I rather think there's well-placed skepticism (at least on HN) of the
decentralize-all-the-things crowd. That crowd gets louder on a monthly basis and makes ever more grand and urgent pronouncements about the need to decentralize all the things. Since there are many open research problems with decentralization in programming, this leads to these posts having more than their fair share of magical thinking.
"We are building distributed filesystems" as a new(ish) practical user control is one such example. IPFS has been around for much longer than a year, so I'm not sure what the author means by "past year." Regardless, it is slow. Regardless, isn't privacy-preserving.
Perhaps the speed issue could be solved with improvements to the software. Let's imagine it's fast, then-- what are we going do with it? We can't house Scihub there-- otherwise the nodes replicating it will get hit with lawsuits. Similarly, we can't build a new Demonoid or any other sharing service that lets students and scholars pool together the various copies of reference materials that they own.
At the same time, college students are constantly bootstrapping all kinds of private content collections on private centralized services, and sharing it among themselves (and, unfortunately, not re-sharing it outside their group so these collections keep getting rebuild opportunistically), jumping from app to app as tastes for apps change. That's achieved with the current internet plus all the walled gardens that people use to get their work done. If our distributed filesystem cannot include the content these users are already sharing amongst themselves, then it's functionally equivalent to a hard-drive that won't allow writing based on certain content hashes. Moreover, if users still have to do cartwheels to use the distributed filesystem in a privacy-preserving way, they may as well use their current walled garden apps that are probably decades ahead in terms of UX.
Yes, blockchain certainly has its share of hot air from finance. But this pairing of blockchain plus decentralize-all-the-things overlays (which will be done baking real soon now!) is particularly prone to magical thinking. One because blockchain fans typically handwave scaling issues (or pivot to a non-blockchain approach which isn't functionally equivalent to blockchain), and two because decentralization zealots are prone to releasing vaporware.
I hope that clarifies things. I've certainly reacted on this list with hostility to low-effort blockchain pitches from people who have a clear conflict-of-interest. My hostility is probably like other people's here-- based in knowledge of the disappearing act that these magic-thinking systems can perform on the money of ordinary folks. That's real damage, and it's bad. The day someone makes a repo for "sloth sort" by pasting a jpeg of a sloth in a readme for sleep sort and performs a pump and dump scheme on it, I'll be sure to be hostile towards that, too. Meanwhile, people seem to only be doing that with blockchain-based cryptocurrencies. That more likely explains the hostility here than the idea that anyone is (unwittingly or otherwise) biased against blockchain.
Edit: typos, clarifications