Only 5 developers, really?
The list is more than 5 https://bitcoindevlist.com/

And you can contribute to Bitcoin Core developments too. https://bitcoin.org/en/bitcoin-core/contribute/

Anyone can make changes. It doesn't matter. What really matters is: who runs your software?

For example, I shared the code for mining coins on a CPU in testnets. And guess what: some people applied all of my changes, including accepting blocks, up to 20 hours in the future. As a result, some nodes saw a chain reorganization of more than 100 blocks, because if you have a lot of CPU-mined blocks, then the chainwork is similar in different chains.

And then, I was quite surprised, when I saw ASIC-mined blocks in my fork. But well, CPU miners only provide some content for testnets, and then, ASIC runners can decide, which chain they want to build on top of, and that chain then becomes real, when the time of the network will reach the time put inside blocks.

So, am I a testnet developer now, because some nodes adopted my changes?

If you are a developer who contribute to Bitcoin Core Developments, like code it, report bug, you are one of maintainers.
If you are a translator who contribute to Bitcoin Core, you are a maintainer.
If you are a Bitcoin node operator, you are a maintainer of Bitcoin blockchain decentralization.
If you are a Bitcoin user, you are a maintainer too because you make transactions, spend fees to keep the network alive.

Surely you can find more than 5 people are doing one of these maintaining contributions.

Oh, that's me. I constantly make transfers in the Bitcoin network and not only. Each has its own advantages and disadvantages. Why choose only one Bitcoin when there are other coins, tokens? Many people just don't know about the existence or just don't trust? What's going on? What to do?

---

Naturally, there are many more Bitcoin network developers than 5. It has already become a truly grandiose project that has proven its viability and effectiveness to the entire world. Entire countries cannot be mistaken about this when they have adopted crypto at the official level.

Van der Laan voluntarily gave up his access to the bitcoin client software after being at the helm for more than nine years. As the second successor to Bitcoin’s mysterious inventor Satoshi Nakamoto, he was one of the few individuals having final commit access to Bitcoin Core’s GitHub repository.

Nakamoto first owned this administrator key and then transferred it to Gavin Andresen. When Andresen, who received direct instruction from Nakamoto on maintaining bitcoin’s code, stepped down from the project nine years ago, Van der Laan took over. Van der Laan has played the lead role in maintaining all aspects of Bitcoin’s development for even longer than Satoshi. He led all operations involving bug fixes, code reviews, upgrades, software maintenance, and dispute resolution.

Only five Bitcoin Core maintainers remain. With Van der Laan’s departure, Bitcoin’s development will now be steered by five people: Hennadii Stepanov, Michael Ford, Andrew Chow, Marko Falke, and Gloria Zhao.

Each is responsible for an aspect of Bitcoin; for example, Stepanov maintains the network’s graphical user interface, while Ford oversees the build system.

Gloria Zhao, the only woman in the maintainers’ team, writes and reviews the code that governs Bitcoin’s transaction validation process. Andrew Chow is in charge of programming for crypto wallets, which allow investors to store their bitcoin, while Marco Falke focuses on testing.

Together, these coders keep Bitcoin’s digital ledger up to date on its network’s thousands of computers. They must ensure that the software remains compatible with the latest versions of operating systems like Windows or MacOS and that it keeps up with transaction volumes.

Many of the cryptocurrency’s proponents claim that its current value and future potential are partly in the hands of these maintainers.

https://crypto.news/bitcoin-core-has-only-5-developers-left-as-key-maintainer-departs/

As far as I understand, in theory, major changes could be proposed by Bitcoin Core developers (which actually has more than 900 people in it not just 5, as many people has mentioned in other replies), but even in the case when most of them are in support of a decision, they can't push it through. All in all, any modification to Bitcoin must achieve broad consensus among the greater Bitcoin community including miners, node operators, and users. Since Bitcoin is designed to be a decentralized model, no single group can force an outcome on it. For any modification or change to be implemented and take effect, it has to be adopted by a considerable number of nodes on the network.

So, while the majority of developers could technically force changes, those changes still must be accepted by the global Bitcoin community through some consensus mechanism, in a decentralized manner such that no single group can force key changes by themselves.

yep the centralisation risk is not the propaganda fear of 'not many nodes exist'.. its actually how many nodes are too reliant on only one brand.

as for the car crash hypothetical, if all core devs died in a car crash then the core brand dies and people then rush to create a new github and fight to win trust of a new brand/group of devs that were not in a car crash. where this new brand suddenly gets to offer proposals for new protocol changes which people blindly follow

however the maintainers of core dont live in the same house and dont carpool to the same office together (unless they were at a convention/conference they all turned up to, that offered a minibus transport they all got on) so odds are if one died the other maintainers choose a replacement and just revoke privileges of the deceased one, and core continues governing the protocol

..
what should happen is multiple brands have proposals where not one single brand wins, but the best idea wins and all brands then follow.. but yea, core dont want to use the byzantine general fault solution. they prefer to keep the master general option

It is the case now, because you have a lot of different altcoins. The main problem with them, is that they don't follow the heaviest chain of SHA-256 hashes, and they abuse 21 million coins limit, by creating new coins, not connected with Bitcoin in any way.

In that case, you would have more than one competing chain. And then, you can have for example three branches, where for example chain A has 6 million BTC, chain B has 7 million BTC, and chain C has 8 million BTC. Because having no single winner means, that some group will say for example "we want Segwit", and another group will say "we don't want Segwit". The same with Taproot, the same with Lightning Network, and the same with OP_CAT, and every other proposal. And then, the only case, where there is no single winner, is when every group can enforce it at the same time.

There is no way to find out, which way is "the best way". For example, someone may argue, that "Segwit won, and all brands then followed". Then, you may disagree, by saying, that "it is not the best idea". However, in consensus rules, you don't have "the best chain". You only have "the heaviest chain", which means, that some idea may be worse than others, and still win (see: Ordinals; it was possible to reject them, but the idea of stopping them didn't reach 51%, and now only some pools reject them, while others endorse them, and have no plans of turning them off).

The first step to change anything, is to know, what is available. So, which alternative client would you recommend? Where it can be downloaded?

you really have no clue..

proposals are not actual changes/forks.. proposals are the idea of a change
there are many ways to show which idea wins before then doing a activation process of the actual code change, where by multiple brands are then all following the same winning proposed change of code.. thus uniting multiple brands on the same blockchain without needing to seek permission from one brand upfront to do the proposal/election part(governing the protocol)

learn more about how the blockchain invention can actually be more decentralised without having to rely on one brand