Quantifying and Identifying Singularity Level Technology

by

The concept of Singularity is popular for futurists and for science fiction writers. However, it is very vague. It is an intelligence or technology “explosion”. What does that mean? We don’t know. How do you measure it? I have a proposal.

I think it is more useful to define and quantify a Singularity as a Technological and economic next leveling of civilization. I would measure this via increased economic doublings over the baseline. Rail and Industrialization did next level the economic growth of civilization. We went from two doubling over 2000 years to a doubling every 25 years.

I disagree with the AI language translation company touting in Popular Mechanics that a more perfected translation of language as heralding the start of a Singularity.

The global baseline since rail and industrialization is doubling once every 25 years. We would expect 3 doublings or 8X the world economy from now to 2100.

There have been almost four doublings since WW2. The world economy was just short of $8 trillion (in today’s dollars) in 1940 and now $100-120 trillion.

$1.2 trillion in 1820 and $1.79 in 1870. $182B in 0AD and $200B in 1000 AD. 1700-1880 was the transition.

Tech prior to 1820 was increased population levels with stagnant per capita output and production. Agriculture improvements were almost all of what really mattered. Sanitation and clean water were the rest.

The World in WW2 was using about 16 times (4 doublings) less steel and oil than today. Germany was trying to get at a few days of our current oil production in the Caucasus.

The post-war baby boom added half of an extra doubling. The technology of society enabling sustainable population growth was the history of economic growth up to 1820. The first 30 doublings were about technology and society enabling population growth.

Computers and Internet Were Not Enough

Personal computing, the internet and smartphones did not add extra growth to the level of an extra doubling but just sustained global growth levels despite slower population growth. This proves that up to this point computing, internet and smartphones were not near Singularity level technology.

Technology has started to give economic doublings without the population doubling. At 8 billion people now and then pretty flat at 10 billion from 2050-2100. There were 2 billion people in 1927 and 4 billion in 1974. A doubling and half of the population since WW2 and economic per capita increased by about 5-6 times.

If all of the new technology and a 1.25X in population just let the economy double by 2050, then we have sustained the post-WW2 growth rate. If per capita income tripled by 2050 then we will have one extra doubling. 1.25 population and 3X economy would be 4X the total economy by 2050. Reminder, doubling the world economy in 25 years is normal for the past 100-150 years. This would be one extra doubling above normal.

1.25 population and 6X economy would be 8X which would be two extra doublings above normal (aka level 2 Singularity).

1.25 population and 12X economy would be 16X which is three doublings above normal baseline. I define this as a level 3 Singularity.

The extra doublings can be spread out of 50 years or 100 years or longer because we have a 150-year baseline for the expected pace of doubling. Sustaining the baseline growth is non-trivial, especially with population growth going from 100% of the economic growth effect to 50% and now to 25% and soon to 0%.

What Technologies Could Be Powerful Enough to Give Extra Doubling(s)?

1. Perfected and deployed self-driving electric cars and trucks. This could reduce the supply chain transportation costs from 10% of goods value to 2%. Electrification can reduce fuel costs by 80% which means per mile truck costs go from about $1.80 per mile to $1.50. But Platooning of vehicles can reduce that further especially if there are no drivers in the following trucks. Following robotic trucks would take per-mile costs down to about $0.6 per mile. Fully robotic vehicles could safely drive at 120-150 mph. This could be an extra doubling over 25-50 years.

2. Teslabot and fully automated factories.

Teslabot has 3% of the mass of a Tesla Model Y. If you can make 20 million EVs of Model Y class per year then the same production could build 6 billion Teslabots per year. Each year adding the equivalent of the human labor force if the Teslabot became as productive as a human.

3. What comes after ChatGPT and Alpha Fold 2 ? If advanced ChatGPT displaces or merges with Google Search. Alpha Fold X becomes the digital biology of humanity, microbiome, and the ecosystem.

4. Aging reversal. Making the 80 the new 35. Europe and Japan’s 30% seniors get back into the workforce and give a one-time doubling of the workforce. Long-term effects unknown based upon fertility impacts.

