking that the process may actually be discontinuous. Much like flight. Prior to 1903 “flying like a bird” was impossible at any price and held the same mythically impossible status as the proverbial fountain of youth. Then within twenty years long flights were available to the wealthy — and that was in the nineteen twenties when the pace of change was much slower. The advent of flight was a rather abrupt event. If a similarly abrupt transition in Actuarial Escape Velocity were to occur today, the whole thing may unfold in five years or so. We could thus go from nearly stagnant lifespans to virtual immortality in just a decade. Immortality seems like an impossible barrier. However, an indication of how this may happen is similar to flight. As with flight we see that nature (evolution, deity or whatever one believes) already knows the secret to longevity. That is why there are moths that live a day and species that live much longer than humans. However, as with flight, nature is not interested in endowing every species with longevity. Nature is interested in overall capability and success of the species, not individuals. So if earth bound rabbits are just as successful as flying crows then be it. The rabbits remain earth bound. Similarly with longevity. If a moth that lives just hours is as successful as a bowhead whale that lives over two hundred years then be it. The survival of the species is the relevant metric of success.

Obviously, just like with flight, humans will not settle for their biological destiny. Longevity could be controlled through something relatively simple, like the telomeres in our DNA. Something simple that, again, nature has no reason to change, but which we humans might be able to change, like placing wings on machines and flying, breaking free of our biological destiny. I’m just mentioning telomeres as an example. The aging mechanism may turn out to be much more complicated or perhaps even entirely different. But the chance that the aging mechanism itself may be abruptly defeated, just like human earthboundness was rather suddenly defeated by flight, is not insignificant. As a matter of fact it may already be before the prediction wall.

In other words, there are indications that limited lifespans are not some sort of hard cosmological destiny like entropy or the speed of light. We may think of it as an insurmountable barrier but that is because we are used to thinking of longevity in terms of technological evolution, not disruption. Like eg. through gradual elimination of disease and patching up old people with laboratory organs. But a disruptor is something that changes the paradigm. When it comes to longevity, interfering with the aging process itself seems to me already on the horizon, before the prediction wall. That’s what a major disruptor is. Something that will surprise and shock everyone.

“Imagine life extension to 150 or 200 - but only for certain people. It could tear society apart.”

It would also teach hippies the same lesson all over again: That today’s luxuries become tomorrow’s necessities. And they too will want them.

Kartik Gada


I’m thinking that the process may actually be discontinuous. Much like flight. Prior to 1903 “flying like a bird” was impossible at any price and held the same mythically impossible status as the proverbial fountain of youth. Then within twenty years long flights were available to the wealthy — and that was in the nineteen twenties when the pace of change was much slower. The advent of flight was a rather abrupt event. If a similarly abrupt transition in Actuarial Escape Velocity were to occur today, the whole thing may unfold in five years or so. We could thus go from nearly stagnant lifespans to virtual immortality in just a decade.

Perhaps. But consider that the people who have being crowing about this for 20 years have zero progress to show for it. Ray Kurzweil, in 1999, predicted that in 2019 average human life expectancy would be 100. In reality, it has hardly increased at all. AEV seemed to be at 20%, but that proved to be a false dawn, and it has reverted back to a noise floor insufficiently far from zero.

In addition to the conversion of 100+ people to 110+, other metrics to look for are whether they can extend a primate's life by 30% or more. They have been tinkering with mice for 20 years, but with limited success.

All of the major researchers and evangelists for this field (Kurzweil, Sonia Arrison, Aubrey de Grey, George Church, Dave Gobel) got a lot of press a decade ago, but now are nowhere to be found.

I will change my tune if any of the aforementioned metrics sees a sudden and indisputable improvement.

Now, it is true that cancer treatments have advanced quickly, and more importantly, cancer detection technologies are finding cancers earlier and earlier (in a true ATOM sense, the technological progress is more on early detection than curing people who are already Stage 4; an ounce of prevention is a pound of cure). This edges longevity up slightly, but nothing that is a major step towards AEV.

The survival of the species is the relevant metric of success.

This is true, and far more relevant than the extension of individual lifespans.

Human extinction is all but impossible now, and total recovery (of the economy, if not numbers of people) from even a 99.9% population loss would take very little time given how much knowledge has been accrued and preserved.

Even the asteroid-impact threat is exaggerated, and is rapidly being minimized through a number of ATOM advancements. Fewer and fewer objects larger than 200m remain undetected and untracked, and there is at the most one such object that needs to be diverted between now and 2050.


I will change my tune when I see a distinct and sudden rise in the percentage of people over 100 who manage to reach 110 (currently very low; under 0.1% of people who cross 100 reach 110).

That's a very robust approach. However, it requires already a massive dissemination and use along the population. In my early forties I care more to reach 100 in reasonable health than the actual chances of AEV/LEV, and worry about that much later ;)

What may happen, however, is the mapping of a person's entire memory and personality (or a better version of their personality) into a VR avatar, for permanent preservation. This could be possible as soon as 2040. I do think a person's entire personality can be simulated from data collected from tests/questionaires that take under a day to complete, combined with consent to use data about their web search history, etc.
Creating a convincing simulation might be possible but that might be quite shallow. I think that mastering mind uploading will be orders of magnitude more harder than keeping the body in good health (except the brain , which might require much deeper understanding) . The moment we can have millimeter-size medical bots to do micro surgeries and, detection and prevention in the bloodstream a lot of diseases and conditions could be miproved significantly : remove plaque buildup, to vastly improve the cardio-vascular health (good for the brain too), remove micro tumors and lesions even before they turn cancerous, targeted delivery of medication, etc..


This all presupposes that present trends continue indefinetely, which they almost never do. I'd predict civilisational fragmentation and or collapse before any tech singularity is reached


"...but singularity technically means a point of infinity."

Technically true, but not precisely. In this instance the term is being used to delineate where technological growth becomes uncontrollable and irreversible, resulting in unfathomable changes to human civilization. Vernor Vinge popularized it's use - he liked it to the edge of a black hole, where all the laws and rules begin to break down.

He believed in 1993 that the creation of greater-than-human intelligence would occur before 2030. That seems unlikely at this point. He imagined it as a great "surprise". It would seem far off, then suddenly it would be here.

Vinge actually credited the term to a discussion in the 1950s between Stan Ulam who paraphrased John von Neumann:

"One conversation centered on the ever-accelerating progress of technology and changes in the mode of human life, which gives the appearance of approaching some essential singularity in the history of the race beyond which human affairs, as we know them, could not continue."


