Construction Physics

Maybe you meant 1900s to 2000s but if you meant the year 1900 to the year 2000 then that century of difference saw a lot more innovation than just the "candle to lightbulb" change of 1800 to 1900.

I'll interpret it as meaning 1800s to 1900s to 2000s. I'd argue that we haven't yet seen the same step change as 1800s to 1900s this century because we're only just beginning the ramp up on the new technology that will drive progress this century similar to how in 1926 they were still ramping up on the use of electricity and internal combustion engines.

Let's take electricity as the primary example though since it's the one you mentioned and it's probably more similar to our current situation with AI. The similarities include the need for central generating stations to supply raw power to end users as well as the need for products designed to make use of that power and provide some utility to the consumer. Efficiency of generation is also a primary concern for both as it's a major cost driver. Both of those required significant investment and effort to solve in the early days of electrification.

We're now solving similar problems with AI, instead of power plants we're building datacenters, instead of lightbulbs and washing machines we're developing chat bot integrations and agents, instead of improving dynamos we're improving GPUs and TPUs. I fully expect we'll follow a similar curve for deployment as we find new uses, improve existing ones and integrate this new power source…

I once spoke to a hedge fund guy who had a large stake in Tesla going back at least 10 years at this point. I was curious... don't you get worried about every other company announcing a push into electric cars? GM and Ford are making big pushes, as are many of the European companies. His response: the best engineers want to build sexy new things and they don't want to live in Detroit.

Construction Physics recently ran a great article on why companies don't design new aircraft from scratch (rather, you see 5 or 6 iterations of the 737) [1]. The gist of it is that it's extraordinarily expensive to design a new aircraft from the ground up and that changes to regulation or the global economy can have a huge impact on which planes are desirable (e.g., big vs small, fuel efficient vs fast), so any new plane is an enormous gamble. So I understand why the bean counters look at a build-from-scratch approach as a bad idea. At the same time, it is very hard to quantify the long term impact of losing the engineering culture that comes along with building new planes from scratch. There's a wealth of knowledge that you lose if you don't do it very often. You also get happier and better engineers when you give them tasks that are more interesting. This stuff doesn't show up on accounting statements, but it's hard to deny that there's a long term effect.

Unfortunately, it doesn't really seem like a return to engineering culture is what they are going to do. The head of their commercial…

The author previously did a great blog post breaking down the overall costs of building new homes:

https://www.construction-physics.com/p/what-makes-housing-so-expensive

tl;dr: For new developments, land is only about 20% of the total sale price of the home, and this share has been dropping since 2005.

However, he also notes an important selection bias to this stat: new homes do not get constructed in areas where land is expensive, unless it's a custom build for a specific wealthy buyer, because developers cannot make a profit building SFHs if they spend a majority of the purchase price just buying the land. He notes that in wealthy areas like eg. Silicon Valley, 80% of the purchase price is land value, but developers are not building homes there, which is part of why the purchase price is so high.

And sure enough, if you go out to Mountain House or Discovery Bay, prices are like 25% of what they are in Silicon Valley. But then you have to commute from Mountain House. America's housing problem is as much a commuting or job distribution problem as housing, but those problems are even harder to fix than housing.

The times article hand-waves away distracted driving by saying that other countries haven't seen a similarly large increase. The problem with that is that vehicle sizes in all other countries have also been increasing, and other countries like Canada and Australia have seen almost identical market shares of those large vehicles without the surge in pedestrian fatalities.

The large factors are phone use (more prevalent among american drivers and there's data to show this), and homelessness - the homeless are dramatically overrepresented in US pedestrian deaths and the population has increased dramatically in the US over the past decade. Though no single factor is a smoking gun, vehicle size is one of the least convincing. Getting hit by a 4000lb car or an 8000lb truck matters much less than how fast the vehicle's going (let's all remember our high school physics class).

This blog post is the best deep dive I've seen on the data: https://www.construction-physics.com/p/more-on-us-pedestrian-deaths

Building a new factory would cost $20 billion and take 3 to 4 years [0,1]. With chip output capacity and AI boom profit margins, it would take just under a decade to break even. If the bubble bursts and chips return to pre-boom levels, then it would take over 30 years to break even.

Ford had almost $20 billion in EV car manufacturing investments planned for the mid-late 2020s and the abrupt end of the EV subsidies cost Ford billions of dollars and they have abandoned multiple investments.

If you do nothing, you still are rewarded because you are making pure profit in either scenario. If you invest billions then you are digging out of that for years regardless, and could be in the whole for decades if you bet wrong.

[0] https://www.construction-physics.com/p/how-to-build-a-20-billion-semiconductor

[1] https://techovedas.com/what-does-it-take-to-build-a-semiconductor-cmos-fab/

Don't get me wrong, it is great that we can harness wind and solar to generate tons of electricity. But as the article points out, they aren't consistent enough to provide electricity that's always on, which is something our society is built to expect. Last year, Contruction-Physics.com put out a great 3-part article on why nuclear is so expensive. It's worth a read [1]. But long story short, we hardly build them anymore (we last finished a new plant in 2016 and there are two more currently under construction, which takes many years) and regulations change all the time so there are very few people with real expertise at putting them together. And talented people certainly don't want to go into the field, because there's no guarantee of consistent work.

[1] https://www.construction-physics.com/p/why-are-nuclear-power-construction?utm_source=%2Fsearch%2Fnuclear&utm_medium=reader2