5. Fleet of fully reusable Starship Super Heavy deploying Teslabots, mining bots and automated replicating factories to Mars, Moon and asteroid belt and then Kuiper and Oort Cloud. This would let civilization rapidly access millions times the physical material (steel etc…) and trillions of times the energy. With rapid exponential expansion via self-replicating factories, the solar system could be accessed in hundreds of years to provide 20-40 extra doublings.

I had some videos that explain these ideas.

Being able to measure a Singularity and being able to quantify Singularity lets us think more clearly about what technology or changes would actually count as a technology that could up-level civilization.

The doubling scale and analysis works from the beginning of civilization to the far future.

42 thoughts on “Quantifying and Identifying Singularity Level Technology”

  1. In the current inflationary environment it is difficult to know if the increments are real (inflation-adjusted) or nominal.
    Perhaps you could restate your analysis in terms of a stable benchmark, such as troy ounces of gold.
    The other consideration is population growth, linearly dependent on agricultural output. With 8 billion of us I would hope for 4 times the production and innovation as when we were only 2 billion ( I remember 3 billion as the answer in school.)

  2. I am immediately filled with suspicion when I see a graph giving the world GDP in the year 0 to 5 significant figures.

  3. I love all the ideas. Utopia! Elon Musk! Yay!

    Lots of great possibilities.

    How do we keep most of humanity striving to achieve even more? What happens when everyone is fat and lazy, and some evil madman uses these technologies against us?

    I guess I need to quit reading the news about China, Russia/Ukraine, the Congo, Eastern Nigeria, Iran, and Pakistan, and just dream a little.

    It’s all gonna work itself out, right?

    Right?

  4. Well, there’s The Singularity and there are technological singularities. Technological singularities are those points where a disruptive technology becomes well enough established to start having an effect on everything.

    They are called singularities (small s) because people who have not passed through them are generally unable to “see” what things may be like on the other side. Like how the nomads that discovered throwing down some grass seed so they could get even more grass seed (to ferment and drink) might never have imagined the rise of city-states.

    These generally seem to be coming faster and faster throughout history, going back to things like stone tools, fire, agriculture, etc. all the way up to things like the printing press, the industrial revolution, electronics, the internet, and whatever seems to be happening right now. In other words, each one is coming in about half the time the one before it took to arrive after the one before it.

    Yes, the internet qualifies as a technological singularity (although not as The Singularity). Consider just one facet of it being that scientists used to have to labor years to get published, and it might be years more before someone came across that work and realized it could be helpful in their work. This can happen at the speed of heat nowadays. That has a real impact.

    The current singularity, due this decade and which we are probably already in might someday be termed as “cognitive automation,” arriving around 30 years after the internet. At least, I’ve been predicting it for at least ten years and I think that may be what is happening.

    That would indicate further technological singularities might be forth coming next decade, and the one after it, with multiples in the 2050s, if that is really possible. Candidates would seem to include strong AI (either AGI or narrow or both), radical life extension, true man/machine mental interfaces, and self-replicating manufacturing facilities. Or something else, possibly something we haven’t even considered yet. And how will we react to all of this? We can’t foresee that yet. It’s why they are called technological singularities (although “technological event horizons” might have been more descriptive).

    • Well, there is regenerative AI:

      “The breakthroughs are all underpinned by a new class of AI models that are more flexible and powerful than anything that has come before. Because they were first used for language tasks like answering questions and writing essays, they’re often known as large language models (LLMs). OpenAI’s GPT3, Google’s BERT, and so on are all LLMs.

      But these models are extremely flexible and adaptable. The same mathematical structures have been so useful in computer vision, biology, and more that some researchers have taken to calling them “foundation models” to better articulate their role in modern AI.”