Just to ground us, I’m assuming we are using the definition of singularity that we are using is either artificial intelligence that is more capable than a human and continuing to improve or technology increasing faster than we can understand. It’s interesting that KG I skeptical about significant longevity improvements, because without a positive impact on human transformation (extended lives, enhancements) in already rich countries the route to the singularity may come across as economic turmoil more than anything else.
I finally got around to reading Robert Gordon’s _The Rise and Fall of American Growth_ about the transformation and economic growth in America from 1870-1970. He describes the great changes in American lives in that time period and the relative lack of change since then, in that the radical improvements since then have been in information technology and related improvements in entertainment, but not in other areas.
That did resonate with me as I compare my life growing up on a midwestern farm after being born in 1966 with that of my grandparents who were born in the 1890s or 1900s and grew up on farms as well. For example, while the upcoming change to electrical vehicles with automated piloting will be huge, it will not be as big of improvement to lives as the change from horses to internal combustion vehicles. My smart phone and its supporting system may be a huge technological advance over land phones, tvs and radios, but once again, not as big of an impact as going from not having those technologies at all to having them.
My label for these earlier changes would be “Verneian” in that they were predicted by Jules Verne and others because they were straight forward enhancements to what people already had – having running water in the home in the temperature desired, rooms being the temperature desired, easily available light and food, etc.
We may be 3% to the eventual technology and while I will certainly enjoy the delivery of inexpensive food with the quality of 5 Michelin stars exactly when I want it, it’s not that great of change to human capabilities beyond having a freezer and microwave. Especially when you compare the freezer and microwave to having a wood stove and no cold storage for food.
What I am curious about is where the technology shifts things to a transformative difference – such as the cheaper computing power allowing rockets to be bundled together to lower the cost of space flight that may then make other things possible. My rule of thumb is that if we are talking only cost, a 10x reduction is when I will consider it transformative. I understand automation throughout all parts of production cycle may lead to that reduction in cost in other areas, but we haven’t it yet, outside IT and entertainment. I also understand that human factors can also slow this – such as we see in education - but I’m still waiting for those changes to happen that overcome those human factors.
Apologies for the long post. Drew

Kartik Gada

Ragnarok said :

This all presupposes that present trends continue indefinetely, which they almost never do. I'd predict civilisational fragmentation and or collapse before any tech singularity is reached.

It does not presuppose that. Technology progresses even in the face of nation-state turnover.

Note that the evolutionary trend of ever-rising intelligence was entirely uninterrupted even through the two biggest extinction events ever (252 million years ago and 66 million years ago).

Kartik Gada


I come across the Robert Gordon point from time to time. The flaw there is that he assumes innovation is binary, and ignores subsequent product improvements as well as ever-rising speeds of diffusion. Secondly, his criteria are US-centric.

For one thing, sure, early automobiles were like the Model T, but continuous improvement relative to cost has been immense. 2019 cars are vastly better than the Model T by any engineering metric. Similarly, while light bulbs were invented in 1879, they did not become inexpensive until the 1980s, and now LEDs are a more advanced technology that improves at an exponential, Moore's Law-type rate (Haitz's Law). The price of passenger airline seats continues to fall. Even ball-point pens were pretty expensive until the 1970s.

The term 'Verneian' is a good one, but remember that the invention of the first market-ready product is only a fraction of the story. How quickly does it diffuse to 25% of the market (i.e. the *world* market)? The speed of market penetration is an extremely important part of measuring any sort of technological progress.

Smartphones went from zero to 2 billion in a short time. China and India both generate a greater percentage of their electricity from Solar than the US does. Only 10-12% of EV sales in the world are in the US. China builds skyscrapers at about 6 times the speed as the US. India has wireless data rates that are the least expensive in the world.


All log scale improvements are really S curves. Or step graphs. The real question is where is the inflection point of the curve? Where is the next leap of punctuated equilibrium?

My own theory, based on exactly nothing, is society is like a snake swallowing a rat. It can't eat another rat until the first one has been fully digested, and the bigger the rat, the longer it takes. And it has been eating a lot of really big rats lately.

Think about this - No one in 2007 was predicting the U.S. would be the largest oil producer in the world and exporting LNG in 12 years - heck, we were busy building LNG import terminals. Recall in 2008 when Obama tritely said we couldn't drill our way out of our energy problems? Yeah - he was as absolutely wrong as it is possible to be wrong. So many elites, so much accepted knowledge, all of it wrong. Most people don't see that as a revolution but it is - fracking has changed the entire geopolitical landscape, changed how we view energy and the future, changed our finances. Next to that cell phones look like a joke.

So what is the NEXT big scary society altering thing that people are not aware of?

I keep saying something like a plasma/laser drill bit could be bigger than fracking. It would reduce drilling costs by a full order of magnitude, and allow geothermal power to be built anywhere in the world.

How about robot lawyers, followed soon by robot judges? DoNotPay is a Facebook Messenger bot that helps people challenge parking fines, gain compensation for delayed flights or apply for asylum, among other tasks. California startup HelpSelfLegal automates simple legal tasks like clearing marijuana convictions or taking out restraining orders. Legal Utopia helps consumers understand legal problems by explaining their issue in plain language, without jargon.

How about dynamic pricing for everything? Hotels, cars, airlines...computers are constantly assessing the exact price that is optimal for each consumer to buy a product, and the price of everything changes...by the hour...but it will be fine because your shopping computer will always be prepared to purchase exactly what you need the moment it is cheapest - think of the stock market as coming to every retail store in America.....

Imagine electric hybrid airliners that cost half as much to operate. Given more than half the cost of you ticket is fuel, that might be a 30% decline in the cost of flying....what does that do for tourism? Hotels? Car rentals?

Anyway, it is the left field stuff that fascinates and frightens me. the changing and undermining of things long accepted as "fact".

As I've said many time, the problem with futurism is we always imagine our future as the same as now, but with one thing different. but that is not how the future works - it is many things different, and those things interacting with one another in weird ways.


My own recent experience - my company has decided to do away with private offices altogether - from the CEO down to the lowest staff member. We all work in cubes. no storage, no desk space. paperless office. No phone - everything works through the computer. All calls are transcribed and digitally stored automatically. We have luxuriously appointed Team rooms will wall size screens we can wirelessly hook to our laptops. Software we need is automatically purchased and uploaded. Working from home is encouraged and seamless. All our IT needs are met remotely...from the Philippines where staffers are on duty 24/7..

This happened three weeks ago....and it's not terrible. It is a radically different way of working, but it has certain advantages. In some ways it's fun.

