There are many ways to batch things: you can have sidechains, you can use Lightning Network, and you can also replace mempool transactions with full-RBF, and combine two or more low-fee on-chain transactions into a single one, with a higher feerate. There are many possible options to scale things.

1. I said "second layers and batching". If you know, what are "second layers", and you know, what is "batching", then it is all you need.
2. The main idea behind all of that is quite simple: if you have one on-chain transaction per user, then you have famous "seven transactions per second", unless you are a big blocker, and want to lift limits, like maximum block size. Which means, that if you want to scale, then you have to handle more than one user in a single transaction; maybe even more than one user in a single UTXO. LN can use two users per UTXO here and now, in the future there may be more users per UTXO, because then, they wouldn't need to do cross-channel transactions to send payments between themselves.

Fees can go up and down, but sooner or later, minimal fees will be accepted. However, if you allow too low fees, then there will be times, when miners will get less coins than usual, and then you could have an option to send something with 0.001 sat/vB, but you will have higher chances of getting your transaction rejected, or less propagated between nodes.

Also note, that even if here and now, you will convince Bitcoin Core developers to release a new version, where 0.001 sat/vB will be the default, then people won't upgrade instantly, and some of them may configure their nodes differently. If there are pools, which can here and now accept non-standard transactions, even though Bitcoin Core disabled that option for mainnet, then there will probably also be pools, which will require higher fees, than allowed by default (by the way: some accelerators required at least 10 sat/vB, to bump your transaction, so people paid 10 sat/vB instead of 1 sat/vB, to have an option to accelerate things when needed).

Oh, even on bitcointalk, you can still find many users, who still think, that block size should be increased in the future. And you can even find articles like that: https://blog.lopp.net/treatise-bitcoin-block-space-economics/

So yes, in 2017, BTC decided to not scale on-chain, and to focus on off-chain scaling. But many users still disagree with that approach, even if they keep using BTC for years. And as long as some tools does not allow to batch transactions in a non-interactive way for those, who will opt into that, people will keep complaining, because they can see, that off-chain scaling is still in its infancy (for example because decentralized sidechains were rejected in the past, and many people are still actively against them; even though some opcodes like OP_CAT are seriously discussed, which would allow making such sidechains anyway).