The Bitcoin rollout in El Salvador was rushed. It's actually pretty incredible that it worked as well as it did with only months of development. I'm not sure what the hurry was, but it ended up not mattering too much since consumers could just keep using USD as usual.
Bitcoin doesn't scale at all, and there is no infrastructure that can fix that.
This isn't correct. Bitcoin scales amazingly, both as a store of value and as a currency for daily transactions. I've found people who claim it doesn't scale are simply unaware of how Bitcoin is architected. So I'm happy to provide that knowledge.
When someone says "Bitcoin" they might mean a number of different things, including the currency, the network, the software that runs the network or the entire ecosystem. I'm going to talk about Bitcoin the ecosystem.
Bitcoin includes a number of components. It is the blockchain, sometimes called "layer one" or the base layer. It is also the additional layers on top that allow you to transact BTC in different ways.
Separation of concerns is a cherished software design principle that Bitcoin follows. The base layer's concern is final settlement, security and privacy.
Speed of transaction isn't one of the base layer's goals, although it is much faster than other settlement networks. On the base layer you can have final settlement in under an hour. People might not realize that it takes banks and payment processors days or weeks to settle transactions. They hide this by sometimes giving you access to your money before settlement under the assumption that it will settle. These settlement transactions are grouped together into large transactions between banks. Sometimes banks have to move physical reserves in order to finalize which takes a lot of time. With Bitcoin, it's always under an hour.
Maximizing transaction throughput also isn't a base layer goal. Only ~7 transactions per second is the maximum on the base layer (it might be more with recent upgrades incl. taproot). This is a limit imposed by the way the blockchain is secured. By contrast, centralized payment networks like VISA can handle many thousands of transactions per second. And because base layer transactions are themselves a limited resource, transactions can be expensive for moving small amounts (although still cheap for large amounts; the fees are fixed per transaction not a percentage).
If this were the end of the story, Bitcoin would still be useful for infrequent aggregated transactions between financial institutions and as a store of value, but it would be a nonstarter as a payment network. Transactions are too expensive and throughput is too small to accommodate all of the daily transactions humans like to do. But it's not the end of the story.
As I said, Bitcoin is built in layers. Like the internet itself. With the internet, the base layer is the physical layer. It has certain requirements, responsibilities, and properties. On top of that are layers of protocols that add functionality, including IP, TCP, and specific applications. Bitcoin is like this.
The base layer, the blockchain, has the requirements and properties previously described. But there are layers and applications built on top of the base layer. Just because you're using an application or layer built on top of the base layer doesn't mean you're "no longer using Bitcoin". Just because you're using a web browser doesn't mean you're not using the physical and routing layers of the internet.
One of these Bitcoin layers is called the Lightning Network. This is the one that allows BTC to be transacted at scale, and it does so while also remaining distributed not centralized. In terms of speed and throughput, it will support millions of transactions per second. This is what El Salvador is using. And again, if you're using the Lightning Network (or any layer 2 or layer 3), you are using Bitcoin. You're transacting in BTC, not any other currency or token.
There have been disagreements about how to scale Bitcoin and cryptocurrencies generally. Programmers and software architects sometimes squabble. Some have argued that the base layer needs to have all the things, including scalability. Some set out to build what they think are better base layers. But this cuts against separation of concerns. The base layer doesn't need to be the be all end all. Just like we didn't cram HTTP into the physical layer or routing layer of the internet. Those fretting about the capabilities of the base layer are missing the point.
Are these second and third layers on Bitcoin just "tacked-on intermediary systems"? Maybe, maybe not. Some of the internet layers were also probably tacked-on to solve specific problems rather than being conceived all at once. Who cares? It works. It scales well. Could we build a better internet today if we rearchitected it with all of the collective experience we've gained? Could we build a better Bitcoin? Yes to both. Again, who cares? Bitcoin works well today. And there are many more layers and applications than just the Lightning Network, expanding the capability of Bitcoin. Because they are "Bitcoin", every bit as much as the base layer.
So there you go, free knowledge for you. El Salvador's rocky rollout was due to their own infrastructure being rushed, not any problem with Bitcoin itself.