Kartik Gada


The quantity of technological disruption going on at any given time is measurable and predictable, under ATOM principles. What is harder to predict is where it is next, and whether the next phase will be a dozen small disruptions in various industries, or one huge disruption (which requires the entire capacity of the ATOM for the period of disruption).

The fracking disruption did occupy a huge fraction of the ATOM for a brief period (during which very few other technologies advanced). We may not see such a large percentage of the ATOM occupied in this manner again, given that the ATOM itself rises in size exponentially. The wholesale disruption of government is one of the only things that is so large that it could occupy the entire ATOM for a multi-year period.


When you say "large percentage of the ATOM occupied" it sounds like you are treating ATOM as a person, I think what you mean is that is where a lot of capital, both human and financial was focused because there was a straight forward path for reward. Sort of like for job searching, it is recommended that you focus on what you do well and what is in demand (bonus if you enjoy it). Oil (transportation structure allowing) does have a straightforward way to get returns, so you would see "ATOM" focusing on similar commodities, or items that can become commodities (Amazon Web Services for computer services). And of course advertising has a straight-forward way to get paid, so an economy focused on attention gathering, commodities and making services into commodities. While there is plenty of money sloshing around medicine and education, access to it outside the established channels is difficult. However, once an opportunity like fracking shows up in those areas with a way to get paid for it, then a tidal wave of human and financial capital will come in and there is definitely a lot of investments in both supporting roles in those areas as well as looking for that big opportunity. Thanks, Drew


Hi Kartik,
Recently I have stumbled upon and now finally manged to read an interesting book by Eric Helland and Alexander Tabarrok titled "WHY ARE THE PRICES
SO DAMN HIGH?". There it is shown that medical and education prices go up consistently during many decades and various countries chiefly because they are not automated like other industries, yet require a lot of qualified human effort. And they said imply that the overhead of regulation, unions, and other inefficiencies while high, don't explain the trend of consistent and steady price rise. Their explanation is the Beaumont effect. Since other industries are more efficient they can afford the top educated talent and raise their wages. On the other hand, sectors which are less efficient need to offer higher pay to match the opportunity cost of people choosing to become teachers, lawyers or doctors instead of traders, programmers and so on. They also mention that there are possible, if not looming , technological improvements that can reduce the costs in education and service sectors. The most important thought, though, is that the rising costs of the "stagnant" industries are a sign of general, yet, uneven improvement the economy. The book is available online for free and is not that long. It add another dimension to your ATOM analysis. By some of the wording it seems that the authors are familiar with, but do not ouvertly subscribe to, the concept of accelerating technological change.

Kartik Gada


The ATOM is of a measurable (but ever-rising size) and a certain percentage of the world economy. All ongoing technological disruptions are interconnected.

For this reason, the quantity of disruption to happen in the near future, combined against places where disruption seems overdue on account of cost of incumbents relative to innovation happening (such as Oil was in 2011 at the time of my prediction) enables some ability to predict future disruptions.

Kartik Gada


Remember that healthcare is a much higher percentage of GDP in the US than in most other advanced economies. This is partly due to the bias towards extending life for a few unhappy months at great expense. In America, half of a person's lifetime healthcare expenditure is in the final 14 months of life.

But it is also true, as per the Beaumont effect, that there is an emotional component to both healthcare and education that keeps their demand unshakeable, and hence productivity improvements low above and beyond government entanglement in the industries. People want to extend life by three months because the costs are 'paid for by insurance or Medicare'. People don't question whether college is worth the cost because 'you have to go'. We just saw a bust where people were paying bribes of $500K to get their children in USC (not even Stanford or Harvard). Private Schools are more about bragging rights for parents than educational outcomes for children.

The rising costs of these industries are a sign of general prosperity, yes, mainly as an indicator that discretionary spending is sufficient to keep their existing cost structures afloat.


There is a certain sunk capital argument. We have invested an enormous amount of money into the current system of colleges and medical care. For real change to happen, those assets will need to be written down. The only thing that causes that write down to happen is if the potential savings/benefits greatly exceeds the value of the write down.

Internal write downs happen when the institution realizes the need to change. But the benefits have to be huge, and accrue to the people making the change.

Externally forced write-downs require an upstart provider. Expect the entrenched provider, the person who will get the write down, but not the benefits, to fight tooth and nail.

There is no reason in the world why, let's say Hilton, couldn't invent Airbnb. They have the booking software for their hotels, experience managing guests, and name recognition. AirBNB is worth $31 billion. Hilton is worth maybe $30 billion. Hilton could have doubled their value at the cost of literally pennies. Why didn't they? Well Airbnb undercuts the value of their own brand, their own assets. It might force a write down on the value of their traditional hotels. So instead they fought it.

Netflix realized it needed to change - that the future was streaming. And with streaming, they would need to own content, since the delivery mechanism was so cheap all the providers could eventually not need Netflix. But the market as content provider was huge! As the joke goes, Netflix needed to become HBO before HBO could become Netflix. Jury is still out on if they manged to pull it off. probably not, but it was a nice try.

When you look at medical costs, why? Well the there is the physical infrastructure. But there is the cost of all those doctors and nurses with all those degrees. And all the fancy equipment they bought. And labs. etc etc. The entrenched capital is enormous.

The revolution in medicine will happen when smart people are able to become doctors and nurses much faster, at higher numbers, and at lower cost. Once a lot of "cheap" doctors enter the field, you'll see price cuts to get patients. And they can afford to cut prices because their own costs are lower (less than 7 years of differed income and <$200,000 in college debt!). So my bet is the robo-doc system, where a lot of the diagnostic duties are performed by computers asking questions and collecting data. People will realize you can open a basic clinic with a few nurses to manage the situation and a bunch of robo-docs doing the work. And just refer the 20% of patients with serious problems to regular hospitals. Just get out of the idea you have to treat everyone for everything at one location.

Make it a walk in visit for $20.

With colleges the write down needed is prestige. Look, most employers want a degree not because they expect you to know anything (you won't). What the degree confers is "I am certified as a person who will work reasonably hard and finish what I start". It just says "I am trainable". We don't have any other mechanism to signal that to an employer, and until that is developed, nothing will change.

Think about the hiring process - the only independent measure of your worth is the college degree. BS Geology from Arizona state? They have a good program. You needed to work hard to complete it. You're hired! I want someone smart, but also someone who will stick with it and work hard. Only college gives me that answer. Which is why they have been resistant to change. On-line colleges are smart...if I actually cared about your degree, which I don't, not really. Doesn't tell me a damn thing about whether your are worth anything as an employee.

