Comment by austinl | Hacker News Reader

Comment by austinl | Hacker News Reader

While not directly tied to economic growth, one factor this article leaves out is that global human population growth could possibly peak within our lifetime (although more likely around 2100) [1].

In the last 60 years, total fertility rate has dropped from 5 to 2.5 [2], and most industrialized nations are hovering right around replacement rate of 2.1 or actively shrinking (Japan, 1.4, Germany, 1.6). Albeit during COVID, the 2020 TFR in the United States was only 1.64, and has declined for the last four years in a row [3].

With technology, I'd still expect the overall "size of the economy" to grow, but it will be interesting to see how growth is affected by substantial changes in demography that play out over the next 100-200 years (if only I could stick around to watch!).

[1] https://en.wikipedia.org/wiki/Projections_of_population_grow... [2] https://en.wikipedia.org/wiki/Total_fertility_rate [3] https://www.nytimes.com/2021/05/05/us/us-birthrate-falls-cov...

Demographic leveling or collapse is in a category I've heard called "demand limits to growth."

Everyone always thinks about supply limits, with some making dire Malthusian prophecies of what happens if we exhaust our resources with runaway growth. I've seen surprisingly little discussion of demand limits.

Other possible sources of demand limits include: satiation (people just feeling satisfied and not wanting much more), diminishing marginal utility of wealth, diminishing impact of new technologies (compare the next CPU density node to jet aircraft or antibiotics), consolidation of functionality into fewer more versatile products (e.g. computers and smart phones eating TVs, tape decks, radios, book shelves, CD collections, etc.), cultural shifts toward simplicity, replacement of complex labor-intensive technology with simple low-maintenance technology (e.g. EVs have 1/10th the moving parts of typical ICE cars), etc.

The eternal growth assumption assumes both infinite supply and infinite demand.

As a civilization, we are good at inducing demand, though. For example:

> satiation (people just feeling satisfied and not wanting much more)

Advertising industry literally exists to mitigate this problem. I don't envision it hitting a limit any time soon.

> cultural shifts toward simplicity

This has not happened yet. Our existing "trends" for simplicity and minimalism are just thinly-veiled attempts to get people to buy more shit they don't need (like "simpler" items, to replace the perfectly fine items they already have, or books and videos and conference tickets of minimalism thought leaders).

> replacement of complex labor-intensive technology with simple low-maintenance technology (e.g. EVs have 1/10th the moving parts of typical ICE cars)

I don't believe this is stable state. EVs are more reliable because they're new. They haven't gone through enough value engineering cycles. The market has a structural incentive to optimize away all quality, until the result is barely fit for advertised function - so I expect that once EVs start dominating, they'll also start breaking at the rate people are used to with ICE cars, because manufacturers will optimize away spare capacity and surplus material of every component.

I think the risk is still dominated by the supply side.

People demonstrate more and more indifference and sometimes outright hostility to ads that offer what they think they don't need.

The genuinely useful part of the ad industry is connecting those who need stuff to suppliers of such stuff, that is, discovery.

ICE cars are inherently more complex than EV cars, much like a Swiss army knife is inherently more complex than a kitchen knife.

If EVs quality will degrade to the quality of current ICE cars, that would mean that production if EV cars will use less resources, not more. It will not have to repeat the ICEs' complexity.

But frankly popular EV cars like Tesla are often considered sub-par compared to modern ICE cars with regard of build quality. They are used despite these shortcomings, and will actually need to rise tp the quality level of e.g. ICE Toyota cars.

Nobody is going to open up and fix EV's in the near future. When they break down people will be expected to just go buy a new one. Especially if they are much cheaper to produce and buy. If you can buy a car for 3-5k nobody is going to fix them.

You will own nothing and be happy.

By all means, show me a new car with a list price of 3-5k. Current ICE cars already cost 10x that amount, and EV cars are generally even more expensive. On what basis do you expect new cars to ever cost 3-5k? That's the range of an electric bicycle, not a car.

https://en.wikipedia.org/wiki/Wuling_Hongguang_Mini_EV

But sure, I was exaggerating. I should have said 10-15k where you will find more examples that are less deadly to drive :)

But actually, I agree with the other comment. I would be surprised if the automotive industry doesn't manage to force us all into car subscriptions.

For the first little while you'll have to buy the car and the subscription, but over time the upfront cost will be less and the subscription will get bigger.

When a my 1980s Ford, my one friend who pays his electrical bill + has a craigslist-sourced machine shop can fix it from steel+aluminum stock, and also teach me how to fix it myself.

When something with power electronics breaks, my option is to buy new power electronics from a big company or trash it. My many friends with PhD/industry EE experience are unable to fix it nor tell me how to fix it with simple tools.

