The problem isn't anything to do with the software. It's to do with the fact that if consensus fails, you need a failsafe, and there isn't one in decentralised blockchains.

Say for example, a group of entities gets 51% mining power in Ethereum and starts abusing that power. What do you do? You have only two choices: pray, or get the developers to forcibly patch out that 51%.

You are then trusting that the developers, who have this control, are in fact genuinely acting on the 51% - when you have no guarantee this is the case. All you need is to make people think there is a bad 51%, it doesn't have to actually be true.

Consensus systems that have no failsafe don't work by definition. It's got absolutely nothing to do with how you implement it; they fail because the people who are making the consensus have a motive for it to fail.

If a rich group or entity gets 51% control of a blockchain, you can't vote them out or hold them to account (and most of the time, you will never know they have 51% because it won't be on one account anyway). Your only failsafe is a group of developers who have no incentive to play fair when it comes to patching out bad actors.

On top of all of this, that e.g. Bitcoin software is open source doesn't mean everyone who mines Bitcoin is using an unmodified version of that software. You can easily make one that sends dodgy validations if you want to, because the only failsafe is that there are more other people giving correct validations; if that fails your whole system crumbles.