An alternative to college might be a series of low paying apprenticeships coupled with an intelligence test.

Anyway, if you drill down, you can see why some sector have been resistant to change, while others have surged ahead.


I just looked something up - the robo-doc thing? yeah, that is pretty much already CVS and Walgreen's long term plan. They are already implementing it. News this week that CVS Health is expanding its new "HealthHub" concept to 1,500 stores across the U.S. within three years provides clarity to the drugstore chain’s plan to fill unused and underperforming retail space with more healthcare services, medical supplies and equipment.

"CVS earlier this year began to pilot three “HealthHUB” locations in Houston that dedicate more than 20% of the store to health services, adding additional services at its MinuteClinics in the health hub stores. New CVS employees in the HealthHUB stores will include a “health concierge” and a dietician, executives said this week."

"Walgreens in April announced a deal with the primary care provider VillageMD, which is developing clinics next to Walgreens in Houston as a precursor to a potential national rollout. And in the Kansas City market, Walgreens has a joint venture with the health insurer Humana to develop senior health clinics inside the drugstores.

For the last two years, such partnerships have been testing concepts and piloting healthcare services and technology to develop what executives have called the "drugstore of the future."

I'll get my medical care at the same place I buy a Slurpee.

Note this competition is not coming from hospitals or doctor groups...but from independent pharmacies who have nothing that needs writing down.

Kartik Gada


An alternative to college might be a series of low paying apprenticeships coupled with an intelligence test.

I think a certain certification of completion from Coursera or similar MOOC site would suffice. Plus, AI for hiring matching is a must, since the current interview process is a horrendously inefficient and ineffective screening tool (despite three rounds of interviews, only half of hired people are considered good hires).

For the Robo-doc, the fact that it has an AI element and is in the cloud means that it learns much faster than any human doc, and learns from ALL the patients seen by all the cloud-connected interfaces. The Walgreen's Robo-Doc might see 2000 patients in a day across all stores. Hence, the speed it trains at necessarily becomes orders of magnitude faster.


We "know" the time of it, which just don't know what it is, or what it'll look like

You guys crack me up sometimes.

Kartik Gada


Do you have anything useful to contribute?

The Singularity, by its nature, happens whether humans want it to or not. You seem to think humans can control everything, and can decide when to start or re-start technological progress.


Singularity in 2062?

Meanwhile half the nation’s politicians supported by most media, and apparently a rising majority of Americans, held a seven hour marathon public session on post-singularity climate prediction. Apparently they predicted not only the climate post AD 2062 (the easy part of the prediction actually), but also simulated and predicted... ...the state of post singularity humanity — and concluded that: We, short lived and extremely poor (by future standards) citizens, must make big sacrifices to ease the suffering in the apparently wretched lives of our post-singularity descendants...

Has the world gone crazy? En masse?

Kartik Gada


Has the world gone crazy? En masse?

Well, the nation-state and media have too much power and a lot of people just want to be told what to think. Obstruction of ATOM progress will lead to those structures being bypassed or toppled head-on. Note that China, for all its other flaws, does not waste time or resources on such schemes. Their sole focus is economic and technological progress.

In reality, Photovoltaic energy is growing quickly and is already 3% or world electricity consumption. EVs are 3.5% of world automobile sales. Technology is advancing in spite of various schemes to expand government power under the excuse of concern for the climate (which ironically obstructs technology that actually helps the CO2 situation).



Indeed, at first glance a statement like “we know the 'when' of the Singularity.  We just don't know the 'what', nor can we with any certainty” seems silly. But in many cases, including this one, it is not.

Compare the singularity to a celestial black hole, which BTW is pretty much standard science these days. A black hole is a space time singularity and many have been identified in the universe, including at the center of our very own galaxy. If you were approaching a black hole you would know when you’d fall in (at least in the external observer’s time frame of reference), but you would not know what would happen once you fell in. This is because the laws of physics as we know them break down inside a black hole. Similarly with the ATOM singularity one might know when it will happen but not what happens when it arrives. As in the case of the laws of physics breaking down inside the black hole space time singularity, the realities of our human world as we know them break down at the technological singularity.

Earlier I asked whether by ATOM singularity one typically means the singularity itself or what is essentially the equivalent of the event horizon in a black hole. It was more of a rhetorical question since the two events are not that far in time (though they may be from the traveller’s perspective, if the ATOM singularity has attributes similar to the black hole space time singularity where strange and poorly understood things happen to time and space past the event horizon).

So, in the case of the space time black hole singularity, yes, you would know the 'when' (with precision actually) you will enter the Singularity.  You just won’t know the 'what', nor could you with any certainty using current physics.

In fact, there are other prediction similarities. In the case of a celestial black hole singularity you would know from your speed and acceleration gradient that there’s a black hole singularity ahead of you, how far it is, and thus when you would reach it. Similarly, from our current technological speed and acceleration gradient we can speculate that there’s a technological singularity coming and estimate its time frame. There’s quite a bit more uncertainty because, unlike with celestial science, we cannot estimate our technological velocity that well and, perhaps more importantly, we cannot quite estimate the technological acceleration profile in the future.

For the record, I’m still a bit of a skeptic whether technological progress will culminate in a true singularity, simply because I think it would be prudent to leave a small window of doubt for something that has never happened before. However, most of the things Kartik presents and their implications apply even under accelerating exponential change which does not quite culminate in a singularity, or even under simply exponential change indeed.

For example, the fact that exponentially growing technologies will occupy an ever larger percentage of the economy is true even if one assumes a constant four percent annual growth without acceleration in the exponent. Indeed a four percent annual growth is an exponent itself. However, just like we went from less than 0.1% growth in the Middle Ages to say 0.5% in the Renaissance, then reaching 1-2% by the nineteenth century, 2-3% in the twentieth, and now around 4% in the early twenty first century (whatever the exact numbers might be), there’s clear evidence of an accelerating exponent. But even in the odd case where growth acceleration stops and we stay at 4% growth long term, our wealth and technological capabilities in the next century increase 50 fold (1.04^100=50.5=5050%). But, as I said, even that is a very defeatist scenario as all indications are that the growth exponent itself is accelerating. This is more than the second derivative of total wealth being positive, as any exponential, even a constant exponent exponential, has a positive second derivative.


Nicely stated HB

I know I have a birthday coming up, but I have no idea what gifts I might receive. Maybe it will pass without remark by my family. But likely not. We know all sorts of things that are likely to happen, and have little understanding regarding their shape and form.