      See:

      https://arstechnica.com/gadgets/2023/01/the-generative-ai-revolution-has-begun-how-did-we-get-here/

  5. If you’re going to count Elon Musk and Tesla as “central to taking us to the next level of artificial intelligence, or anything else, then you’d better factor in Key Man risk in your calculations. Musk is 51 years old, works a punishing schedule that would overwhelm a man half his age (let alone women, who still juggle most of child-rearing), and has lawsuits, parental and ex-mate demands too.
    How many productive years does he have left, assuming no great breakthroughs in Longevity Science? 10? 15? I’m 64. I can tell you it’s almost impossible for a 64-year old man to work at the pace and depth of a 51-year old. We prioritize and grow wiser with our time, but that is not enough if Musk wants to continue to be the Key Man in half a dozen businesses.
    I notice that Twitter is nowhere in your estimate of Singularity-causing businesses, and Musk is spending an awful lot of time and resources at Twitter. It may even retard the Singularity as a time sink. Computers and the internet can partly be timesinks too, raising the bar of expectations, without really producing anything productive, just make-work. A.I. looks like it’ll do more of that too.

    Think of Apple after Steve Jobs; nothing evolutionary has been made, just incremental improvements on a product that Jobs invented over a decade ago, and a nearly forgotten line of desktop computers and related items that cannibalize each other. That may be Tesla too, post Musk.

      • And the Motorola 6800 series was superior in every way to the Intel 8080 series, but look who won that fight. Technical superiority vs resources thrown at, the latter often wins.

  6. In my opinion, intelligence enhancement techs are keys to rapid economic development. In 5 years, we will have chatgpt for intelligence amplification, in 30 years we likely have brain-machine interface techs like neuralink, in 50 years we will bold enough to use designed baby or genetic techs.

    • Dunno. All these improvements, like Teslabots building Teslabots, seem to be based on removing humans from the equation entirely. We’re going to become like cows while The Machine gets on with it, with a plan thats impossible for us to comprehend even if plugged directly into our brains.

      The plot of 2001 is something similar.

      • Meh, getting stuff done isn’t about just brawns and brains, it’s brawn, brains, and motivation/direction.

        People pre-industrial age had to provide the brains, the brawn, and the motivation/direction.

        Afterwards they had the machines to provide the brawn, while they still had to provide the brains, and the motivation/direction.

        After cognitive automation, the machines will provide most of the brawn and most of the brains, but humans will still be choosing what they want the machines motivated/directed to do. I don’t see us giving that up any time soon.

        • I’m not so sanguine about human motivation as you here, unfortunately.

          10 years ago how many of us would’ve been OK with a big corporation knowing virtually everything about us? Now, almost all of us tolerate it. Run that process forward for a while with new, flashy tech and it doesn’t look good for us.

          To me, it seems there are very powerful forces working to make humans into “zoo animals”, and I’m not sure how successful that process will be.

  7. I like discussing these abstractions –and I think that they have real-world applicability– though Singularity and socio-economic accelerations ‘doublings’ and such seem to dwell on very specific societies, industries, and lifestyles — and their historical precedents.
    Fine. But I think that focusing on a more Universal goal, which may not be as specific as hoped, and may still be a moving target…
    Post-Scarcity.
    Often discussed, occasionally modelled, very much theorized, featured in some glorious scifi: Star Trek, Bank’ Culture series, etc.
    How could we make it that each and every person had unlimited ability to have all the knowledge, commercial goods/property, travel opportunity, life span, etc., that they could conceivably fit into an individual Life? (and let’s not for the moment consider those ridiculous celebrity/ oligarch individuals who have like 20 houses they visit for one week a a year). A full, rich, productive, ambitious life creating and consuming without restriction.
    Post-Scarcity. Consider population and its change. Consider basic resources and their limits (but also infinite recyclability/closed-loop). Consider off-Earth travel. Consider consumable entretainment, knowledge, research access…
    Post-Scarcity Level One? When? Something within this century. 130 years life span. Full access to every piece of entertainment, knowledge, research assistance so as to as many personal postPhD-equivalents as one wants. Resources that allows one week a year in orbit/ one week in 5 years on the Moon. Every possible foodstuff, commercial object, and consumable experience/ service. No need for money – just make an appointment for all.
    My 2c.

    • Could be diluted or de-railed by people trying to figure out how they can stop working rather than being more productive, more ambitious, more technological, etc. A decade long Sabbatical.

