you can hard reset a git branch and erase all the data after that

First of all, that’s misleading. Yes you can do that, but only for your copy of the repo, or the remote if you can force-push to it. Other people’s copies can still contain the data you were trying to remove. Hilariously, this leads to certain difficulties Github tries very hard to ignore.

a git repository is usually centrally controlled by the few ppl who own it or are contributors to it

Again, demonstrably untrue. It’s trivial to fork and every copy is a full repository, with its complete history. This is exactly how people save projects that are removed from major platforms like Github, for example.

git repositories don’t have any inherent hosting in and of themselves (your git repo can be created offline and never be reachable by anyone, in fact, most git repos are like this)

How is this a bad thing? Not everything has to be always constantly accessible on-line.

Also, there is no “inherent hosting” for NFTs etc either: the images they are minted for are kept off-chain, after all! Which also leads to hilarity.

so, unlike a git repository, a blockchain could allow you store data that is

• immutable (undeletable data)

Demonstrably not true:

• https://en.bitcoin.it/wiki/Value_overflow_incident
• https://www.gemini.com/cryptopedia/the-dao-hack-makerdao

Is it more difficult to do than with git repos? Hard to say, because git repos and their copies are independent of each other, so there is no way to enforce deletion of some data as effectively, as it was done with the cases above.

It’s also interesting to consider who and how gets to make these decisions. At least in case of Bitcoin above it was developers pushing the change in code, and users just updating their software. Many might not have even known or noticed such a change came about.

It’s not hard to imagine a scenario where developers push such crucial changes (or a malicious actor tricks them to do so) and users just upgrade without knowing that now history was perhaps altered.

• publicly available for anyone to read

Git repos can also be public, and many are. Not sure how this is a game-changer you make it seem to be.

• publicly available for anyone to write (in a way so that it can’t be spammed to infinity and rendered useless)

No, not everyone. On a lot of blockchains the gas fees are becoming prohibitively high. So it’s becoming a “pay to play” situation.

• always online

Again, demonstrably not true, even for Ethereum.

which enables you to do things like authentication 🤷‍♀️, which you can’t do with git

Git is not a tool for authentication. But there are better tools for authentication that do not require a blockchain at all.

In the end, there’s going to be a database somewhere, whether on a blockchain, or a more traditional one. Blockchains are not necessary for this, and I don’t see how they bring anything useful to this particular field.

could you elaborate here? for example, how would you replace something like filecoin, which helps track where/how much/how long etc a particular piece of data is, and rewards ppl storing it with tokens?

You are aware that there is a whole data storage industry, right? As in, companies you can buy such service from? On plenty of levels, using plenty of protocols and software tools, with plenty different rates and plans and whatnot.

There is little reason to bolt a blockchain to this. It’s simply wasteful, and doesn’t actually solve any problem.

it’s true for any piece of software, and there have always been dangerous exploits, this doesn’t mean we shouldn’t use software

No, but knowing how error-prone programming is, we should really not pretend it makes sense to make software (smart contracts are software) and deploy it in a way that updating it to fix bugs is extremely hard and prohibitively costly. Which is exactly what happens with DAOs and other blockchain-based smart-contracts.

The problem is not that software has bugs. The problem is in insisting on trying to make software immutable and pretending this is a good idea.

you’re right about the distinction that with DAOs and blockchains, bc everything exists within the digital realm and most actions are completely irreversible (unlike, say, a server, which can be hacked, but you will still always have the physical machine with you), but this is the price you pay for decentralization, and in my opinion, it’s worth it in certain scenarios

Decentralization can be achieved without blockchains. Decentralized, federated protocols exists and have existed for almost half a century now. We’re using one right now to have this conversation.

And yes, “in certain scenarios” perhaps, but there are very very few scenarios where it actually makes sense. And so far blockchains with distributed consensus have been crammed into everything, for no good reason.

it’s a subjective point whether such risks are worth the benefits, and i think it’s fine to disagree about this one

Sure. Although I would posit that these risks (including the slow suffocating death of humanity on this planet that we are slowly cooking ourselves on) are very very hard to justify with any perceived gains cryptocurrencies/NFTs/web3 have actually delivered so far. Meanwhile, the CO2 emitted to mine all this remains very very real.

right of my mind would be radicle, which is a decentralized code hosting platform, or 3box, which provides authentication services, i’m a blockchain developer, so you prolly shouldn’t ask me

I really don’t see the benefit. The problem with centralized code hosting on platforms like Microsoft Github can be solved by… federated code and issue hosting. Git already lends itself to federated MRs/PRs, and there is work being done on issue federation too.

Blockchain-based authentication services just strike me as a solution in search of a problem. I can host my own OIDC/SAML identity service easily if I don’t want to use any of the large providers, and plenty of services are already compatible with OIDC/SAML.