Seems like the climate will change. Will Miami flood? Who knows?

And the singularity is not as puzzling as it might seem - we simply posit that computers will become fast and sophisticated enough that they can design and improve themselves. really that they will replicate us. We are not trying to make something new, but copy something we already have billions of examples of. And naturally, if we are picking a human, why not Albert Einstein? So we can manufacture identical Einsteins. Millions of them. And given electronic speeds, even if they are no MORE intelligent than AE, poof, singularity.

It's hard to see how it doesn't happen.

Kartik Gada


For example, the fact that exponentially growing technologies will occupy an ever larger percentage of the economy is true even if one assumes a constant four percent annual growth without acceleration in the exponent. Indeed a four percent annual growth is an exponent itself. However, just like we went from less than 0.1% growth in the Middle Ages to say 0.5% in the Renaissance, then reaching 1-2% by the nineteenth century, 2-3% in the twentieth, and now around 4% in the early twenty first century (whatever the exact numbers might be), there’s clear evidence of an accelerating exponent. But even in the odd case where growth acceleration stops and we stay at 4% growth long term, our wealth and technological capabilities in the next century increase 50 fold (1.04^100=50.5=5050%). But, as I said, even that is a very defeatist scenario as all indications are that the growth exponent itself is accelerating. This is more than the second derivative of total wealth being positive, as any exponential, even a constant exponent exponential, has a positive second derivative.

This paragraph is very good.

I have long thought that asset prices are now a better indicator of economic growth than GDP (even NGDP), since both the positive and negative details of economic progress are captured more accurately.

Asset growth has decoupled from GDP in wealthy countries, since people save a large portion of their income. Before the modern era, this was not possible for many people due to necessities consuming the entire income. Japan is an extreme example of an entire country reaching this point. It is old and prosperous, so the collective net worth of the Japanese citizenry is rising even as GDP growth is not. A large portion of the population are savers.

On the flip side, since assets are more concentrated than income, issues such as uneven distribution of progress are captured more accurately.

GDP growth is a byproduct from an era where almost everyone had to spend all of their income.

The good news is that the exponential, accelerating rate of progress is fully evident in non-inflationary asset price growth.


Geoman: “Seems like the climate will change. Will Miami flood? Who knows?”

Miami will likely flood on current trajectory, ... just like the world today would have been devoid of telecommunications if we kept the trajectory of burning logs in our fireplaces at the rate we did at the turn of the twentieth century ... having burned all the available telegraph poles by 2019. The population growth and firewood utilization increase models of the time would have predicted that. Modeling of those trendlines was within the capabilities and scope of the science of 1900. The “science” and scientific consensus of the time would have thus been correct in warning us and its calculations would have been more or less factual. What was not in scope was the futurology of predicting the human condition in 2019. Telegraph poles are completely out of context in our days. And even if they were still somehow absurdly in context we would have dozens of ways to replace them. That is the same myopic mistake made by climate activists (and most of their detractors too), because the issue is not about science. It is about the futurology of predicting the human condition in eighty to a hundred years; an utterly foolish endeavor. To inconvenience the relatively wretched lives of today’s pre-singularity humans to ensure availability of “telegraph poles” for jubilant post-singularity individuals.

When presented with exponential technological change climate activists react like Palamas wondering how can we be sure that technology will reach stellar levels in the next two generations when we cannot predict specific technologies in twenty years. But this is like ridiculing climate prediction science six decades into the future when we cannot predict the weather in six weeks. Weather and climate are somewhat different concepts. So is prediction of more specific technologies vs the general technological progress trendline. But, as I said, the issue of climate activism is not about climate science or climate model science. It is about the futurology of the human race five and ten decades into the future. Even if one takes a pessimistic view of the singularity or simply rising exponential growth, it seems silly to worry about climate change. We will face much direr catastrophes under a pessimistic singularity scenario. If you want something dire to worry about then worry about nuclear proliferation which seems rather unstoppable and for which we don’t seem to have any mitigating factor in the short to medium term.

I believe in a positive singularity scenario but predicting the trajectory of its unfolding is a fool’s errand. Climate doomsday scenarios are as silly as doomsday scenarios in 1900 discouraging wood burning, in houses that are sixty degrees in the winter despite the fireplace burning all the time, to save us from a future fate of no telegraph poles and thus no communications in 2019. Fools who self righteously declare themselves “progressive minds”, that is, thinking about progress. Yes, sure.


HB, Speaking of great paragraphs, your first one is a doozy.

I agree wholeheartedly. I really hate the lazy sort of futurism you so often see - just extend the trend line out! Just like today but even more so!

At the end of the day the future is driven more by economics more than anything else. Petroleum didn't displace whale oil because people wanted to save the whales, it displaced it because the cost of whale oil was too high, and there was not enough to meet demand. Money drives invention. Figure out the economics of things, and it is relatively easy to predict what will happen next.

We don't "run out" of things. Once what was once abundant and cheap becomes rare and precious, the hunt is on for a new source or substitute. And the longer the shortage goes on, the stronger the effort to fill the void is. Running out of materiel things is like saying the earth will someday run out of atoms. It's just not possible.

We used wood for telephone poles because wood was cheap and plentiful. When wood became more expensive we found substitutes.

Or look at examples of silk, spices, and salt. Rare and precious till suddenly...not.

Atoms and energy and time - everything is made of atoms and energy and time. And time can be substituted for energy, and energy can substitute for atoms, and so on. You get to change the ratio.

Want more oil? Oil is a creation of time, atoms, and energy. I can reduce the time, but only at the cost of energy - I can build a nuclear power plant, suck carbon dioxide out of the air, and make all the fuel you want. What would have taken millions of years can be accomplished in weeks and months. But I have to pay an energy price.

The oil shortages of the 1970, or the early 2000s drove a massive effort to find substitutes. Both eras of shortages were immediately followed by eras of incredibly cheap energy. So cheap the hunt for substitutes slowed, allowing another shortage to develop. The ATOM attacked the problem from multiple directions - exploration, pipelines, fracking, electric cars, solar panels....until the problem was defeated.

The inflation adjusted average price of oil is $41.70/bbl. That is since 1946. It spiked in 1980 to $116/bbl, the dropped to $38/bbl. in 2008 it spiked to $136/bbl. Now it is right back to the long term average. 70 years of extracting oil...and the average prices inflation adjusted has changed very little. When the price spikes the ATOM focuses on the problem...causing the next crash.

And the thing is, the ATOM grows every stronger every year. The resources that can be directed at a problem become ever more powerful. What once took a generation to solve takes a decade, what once took a decade takes an app on your phone.

