"No, starlink is aiming to serve musk’s ego. Just like the hyperloop and colonizing mars, musk will spend lots of taxpayer’s dollars on "visionary" (as he sees himself) projects…"

True enough but say what you will about Musk – and there’s a lot to be said – at least the man dares to dream. This is something of a novelty in a US where "No we can’t!" replaced "Can Do!" some time ago, on every topic.

Starlink will remain a novelty since, as predicted, it doesn’t scale. The engineering experience of a satellite constellation in practice will be invaluable, however, for space exploration in general.

If there’s a single utter pie-in-the-sky project here it’d be Mars. That planet has no magnetosphere and so is exposed to the solar wind. Any atmosphere extending too far from the planet surface will be ionized and ripped away – so the idea that humans will ever walk that surface without a spacesuit is a pipe dream.

With that being the case any potential future Mars colony will consist of a very small amount of humans living underground for keeps.

"(A) Launch rockets with complex satellites and solve all the problems with collisions with other objects, power, bit rates etc.; or (B) Purchase a lot of fiber and just do the unexciting job of laying it and using mundane equipment."

Neither of which is sexy enough and both of which require far more initial investment than him using every launch to leapfrog a boatload of small, cheap and simple cubesats into a barely suborbital velocity around the globe.

"Personally, I think (B) wins without debate and is also a sure thing and doesn’t interfere with telescopes or existing objects floating around in space, not to mention the (literal) fallout when those thousands of satellites "fallout" of the sky. "

A land line of well-placed fiber can serve for upwards of a century without much maintenance and is cheap to fix. The fallout is less of a problem since those cubesats will literally evaporate upon reentry. I’m more concerned that the amount of launches required to maintain the network will exceed by orders of magnitude those which already in 2020 prompted environmental agencies to take a closer look on the impact on the environment.

The Falcon-9, for instance, generates massive amounts of CO2 in itself. Other launch vehicle propellants punch holes in the ozone layer. Others again are just highly toxic and slow to degrade.

The main issue, though, remains that the satellite constellation fscks up wide-range astronomy. Orbital telescopes like the Hubble are hideously expensive to put in orbit and thus the few we have are focused on pinpoint examination of objects of extreme interest. Wide ground-based arrays do most of the heavy lifting here, and that includes trying to find large rocks about to pull a dinosaur killer on the human race. I can just imagine alien archeologists pinning the label of "death by stupid" on the long extinct human race once they found out we screwed up our early warning capacity by putting on an orbital blindfold over commercial hype.

"Real innovation and engineering involves solving complex problems in the simplest way possible. "

Well, there always has to be That One Guy (not the account on these boards) building a Rube Goldberg engine just so that any sane engineer with means feels personally offended enough to design a better version. Like early DOS/Windows prompted Torvald to build the ancestor of what currently powers 90% of the world’s servers. Musk fulfills the role of MS there eminently. A shame that his only real adversaries in the field are China, the US largely having given up on the future.