The complexity of the latter therefore seems greater in he colloquial sense than the complexity of the former.

A 1901 electric car has batteries in glass jars and plates of iron and nickel, speed control of just carbon and copper, easy to make with a vice, hacksaw, and file.

Let's not be comparing obsolete technology to modern.

With your 1980's ford, you outsource all off the complexity to a local shop. You're enjoying 100+ years of investment in internal combustion engines. In the future, with wide-scale adoption of EV's, local shops will respond to market demand. We're in early phases. The Model T was launched in 1908. In 2121, I think it's a strong possibility you'll be able to get your 2080's Ford serviced at a local shop.

"Advertising solves satiation"

You're not thinking big enough. A star trek style holodeck (or matrix-style brain port) has the possibility to simulate literally any experience with roughly constant resource inputs.

Short of literal deception, advertising cannot offer someone with access to those technologies anything.

You don't think holodecks will come with freemiun subscription plans or micro-transactions, a la video game skins and upgrades today?

Why would someone sell you a holodeck once, when they could sell you plugins and content modules for your holodeck continuously forever? You don't think you'd get bored looking at the same trial plan holo-beach for years?

Of course, piracy and/or open source will hopefully be around to offer free alternatives.

At first sure. But in this hypothetical future I think holodecks would eventually become widespread enough that an upstart vendor would differentiate itself by selling an unencumbered holodeck.

At that point, the designs themselves become the currency, and spread of them is limited by brainpower of the authors.

People cannot be easily prevented from using (or trading) a thing they make themselves. Enforcement of it requires serious violence. Previously, that resulted in communist revolutions.

No matter how clever/appealing the Smirnoff advertisement becomes, there is a point where drinking more vodka just kills the customer. There are analogous limits for any product that one can eat, drink, inject, or inhale. For products that aren't directly consumed by the body, the limit is time if nothing else. An hour spent watching Marvel® characters save the world is an hour you can't spend driving your Acura MDX with A-Spec®. Even the best ad agencies can't create enough hours in a day to indulge every urge they cultivate.

On top of that, you cannot watch infinite number of advertisements at the same time.

Attention and time is a limited resource. Even shortening work day we get to the point where work is dominated by travel time, further reductions there require restructuring cities or removing whole swathes of work while still paying people to consume.

That's where population growth comes in. You can't add hours to the day, but you can add person-hours.

> I don't believe this is stable state. EVs are more reliable because they're new. They haven't gone through enough value engineering cycles. The market has a structural incentive to optimize away all quality, until the result is barely fit for advertised function - so I expect that once EVs start dominating, they'll also start breaking at the rate people are used to with ICE cars, because manufacturers will optimize away spare capacity and surplus material of every component.

This is a completely nonsensical argument that doesn't fit the real world data. ICE cars have steadily improved in reliability over time. A car that made it to 100k miles was notable in the 80s. A car model from 2010 that didn't routinely make it to 100k miles would probably be considered defective.

Whether it makes to 100k miles is not the metric.

The metric is, how long before you have to get it to a mechanic to fix a component failure, and how difficult/expensive it is. Engine durability is good in ICE cars these days - it's everything else that falls apart much quicker.

This is completely anecdotal and could be explainable by survivorship bias, but in my experience, past 1990 or so, the newer the car is, the more likely it is to cause regular and expensive maintenance burden on the owner.

Also, think of it this way: people are already used to having to do some serious maintenance work on their cars around 100k miles - why would EV manufacturers not want to eventually target similar repair schedule? In complex projects like these, reliability is a controlled variable.

> As a civilization, we are good at inducing demand, though. ...

This right after GP mentions the unreasonability of the infinite demand assumption. Is it? I think so. How is that wrong?

I should've expressed it more clearly in the original comment. My point wasn't that demand limits don't exist - they do, and GGP is right by calling attention to them, they're definitely not talked about enough. I only wanted to say why I think we'll hit the supply limits first.

I think it's the inverse. Demand limits first.

Wealth inequality also exhibits a form of demand limit.

Not that there's no demand per se but if there's no spare cash to back it up...

At some point people can't come up with enough ideas to spend it all even if you forced them.

Pretty much every aging country is seeing the demand side of their economy die. I'd say education increases the minimum standard of living that people expect. If having more children threatens that standard of living they will refuse to have them.

I agree. The book "Factfulness" by Hans Rosling is an excellent read about the topic.

Yes, growth in economy and population can go on and will likely result in an S-bend curve in the distant future.

Hans Rosling's presentation "Don't Panic" (produced by the BBC) is pretty good:

* https://www.youtube.com/watch?v=FACK2knC08E

It's an hour long; really interesting.

Reply on news.ycombinator.com