So when it comes to climate change, especially problems we might have in 50 or 100 years, you have it exactly right. Like the oil shortages of the 1970s, or whale oil, or telephone poles, it will all seem a little bizarre and laughable in retrospect.

And we never run out of oil before that happens. In fact we will leave a substantial fraction of the earth's total oil supply in the ground, just like we left whales in the sea, and trees in the forest.


Pepper was commonly used as a monetary source between 600 and 1200 AD. Eastern Europeans paid 10 pounds of pepper in order to gain access to trading with London merchants. Throughout Europe, peppercorns were accepted as a substitute for money (some landlords would get paid as a “peppercorn rent”). Peppercorns, counted out one by one, were accepted as currency to pay taxes, tolls, and rents (partly because of a coin shortage). Many European towns kept their accounts in pepper. Wealthy brides received pepper as a dowry.

Today a pound of pepper costs $17. Ground pepper sits on the table of even the cheapest diner.

Imagine a peppercorn shortage today - how long would it persist?
What took them 600 years to solve could today be solved practically overnight. In at most 2-3 years.

Kartik Gada

Very good comments, HB and Geoman.

Today a pound of pepper costs $17. Ground pepper sits on the table of even the cheapest diner.

Indeed. Many other examples exist.

i) Paper was so difficult to make that it was kept under guard and owned only by the Chinese Emperor.
ii) As recently as the 1970s, incandescent lightbulbs were expensive enough that hotels had a major problem of lightbulb theft.
iii) Ball point pens were very expensive in the 1950s and did not become inexpensive until the late 1970s.


I think what gets confusing is we abandon some things for others. We don't realize the substitute makes the price of the original thing irreverent.

No doubt I'd pay a pretty penny for some whale oil these days. Is that because there are no whales, or that it would be hard to harvest them? Just the opposite. Modern ships, GPS, satellites. I could find and harvest them with frightening efficiency. There are more whales now than in the 1950s.

Whale oil is more expensive today because whales became a luxury good. We value whales in the wild more than we value their meat and oil. We became rich enough that we bid up the value of the whales beyond the price any nation was willing to pay.

Climate change will be the same way - it won't get fixed until the climate becomes a luxury good that most nations will happily pay money for. Why won't those poor Indians/Africans/Chinese stop burning coal and switch to solar? Might as well ask why they don't all drive luxury cars and eat cake for breakfast.

Which circles back around to climate activists, who seem to overwhelming favor socialist/government solutions, that can't possibly be made to work. They won't make poor people rich, and only the rich value the environment because environmental protection is a luxury good. So only a totalitarian solution will protect the environment under a socialist system.

Want to stop the rain forest from burning? Easy - get Brazil's economy into high gear and make them rich ASAP. Those ain't millionaires setting those fires. Environmental groups would best spend their time leaning into that goal - a goal of economic development. And yet in almost every instance their dogma is exactly the opposite of that goal.

It's not like we don't know how nations get rich quickly, there's a formula for it and everything.

Kartik Gada


All good points.

One detail : the number of Blue Whales, technically, fell by 99% due to whaling. Numbers have improved since the 1950s, but are still 98% below the 19th century level.


The world came perilously close to forcing the largest and grandest creature to total extinction (population below critical mass means no more reproduction).

But smaller baleen whales were not targeted to the same extent and did not fall in number. If anything, they had more krill available to them due to the absence of blue whales, so have been in buffet-mode for a century. Total biomass of all baleen whales might just be 25% lower than the 19th century, even if blue whales are 98% lower.

Which circles back around to climate activists, who seem to overwhelming favor socialist/government solutions, that can't possibly be made to work. They won't make poor people rich, and only the rich value the environment because environmental protection is a luxury good. So only a totalitarian solution will protect the environment under a socialist system.

Excellent point. They think reducing prosperity will reduce CO2 emissions, when exactly the opposite is true.

This is why an ATOM disruption of government is on the horizon. It is the single biggest obstructor, not the least because anti-progress people are given too much power, indirectly, by being clients of government.

Joe Blow

I have a vague suspicion that bitcoin, perhaps coupled with side chains, may just be the wedge that eventually disrupts a large part of government activity. Decentralised atomic transfers of all forms of property and things like that. If it did happen there would be an incredible amount of resistance at first but once it started could be unstoppable. Probably just wishful thinking on my part.

Kartik Gada

Joe Blow,

It is true that distributed ledgers, a byproduct of cryptocurrencies, serve the function of pressuring government institutions into modernizing many aspects of transaction costs and security. It is early, but if the combined market cap of cryptocurrencies rises, that is a measurement of the pressure being inflicted on the established system.

Kartik Gada

In other news :

A huge disruption in oil supply has caused almost no trouble for the US stock market. Oil has well and truly been overpowered by the ATOM.

A similar supply disruption even 10 years ago (let alone in the 1970s) would have been enough to cause a recession.

The ATOM is making more of the world anti-fragile.


And such an attack will only accelerate current trends, more electric cars, more solar, more wind, more nuclear, more fracking, more irrelevance for the middle east.

Look how calm the markets are to destruction of 5% of the world's energy supply, versus a few tariffs with China. Interconnections of supply chains has a greater value than oil.

20 years ago, such an attack would have prompted an immediate response by the U.S., and a war with Iran. Now there is much more of a wait and see attitude. Hell, wait for the world to demand we take action. let Iran provoke and provoke till all parties agree action must be taken.

I can't see why people don't realize fracking is possibly the single greatest improvement to our economy and environment in the last 30 years. It has completely upended all our assumptions regarding energy, and world politics. It has also ended the U.S. need to police certain despotic parts of the world.

How desperate must Iran be to stage such an attack? Why would they do it? They do it because the world is accelerating past them, faster and faster. The gap between us is becoming a gulf. It is the arsonists vote - better to burn it all down than be made irrelevant. But it is already too late...

The secret of Iran is they are dying. Iran may well become the first country in the world that will get old before it gets rich. Its fertility rate fell from 7 in 1979 to perhaps 1.7 today. That produced an enormous generation of people now in their 30s to 40s who have very few children and are largely unmarried. As this generation ages, the proportion of Iranians over the age of 60 will soar from about 7% today to around 40% by mid-century. No poor country can survive this. Iran will undergo an economic disaster unprecedented in history. That is baked in the cake, and nothing its government can do will make much different.

So it lashes out because it has no internal solution to its problems, it will solve its economic problems by seizing its neighbors' oil and becoming a shi-ite empire...or die trying. It is an insanely desperate move and is almost certain to fail. But what choice do they have?

