He gets how important the Liverpool and Manchester Railway (1830) was. I've pointed to that for years as the moment when the Industrial Revolution got out of beta. There were railroads and steam locomotives before that, but they were one-off demos or in-house operations. The Liverpool and Manchester was all steam locomotives (no horse-drawn vehicles), with double tracks, signals, switches, schedules, and tickets. At last, people could buy a ticket and go someplace. As the author points out, once it got running, it was quite profitable.
So, of course, everybody wanted to copy it. Hence the railway mania. I'm amazed that the author had trouble finding materials on that subject, but apparently he searched for the phrase "railway mania" and didn't get enough hits. Strange. Oh, and on page 189, there's "(in fact, got continued growth)", a phrase from an outline never filled in.
Brunel was indeed a high-cost builder. Most of his works are still standing and in regular use, even those coming up on two centuries old. Railway mania was enormously profitable for Britain, although many railroad stockholders did not do as well. Railroads probably paid off better than aviation - commercial aviation, over its entire life, has been a net loss to investors.
If you want to read about 19th century railways, both Stanford University and the San Francisco Mechanics Institute have good collections.
You could say the same about the investors who funded the city-to-city fiber network in the US in the 1990s.
In both cases, the investors left us with a legacy of useful infrastructure.
I'm trying to figure out what legacy the cryptocurrency boom will leave us. Buildings full of rapidly obsolescing ASICs and GPUs in remote corners of the world?
Optimistically? One of the newly minted millionaires will go on to do something great, like the people who built the fiber net.
Realistically? Who knows...
And a big part of the system isn't about 2-3 hour train rides between LA and SF. More half hour commuter train service in the LA basin and between Morgan Hill, San Jose, SF. The fast rail between LA and SF is really about not having to expand a dozen airports which are near or over capacity. 10 X 2 billion ea is ~$20 billion.
Yeah so big project with a lot of facets with economic effects on people and other infrastructure that doesn't exist yet. Make it hard to do a realist balance sheet analysis. Even it is a bit of white elephant, better than the F35 or the Nuclear Weapons moderation program at $1T and $700M respectively.
 There is a water tunnel running through the same section, so frankly doable. Just $$$.
* The useful part of the project is doing the rail connections across the mountains, which is unfortunately the expensive part. For various reasons, the project started with building the section in the Central Valley, which is the most useless part of the project by itself. If the project is killed without connecting to either SF or LA, then it really will be a useless white elephant.
* The best routing is far from obvious. The only easy way out of LA is roughly I-10 to Palm Springs, but that's the wrong direction for anything other than Phoenix and maybe Las Vegas. This means the decision is going to be contentious, and every person drawing the transit fantasy maps is going to be upset at it, thus spending more time bickering and less time supporting the project.
* The cities aren't very well set up for mass transit to take advantage of HSR. This is particularly true for non-SF/non-LA--Sacramento, San Jose, San Diego, Las Vegas, Phoenix (latter two aren't in CAHSR's plans, but they're easily close enough to have branches and are constantly suggested as useful extensions--see above point). Even LA and SF are fairly bad in where the "downtown" station ends up relative to major business districts,
* As a megaproject, CAHSR ends up attracting lots of little related projects that balloons the cost. Not only is it about creating a grade-separated electrified rail from SF to LA, it's about building an entirely new underground rail station, etc. The project becomes about trying to justify why you need to spend billions more on other desiderata rather than thinking about how to make do without those add-ons.
No, it's really about reducing the growth in North-South personal and passenger freeway traffic and reducing expected capital and maintenance costs to handle that; that's pretty much the entire basis of the cost justification.