      Though having unlimited job choice, as long as qualified, and not limited positions available – would be a feature of post-scarcity. How to balance a person’s contribution vs their consumption – a minimum level of work equals one post-scarcity ‘share’?

      • The problem with that is that much of the wealth creation, technology development, skill enabling comes from the type of people who want scarcity. Coveting intellectual property. Coveting product breakthroughs and limited early release. Limiting access to maximize impact and profit. Incentives matter. Winning matters. Trickle-down, early adopter, and exclusivity-first is the only way forward. Most real advances came from people who were most interested in getting rich.

        • Yeah, you’re right. Nothing much gets accomplished unless there is an incentive. Offering unlimited vacation and artistic/intellectual pursuit means limiting humanity’s potential to keep achieving…

  8. Antiaging is coming, but at first, it will be for the rich:

    “According to a recent Bloomberg profile of the CEO, Johnson could spend up to $2 million on his body this year and there are early glimpses that show he may be on track to unlocking the secret to age reversal.”

    See:

    https://www.msn.com/en-us/money/other/a-45-year-old-biotech-ceo-may-have-reduced-his-biological-age-by-at-least-5-years-through-a-rigorous-medical-program-that-can-cost-up-to-2-million-a-year-bloomberg-reported/ar-AA16KoFu

    • The world has long been about “have nows” and “have laters” not “haves” and “have nots.” Misguided efforts that focus on “haves” and “have nots” just lead to everyone being “have laters.”

    • Not necessarily.
      It costs 100s of billions to keep old people relatively healthy and alive. Already, Republicans and even some Democrats like Joe Manchin are discussing how to cut Social Security and Medicare (SS should not be cut and is actually economically stimulative, putting money INTO the economy. Medicare/Medicaid are less justifiable, from an economic POV at least).
      If there is a treatment that would essentially eliminate 50-90% of disease by de-aging people, there would be a lot of pressure to convince people to take it, even if it cost a million per person. That could end up being cheaper than 5-10 years of end-of-life care, including dementia. It’s a one-time big expense – unless the treatment is for “Chronic Aging Syndrome (CAS)” – but it will pay back big-time in a generation. And that’s before even counting on the extra years of productivity from 60-80 year olds, and beyond in a second generation.

      • Considering that the almost all the Baby Boomers are in retirement or about to retire it puzzles me why the Republicans are pushing cutting Social Security. That used to be known as the Third Rail of politics and now there are more retired people than ever before AND THEY VOTE. Sounds like the planned demolition of a political party. *shrug*

    • I didn’t see your link but read on reddit that this guy isn’t really testing any ground breaking anti aging therapies, but using the usual stuff: fasting, exercise and regular testing.

      • There aren’t really a lot of ground breaking anti-aging therapies available, unless you can worm your way into the right clinical trial. But everybody can do fasting and exercise, and most people not poor can do the regular testing.

  9. “Singularity” means infinite growth speed. So the proposal is that 3x in 27 years is “infinite speed, level 1”. And 12x in 27 years is “infinite speed, level 2”.

    Ok. I guess 12 equals infinity, for small values of infinity.

    • I am correcting. 2X in 25 years is normal for the past 100-150 years. 1.25 population and 3X economy would be 4X. One doubling above normal. 1.25 population and 6X economy would be 8X two doublings above normal. 1.25 population and 12X economy would be 16X which is three doublings above normal baseline.

      Infinite speed growth is meaningless.

      • Yes, a true singularity is impossible. The original people who used the term were envisioning something like 1000x in a single year, or even faster, which they considered “practically infinite”, so they called it a singularity.

        It usually was used for a situation where you had both an advanced nanotechnology (which can increase far more than 1000x in a year) and a self-improving AI (which they thought would be similarly fast).

        I’ve always been skeptical that such a singularity can occur.

        • The definition I’ve heard is that the “singularity” is when the speed of technological development actually outstrips the community’s capability of understanding and absorbing the technological changes, so by the time you hear about the great new thing and get a grip on what it’s all about and what we can do with it, it’s already obsolete and there is something else even more advanced which outstrips it.