If I were them I would have tried to solve the demographic crises with immigration. Become a shia homeland. Liberalize the economy and society, ditch the bomb and terrorism, invite oppressed shia to immigrate to a robust and rapidly growing Iran. But it is much too late for that.

Kartik Gada

If America wasn't so politically correct, we could bring regimes like Iran to their knees with a simple stroke of the pen : Unlimited immigration permitted of young unmarried Iranian women under the age of 35.

Done. The country will surrender immediately rather than have that happen. This should have been done many years ago, but it would still work.


One communication issue for discussing the singularity is talking up economic growth when nominal gdp is not increasing as quickly.
Two links about that – one is Erik Brynjolfsson’s attempt to do GDP Benefits to show improvements not reflected in the GDP numbers.
I’m thinking good luck with that to track all the other improvements like LEDs, but maybe this is something that a combination of Mechanical Turk type surveys and cheap computing power can do to show the improvements where things are getting better and cheaper, but not taking more money because either a new entrant is stealing market share or existing market sharers are defensively improving what they are offering.
A second link is to demographics pushing low to negative interest rates. As I’ve said earlier, technology will eventually overwhelm the predictive impact of even demography, but it’s not there yet.
– a good call out that the increases in population in wealthier countries are the less active older folks, so not driving a lot of demand and activity and that the poorer countries where the economically active younger populations are growing are fractions of the wealth of the wealthier countries so unlikely to drive demand in the economy in the whole.
I will go back to beating the drum of Japan as an example as they are getting to the future first – if people start thinking that life in Japan is getting better despite no increase in GDP, than there will be optimism, if people think Japan is stagnant, there will not be much enthusiasm for the new world. Drew

Kartik Gada


The comment above had gone into the spam folder. I am not sure why. Two URLs usually is not enough to trigger it.

Yes, as time goes on, the current metric of GDP will continue to understate the extent to which technology improves living standards. Many examples abound (such as how YouTube videos that show you how to change your own battery and brake pads save consumers hundreds of dollars, but put mechanics out of business).

You are correct that Japan is misunderstood, as net worth is rising even if GDP is not. That said, young people in Japan are not very optimistic, partly for non-economic reasons.


I suspect the average young person in most wealthier countries is also not optimistic, the appreciating assets include real estate in areas with the best work prospects, so they feel more priced out than wealthier.


This analysis does not seem to factor in the arrival of artificial superintelligence. Do you have a separate metric for that?

Kartik Gada

Hello Dixon, and welcome.

The emergence of an economic-centered ASI sort of goes without saying, but I don't think putting out specific prediction dates about that is a gating factor in the timeframe of a Technological Singularity, nor does it have to be better than a human at ALL things.

For example, there are at least two paths to ASI. One is that the basic AGIs we have today progress up the chain of human progress (newborn, infant, toddler, 5 year old, etc). The other is that the despite the trend towards super-narrow domain-specific AIs today, that someone finds a way to link together hundreds of super-narrow AIs to form what effectively amounts to ASI. Either of these two are possible.

But also remember that a Singularity has to include diffusion to the entire economy. ASI in a laboratory is still a long way from full worldwide diffusion.


You state in your post that the technological singularity will occur "when 50% of the World economy comprises of technologies advancing at Moore's Law-type rates."

I think when you believe ASI will emerge is relevant here because if ASI is not invented before then, it would be debatable as to whether the technological singularity has actually arrived.

Kartik Gada


Perhaps. As I said before, ASI by 2060 goes without saying, i.e. it is almost a foregone conclusion.

But I don't think ASI is as soon as 2030 or anything like that, which is what some people predict.


I'm curious why you believe that ASI by 2060 is almost a certainty. It wouldn't surprise me if it takes until the end of the century for machines to be better than humans at all tasks.


I can't answer for Kartik, but I do have an opinion.

I don't understand why we need ASI at all.

1) Can we build human level AI? Yes. How do I know this? because we are simply trying to recreate something we have billions of examples of. Us. We KNOW our level of intelligence is physically possible.

2) if we are just building replicas of ourselves, why not build it to the highest standards? Meaning we'd build Einsteins.

3) Mass manufacturing means we could be hundreds, thousands, billions.

4) lack of sleep and bodily functions means there are working 24/7.

5) digital systems means instant memory and total recall, and instantaneous communications.

6) a hundred, a thousand, a million Einsteins could collaborate on any problem. Including improving themselves.

I can't find a flaw in the argument myself. once you cross threshold 1, all the rest happens in a few short years. 10 years at most. It never has to cross over to being "super" just a lot of intelligence.

An alternate scenario - we figure out how to direct link our brains to a digital system.

Projecting back, we should see human level AI within 10 years. 2030. But it might a very expensive oddity. ten years to build out in larger numbers. 2040. Anytime after that, singularity.

2060 seems a bit pessimistic and conservative to me. But, it IS the singularity, and one element of the singularity is our inability to predict the singularity. 2040-2060 seems like a likely window, barring catastrophe (nuclear war, biological was, asteroid strike, a Carrington Event, widespread socialism, etc.)

Kartik Gada


Almost everyone says that my 2060 date is too conservative. You are the first who has said it is too optimistic.

Plus, it certainly does not have to be better than humans at ALL tasks. It only has to be better at a sufficient percentage of economically valuable tasks. Tasks that are not economically valuable are not needed at all.

Again, ASI isn't truly a pre-requisite for a Singularity. Massive economic growth (up to the point where the curve is almost vertical) through ultra-inexpensive and ubiquitous AGI is also a path.

Remember that a makeshift 'ASI' is most likely to be formed by aggregating hundreds or thousands of existing narrow hyperspecialized AIs into one, through some basic linkages across these narrow AIs. It does not meet the purest definition of an 'ASI' but if it is better than humans at 500 different things, then it effectively becomes what many would imagine an ASI to be.


Imagine economic growth in the U.S.at say 5%. That is an economy growing at a rate of $1 trillion per year and accelerating. We imagine the intelligence displacing us - what will likely happen is that things where very little intelligence is required will get a huge dose of it.

Let's give an example. With say an automated long haul truck, it doesn't take a huge amount of brains to drive it. I mean it takes a lot, but not a ton of high intellectual firepower. We can almost do it now with automation.

Let's say we automate it, and the automated truck saves 20% on costs - labor, but also efficiency. It manages fuel precisely, doesn't need to take bio breaks, plots the most efficient course, is never two heavy on the throttle, gets in fewer accidents. 20%. per truck. per trip.

