As for sunk costs, this is a development phase, so we should expect changes as the new technologies and techniques are updated due to lessons learned. They begin a new version of starship, but when they realize that the design has problems or will not teach them much if it is launched, they find it easy to abandon in favor of a different new version.

Most of the other large organizations are playing it safe, using established, tried and true technologies and techniques. There is nothing wrong with this, except for the high cost, low launch cadence, and stagnant mediocrity; all three of which are antithetical to free market capitalism, which explains why so little business was conducted in space for the first half century of the space age.

Since these other large organizations are not developing new technologies, they are committed to the existing tech. If they are having trouble with this tech, then they must fix the problem — for them, there is no starting over. Or that is the mindset, and that is where the fallacy comes in, that they must continue on, because they have already spent so much and committed so much time. This is why the Webb Telescope cost so much and took so long, and why we lost many opportunities for additional science that will never be performed because the resources went to Webb instead.*

Being willing to do the development is why Blue Origin, Rocket Lab, and SpaceX are inventing the future. Other new companies are joining in and doing their parts to invent the future, too. Boeing, however, seems to be abandoning its effort to invent the future, as Starliner turned into a disaster for them; Starliner may not do anything other than complete its NASA contract. (Is this a new kind of fallacy? A ‘we spent so much on development that we won’t bother benefitting from the commercialization possibilities’ fallacy?)

It reminds me of an important lesson I learned during a programming class. I had a program that just was not working out, and I spent two days trying to make it work. I had too much time invested in this twisted, convoluted, virtually unreadable tangle of code to even consider abandoning it. Two hours before it was due, I finally decided to abandon it and take the hit for not turning it in — there went my “A” in the class. With two hours of downtime before class started, I chose to just mess around and start from scratch, maybe even hand in the assignment the next day or so; turning it in late was better than nothing, and maybe I could recover and still get that “A.” I decided that the sunk time (cost) was not working out and I should abandon it for a fresh start. Two hours later, just as class was starting, I had a working program and turned it in only five minutes after class started. What I had was simple and straightforward (though not as well documented as I usually did).

Yeah, I was just as shocked as you are now that the second try went so quickly and worked so well. It was as though I had learned what not to do and what to do, or something.
____________
* Sure, Webb is giving us fantastic science, but at what cost? We see the science that we are getting for our $10 billion, but we don’t get to see the science that $9 billion of it would have produced, had it been available and Webb had only cost twice its budget. Could that lost science have answered some, most, or all the questions that Webb is raising? Counterfactual history is hard to determine.

It isn’t the number of engines. It is the cost. A Starship is expected to cost around $20 million to build and a Super Heavy may be a little more, once production lines go into operation. The stack is cheaper than one SLS engine, at $100 million each, and Starship-Super Heavy can be reused many times. Disadvantage: SLS.

“A similar design could have gotten Starlinks up faster in a modified Elonergia set up.”

Energia was not available and still is not. Disadvantage: Russia.

“which means it’s tech might make for a lighter Starship as an aluminum bird.”

If only you understood why SpaceX chose steel rather than aluminum or composites. Disadvantage so far: Aluminum. Maybe another day; maybe another version (e.g. lunar lander); maybe another rocket or spacecraft.

The stage-and-a-half design is sound…it is just the wrong firm doing it.

A similar design could have gotten Starlinks up faster in a modified Elonergia set up.

Ironically, Boeing’s Space Freighter was to be a winged version of Starship SuperHeavy…that’s why I say that each firm really should have been making the other guys rocket.

Starship uses simpler tech that would be fine for expendable, where SLS shaves a lot of weight off—which means it’s tech might make for a lighter Starship as an aluminum bird.

Speaking of rocket tech:

https://www.eurekalert.org/news-releases/1004568
https://forum.nasaspaceflight.com/index.php?topic=59674.0

The first article talks about how hydrogen can be stored in a liquid that is itself flammable…the second link about generators.

Could this mean you could get two rounds of combustion from one liquid?

Maybe with this tech?
https://www.lifetechnology.com/blogs/life-technology-science-news/an-electrical-switch-to-control-chemical-reactions

There’s got to be a way to combine all this into a whole that takes lessons from different disciplines to make the ultimate RLV.

No doubt some or even many of these flaws could be corrected, but the cost would far exceed just starting over with a clean sheet of paper… like Starship.