No, I wrote that in 1998. It was published as the novelette "Lobsters" in Asimov's SF Magazine in 2002, made the Hugo and Nebula shortlists in 2003 (it didn't win), and later became the opening of the novel published in 2005.
I emphasized: the direction things were going in was obvious in the late 90s.
And don't *ever* let anyone tell you that Accelerando is techno-optimistic or pro-AI; by the end of the book our entire species is extinct, surviving only as simulations/memories recalled by something arguably not alive.
We really, really need a second run at Biosphere 2, only this time integrating what the failures of the first run taught us. (Only 25 years ago, with no serious follow-up!)
The next step would then be something like a Bigelow B330 module in LEO, which is close enough to get the astronauts home from in a hurry if something goes badly wrong.
(These steps can be commenced with current tech: Heavy/Starship not required.)
Step 3 would be a bigger test hab out beyond the Van Allen belts, preferably a couple of habs revolving around a hub to provide centrifugal "gravity" at Lunar or Martian levels. Goal is to test systems for use on planetary surfaces exposed to cosmic/solar radiation (because outside our atmosphere). Starship is probably mandatory for this phase, because it's a lot more massive and a lot further away. Alternatively: conduct this experiment on the Lunar surface, once astronaut return capability is available (but why waste expensive reaction mass if you can simulate a gravity well?)
Without a lot of R&D work under these conditions, a closed-circuit life support system for Mars is a huge safety risk for the astronauts who set it up (and who are too far away to rush home in a hurry if it goes badly).
And without closed-loop life support, a Mars "colony" is no more a colony than an Antarctic research station reliant on resupply for everything except air and water.
The path taken in the 1950s for spaceflight was dictated by its funding source -- missile development. Missiles were optimized for high performance and were single use devices.
Even when dedicated non-military rockets began to appear, e.g. the Ariane range and the Space Shuttle, they had dependencies on legacy infrastructure that was originally designed for strategic weapons (i.e. it was intended to work at peak performance, just once). The payloads evolved to reflect the launch system constraints, so very expensive, one-of-a-kind comsats and earth resources satellites, each of them a bespoke design (or at most one of a dozen or so).
SpaceX isn't building missiles; it's optimizing for reliability and cost (which means reusability). More like airliners than missiles. This doesn't mean low performance (the efficiency and power density of a modern civil airliner turbofan would have been a jaw-dropper to 1940s military aviation engineers) but it does mean the performance goals are different.
I am still extremely skeptical that Heavy/Starship will facilitate a Mars colony ... but that's because I'm skeptical about the economics and practicalities of building an off-planet colony when the externalities we normally take for granted (like a compatible biosphere) aren't available and nobody's really done the necessary R&D work on self-contained biospheres -- even the ISS is effectively an open-loop system dependent on constant resupply. (Biology and ecology are much harder than they look to a naive outsider.)
A sunshade is also a strategic weapon, from the POV of those nations whose insolation may be blocked (or even increased, using reflectors) at the whim of the foreign policy of the national power that controls the sunshade.
Not saying it's impossible or impractical, but the political consequences are non-trivial ... imagine being able to shave 10% off a nation's photovoltaic power capacity if they don't knuckle under to trade demands, or to mess with their storm frequency/weather patterns. It's a huge can of worms.
What I want is Banksian fully automated luxury gay space communism.
(You can quote me on that. I hate what tech has turned into.)