          Note that, under this definition, the singularity is a problem, not a solution.

          • Ok, I can see why people might call that a singularity. It isn’t literally infinite speed, but to the people struggling to grasp it, it might seem to be infinite. So you could say it seems like a singularity. That makes sense.

            In math and physics, a singularity is where a curve goes vertical (infinite slope) or the center of a black hole (infinite spacetime curvature). I can see how the situation you describe would seem like infinitely fast changes.

            • If we grow at 11% a year, then that’s more than 16x in 27 years. I think 11% is impressively fast. But I wouldn’t call that “infinite rate” or “singularity”. Not even metaphorically.

              But if the growth is faster than a person can comprehend, then sure, that might be called practically infinite, or practically a singularity.

      • It’s a reasonable way to quantify the detection of an approaching singularity using just economic data. GPT could probably help. Evidence of accelerating economic growth in the most advanced economies – which traditionally are mature and slow growing- is the metric. It ought to show up as a counter-trend in places that now have a demographic collapse going on like Japan. The defining characteristic of a shift toward a singularity is accelerating anomalous growth counter to traditional demographic expectations. If an aging shrinking population is producing accelerating economic growth there is no traditional model explanation.

    • Wow, I liked that. I would rather call those not mini singularities but rather pre or harbingers of the singularity. And technological changes leading to a comparative revolution over a couple of decades are nothing new. Probably as early as the adoption of the printing press and gunpowder

  10. The premise of your article is spot on. my only caveat is that much of what you have written will not become the tech we want or need until we develop a true General AI, that can understand the concept of your orders not just the words you speak that match its programing package. The trucks, bots, and even “space mining bots” all depend on a very strict programing language that makes them very limited in what they can do or handle if things go wrong. ChatGTP may seem great, until you realize that if you put in the same parameters as i do.. we are both likely to get the same response, (or even gibberish that only half makes sense) even if we are sitting miles away knowing nothing about each other. this is not intelligence its biased programing and i dare say is a negative more than a possessive as it restricts critical thinking and problem solving, lacking the intuitive leaps that a human mind can make.

    Last but not least its hard to get excited about the “Future” of these technologies since most of them will be locked behind corporate doors and unavailable to the common worker for the foreseeable future.. space mining, tesla self driving trucks, de-aging. factories that run themselves all sound great until you realize that you wont get your hands on them anytime soon… its like getting excited that your neighbors grand kids are gonna win the lotto 20 years after you’re dead..doesn’t do you a lot of good. O.o

    What are you going to do with all those displaced workers? how do they make a basic living if common labor is done by robots?

  11. A fundamental complaint about your analysis: Leveling up of backwards areas is not remotely the same as growth in advanced areas.

    By the way, I use per capita numbers, because we’re looking for improvement in tech and practices, not mere population growth at a fixed tech level. We want to know if people will be leading better lives, not if more people will be leading unchanged lives.

    Looking at the numbers for the US, in 1945, US per capita, inflation adjusted GDP was $14,213. In 2022, 60,080.

    That’s a bit over two doublings, not four. And the US has had an unusually high growth rate for a developed country. Most of your eight doublings for the world economy has been bringing the bottom up, not the top.

    Not improved practices, but just wider adoption of proven practices.

    The reason this is significant is because, once the best practices are used everywhere, you’re limited to the improvement of best practices, and that bonus growth goes away.

    Now, what do we need to improve the US doubling rate of once every 41 years?

    Basically, we need to further decouple production from human labor. Automation, in all its many forms. These doublings are a product of technological advances reducing the amount of human labor necessary for production.

    And there’s no fundamental reason we can’t totally decouple production from human labor, can’t invent self-replicating factories. Might be a tough job, but that’s what we need to aim for, to achieve a real singularity.

    • Reasonable on GDP … BUT .. how for instance does this blog and these discussions get measured in GDP or growth, how does the free videos / advice / teaching etc etc get measured? ……. by Google profits by power consumption? … a motivated human using GPTchat and google could gain real insight into a thousand subjects how does the value get measured?

Comments are closed.