Also, Archaeologists love the garbage dumps and cesspits of old towns, literally the places where people put their least valuable things, because they weren't raided by earlier visitors and they can derive so much info about it. And it's nicely stratified so it gives some rough chronology.
That's why you need to catalog stuff.
And if you have stored something, you can always get rid of it if you deem it to be unimportant, but if you haven't stored it, most times you can't get it back. Erring on the side of storing unimportant things is an important strategy to cope with that.
It would be polite if ArchiveTeam were to now contact everyone whose profile they have scraped, and ask for permission to retain that data. And then delete all the profiles for which they didn't get affirmative consent.
For two, being public and publicly accessible doesn't mean it isn't gauche to scrape it. It's kind of like how nobody sticks a sign that says "take one only" over the bowl of mints at a restaurant; it's assumed you just know that it's not cool to stick a whole handful of them into your pocket.
IA aren't scraping the data themselves, but they're the customer.
As for your second point: there's merit to that argument, and I've discussed same previously -- I'm very much of mixed minds on this. A few considerations weigh strongly, however.
0. The data were already public, as noted.
1. The system shutdown was not a known factor when most of the data were created. The expectation at that time was that the data would continue to exist.
2. The shutdown itself has occurred in a context in which individuals, and far more importantly groups quite literally could not archive the relevant data themselves. Google's own Data Takeout, whilst fairly remarkable (in a positive sense) within the industry makes many things difficult or impossible. Ordinary users cannot archive Community content, and even Owner and Moderator roles within communities could only archive posts from public communities -- neither comments nor private communities were archiveable. (Third-party tools could provide these capabilties). Moreover, technically, cost, bandwidth, or storage-constrained users or communities largely had no viable options for saving their own legacies.
3. The contents sitting on Google+, indexed and searchable by content on both the site and via the public Web, were far more visible than they will be at the Internet Archive, which does not support full-text search of its archives (at least not yet), and which is not as effectively indexed publicly as Google+ was.
4. The Internet Archive does provide for content removal under the DMCA, as well as other mechanisms. For a G+ user, given how content URLs are constructed (they all include the user's G+ UUID as a common element), requesting removal of an entire tree is trivial.
On balance, this favours the Archive.
Often this argument is used to try to justify things as important as mass-surveillance and as small as logging in a software project. “Just collect and retain everything, who knows what we’ll actually need!” With GDPR and increasing focus on massive & intrusive data collection, I think this mindset is going to have to change. Before deciding to preserve or collect information (particularly information of a personal nature, like social media accounts), you should be prepared to justify the activity. “It might be useful one day, maybe” shouldn’t be good enough.
What seems unimportant and insignificant to us now might become very interesting in the future. People in the future might consider the arrival of the internet as the beginning of a new age. Furthermore, some things may seem obvious to a person born in our times, but might not be obvious to people born in the future.
So many works, including even movies from the 20th century are lost now. E.g. the original version of Metropolis used during its premiere screening. Diogenes wasn't just a guy living in the town square, he also created some written works. They are lost now.