The trucking industry spends $700 billion per year. That is $140 billion saved. In fact your savings are more likely 50%. $350 billion.

Now you can spend this savings, but another way to look at it - the cost of shipping just got 20% cheaper. Meaning everything in the U.S. (all of which must be shipped) went down. And suddenly more opportunities open up - things that were too expensive to ship cross the line to being cheaper.

That $140 billion saved starts to reverberate through the economy. Get multiplied. Adds additional economic growth. New companies will be founded to exploit the lower shipping costs. Companies will realign their business to maximize the savings.

This will happen over and over. In industry after industry. That is what will start pushing growth to a higher and higher level.

And we don't need ASI for any of it to happen. Just near human AI will suffice.

I suspect ASI will appear and it will be a bit ho hum after that.

Kartik Gada


Fully agree. Most AI does one narrow task hundreds or thousands of times faster than a human. But that narrow task comprises only ~0.001% of the human work being done in the workforce.

For example, some computer vision programs can scan MRIs and CT Scans 10,000 times faster than a human radiologist. This lowers costs of MRIs and CT Scans.

Note how much deflation continues to be generated through the examples you provided. Hence, ATOM.

I think the equivalent of an ASI will be when someone ties hundreds or thousands of narrow AIs together. Much like a colony of ants can work in perfect synchronization across thousands of individuals.

Kartik Gada

Here is a nut who says the Singularity will be in 2035, which is patently absurd. As far as I can tell, he has no track record of predictions and is just a marketing personality.


Even Kurzweil's date of 2045 was from 2005, so 40 years ahead from the prediction date (if Kurzweil doubled down on 2045 in his upcoming book, he loses some more credibility). But this person says, in 2019, that it will be 2035, or just 16 years hence.


I was surprised by the attention to the above post because of the vagueness and lack of proof in the arguments and thought it more of a sign of his skill as a marketer than anything else. A post I have enjoyed reading is a prediction of industrial rumination, partly because I have been wondering how well microbe produced food would work since reading _Islands in the Net_ long ago – and just realized by looking in Wikipedia that that novel from 1988 is set in 2023-25.
I think that RethinkX’s prediction of this as well as their predictions of transport are over-optimistic for implementation by 2030, but could easily occur by 2040. Cheers, Drew


The demonstrations in Lebanon and Chile appear (I'm not an expert in either area) tied up in the frustration of the less wealthy that the asset appreciation is not benefiting them and is not seen to be improving their future. If GDP growth per person stalls out and people are not feeling the benefits of deflation (any gains in food and entertainment prices and opportunities may be seen as more than balanced out by the cost of housing and credentialism) we will continue to see people who are both upset for "missing the boat" in past opportunities as well as those who feel that they will face a decline as they get priced out of future opportunities. The economies may be doing well and reflected in the asset prices, but large swathes may feel excluded and act out.


Hey Drew,

Impossible meat has shown there is a market for meatless burgers. A huge one in fact. Expect such products to be essentially everywhere in a few years. And they are exactly what is described - meat produced in a lab using yeast. It could very well change the world in many strange ways. Consider this - none of this happens without a mighty dose of energy, from whatever source that happens to be. In the case of cows, sunlight is converted to grass, which is converted to meat. I know natural plant conversion efficiency to energy is really low - 1%. The the conversion from grass to cow meat is also pretty low. It another way, it takes 36,000 calories of feed to produce 1,000 calories of beef. What if the feed WAS the beef? And you could take pasture, cover it with solar cells are harvest 20% of the sunlight rather than a measly 1%. So 20 times more power to produce 36 times more meat product. So yeah - you could simply re-wild an awful lot of farm land pretty easily. And, in theory, it should cost way less money simply due to massive efficiency.

But your costs become dependent on several strange things - like access to cheap electricity. Warmer climate would help. I'd say rural Texas would be the ideal place for this to take off. New Mexico, Utah, Colorado would be good spots.

In Lebanon the protests are over a political culture of corruption being run by proxies by Iran.

In Chile the problem is that they are reaching the end of a resource export model. Resource export models of economic growth tend to magnify divides between rich and poor - those who own resource get very rich, and everyone else stays poor. Mining represented 59.5% of exports in Chile, while the manufacturing sector accounted for 34% of exports, concentrated mainly in food products, chemicals and pulp, paper and others.

Problem is resource exports grow the economy and appreciate the currency...but they don't make anyone other than the owners of the resources rich. It hurts the countries ability to export other goods and services.

For Chile to improve, more effort must be mad in value added industries, and in improving infrastructure.

A quick check of the index of economic freedom shows Chile to be in pretty good shape (> 70). However, improvements could be made to labor freedom, judicial effectiveness, and property rights. But their main problem is to transition off a resource export model.

Lebanon's score is pretty bad (51). the nominally independent judiciary is susceptible to political pressure. The courts are ineffective and struggle to enforce judgments. Corruption is reportedly pervasive in government contracts (primarily in procurement and public works) and in taxation, elections, judicial rulings, and real estate registration. Anti corruption measures are poorly enforced. Changing that would hugely increase their economic score, and grow the economy rapidly.


Just read in Scientific American: Wiring brains together to form “a biological supercomputer of networked human brains” capable of working at vastly greater speeds and across language barriers. They are completing experiments now on this. With people.

We don't need Moore's law to improve anything - just figure out how to wire together the wet ware we have, connect to a big database, and presto. Instant super intelligence.


Dear Kartik: would you please consider tweak-updating your main thesis pages to eliminate reference to a forthcoming crash in 2017? I think you are making very important points, and I want to refer more people to your work, but it is inconvenient to have that glaring non-starter point right in their faces. Granted that your overarching message remains the same; still, it is distracting. It would not be hard to make these changes, really just tweaks. Thanks!
Alan @alan2102z

Kartik Gada

Hello Alan,

Yes, that update is being done, and the entire book will incorporate new findings as well.

Hopefully around the Xmas timeframe, if you can wait to send it to people until then.


I hope you will make an update to your article "The Misandry bubble also". It will soon be ten years since its publication, and I am very curious about your updated judgment.

You predicted a burst of this bubble by the end of the decade but it seems to me that this bubble is ever inflating, even the rate of growth seem to increase.


Dear Kartik: thanks for your reply, and the upcoming update. I look forward to it. I will be plastering it on twitter and elsewhere from time to time. Great stuff! Methinks austerianism is about to go into a permanent death-spiral. :-)
Thanks for all your contributions.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Your Information

(Name and email address are required. Email address will not be displayed with the comment.)