Twitter as a private company can moderate content as it sees fit. But the government is supposed to be constrained by the First Amendment. And here, they technically are, because they're not forcing Twitter to take action. But there's always the background threat of regulation if Twitter doesn't play nice, so it's not as black and white as this.
Having been on the receiving end of such requests I was pretty happy all of them were made in a confidential manner, it saved everybody a ton of headaches and reduce the amount of grip the various miscreants had on our community. Doing that in the open would have caused massive issues, some of these people were downright dangerous and others simply needed help, keeping their data and our interaction around that data confidential was - in my opinion - a good thing all around.
Just look at COINTELPRO for the most egregious example.
Then there's Waco, Ruby Ridge, the campaign finance corruption, infiltration of latin American governments, and modern operations similar to COINTELPRO to take down various movements like occupy wall street.
They are a mafia, not some benign group that can be blindly trusted with secret carte blanche powers.
But none of that seems to have any bearing on what happened here at Twitter, which is so far - to put it mildly, given the amount of noise made by Musk et al - underwhelming.
If you feel differently then that's fine but personally I think there are far bigger transgressions by the FBI than requesting that Twitter take down objectionable content and rabble rousers.
In NL Twitter was successfully weaponized by our most despicable - outright pro-Nazi - political party and the aftereffects are still felt today. Compared to what happened in the US it obviously pales but you have to wonder how the 6th of January 2020 would have ended if it weren't for the good contacts between the FBI and Twitter.
Not all such contacts are bad.