Nothing about this requires blockchain. There are plenty of Twitter archive sites like nitter.net that are not only open source, but also have plenty[0] of distributed instances on good old "web 2"
For example, you cannot see who liked a Tweet on Nitter.
That's exactly what you'd expect from a Twitter archive site. The author's point was that this is the same tech you'd need for a Twitter replacement, and this doesn't need blockchain or web3.
Reminds me of the comment on HN from a few months back "On web2 music artists only get 18% of the revenue. Spotify gets 33%. The music industry takes 49%. Web3 flips the script." You don't need blockchain or web3 to fix that either - centralised solutions like bandcamp.com already give the artist 80-90%.
The problem is that whoever builds the replacement can just fuck up the user experience any time in the future. Just like Twitter did.
To look at it another way, how would web3 actually help here? As far as I can tell it would only make this example worse - a more user hostile UI (faffing around with wallets and keys and whatnot), less decentralised (everything going through one API to interact with the blockchain), more resistant to change (forking blockchains is a big deal and only tends to happen when rich people at the top of the pyramid don't like something), more expensive (gas fees), less privacy (your wallet address being essentially an undeletable cookie), more likely to incentivise divisiveness and conflict (flamewars driven by the expectation of personal profit rather than just likes), etc. etc.
That's a positive feature.