I think that's crazy reductive though. It's not an either/or, and it's a gross oversimplification that big projects only show benefit within their major focus. LHC has produced benefit in particle physics, yes, but I am 100% sure it also introduced new methods of construction, semiconductors, material science, and other fields in the construction of the facility. I see this argument with space exploration as well, "why not feed the world instead of exploring space?"
Why is it zero sum to begin with? And more importantly, how less effective would be able to feed the world had we had no space race to begin with? If we were still at a 1950s technical level, our farming capacity would be exponentially lower, and most of the big leaps came as a direct result technologies originally developed for space.
It's naive to assume we can see all the benefits of daring to attempt hard things, but attempting hard things is what moves society forward. If all we did was lots of little cars in all directions when we have already figured out the challenges with little cars, mean the ONLY benefit is the direct one. If we've never built a car to explore REALLY far, well, there's a bunch of other problems we need to solve to get there. And honestly, it's those results that will likely move the needle.
In THIS case, a writer for the Atlantic with a bio of "Shannon Stirone is a freelance science writer based in the Bay Area" is criticizing. If we had an expert in planetary exploration doing the criticism, I think it would be a more interesting conversation, even if I feel that conversation would be FAR more nuanced than - "Mars bad. Elon Musk bad." Which is what I read here. If I agree with Elon's vision or not is immaterial - he's executing, I'm not, neither is the author. If I have a better idea, I should do it, or at least get it to a point of doable, then we can discuss if they are mutually exclusive or at cross purposes.
For us, it tends to be driven by the language/logging framework used. When using syslog-type logging messages, we log at critical/fatal for the REALLY bad ones. When using npm-type logging, we try to have the string CRITICAL in the message, triggering the alert on a single occurrence of such messages. General alerts have an x over y time trigger based upon usage/traffic/noise. We also generally encourage using warn as the default error state, only hitting error when things are broken.
In my last three companies, the bulk of the infrastructure was defined and managed via Python scripts (a lot of this predated Ansible being great), so what gets forgotten is the literal billions of lines of custom wrappers and classes that are broken, usually on the print statement v function debate or how string formatting works. I can't justify hiring someone to dig through all that code just to bring it up to snuff and everything needs tweaking. Usually we end up just writing new code in another language and call it from python or the other way around. I can't seem to get comfortable handling both versions in one project without getting REALLY frustrated. So, yeah, a fork with the niceties from python3 that allow my tech debt to still run (and hopefully better), allowing me to replace bits (likely into non python languages - Go is growing on me) at a time and not en masse is pretty frikken awesome.
The truth is, it's Python. I'm not looking for a great language, I'm looking for language that'll help me get work done... quickly and not give me a migraine if I have to come back to it (Perl). The 3 changes are just big enough to catch me (and every other Python programmer I've worked with) so that I can't just program, which tends to push me towards keeping it in 2.7 or switching to another language, where the similarities don't exist to pull me into bad habits. The whole mess really sours me on the language that made me love programming again.
You talk as if the switch doesn't look like a bunch of arbitrary decisions made on the appearance of purity. If more of an attempt had been made a maintaining compatibility, there wouldn't be nearly as much fight. Python 3 becomes a bit uncanny valley for me when I try to code in it and I tend to use a different language as there less of an internal code switch, especially as I still have to maintain large numbers of python2 code.