You do need to filter through some BSers to find a good job but that doesn’t mean any company that claims anything good about themselves is lying.
I guess... for one thing, I have, multiple times in my career, been part of technology organizations that have positively transformed their relationship with business stakeholders by getting the business to buy the underlying truth: building software is a learning exercise; the requirements can and will change.
And the frameworks you have to put in place to make that kind of relationship work, to replace the arbitrary deadlines and deathmarches? Turns out 'agile practice' has a bunch of useful ideas for those.
Teams have limited capacity for 'work in progress'. The upcoming work can be prioritized on a backlog. Loose sizing estimates can help us make decisions about trading off time and features. You can only have one top priority at a time. It helps to let teams understand what problem they are trying to solve, not just tell them to implement a piece of a solution.
That's what I value when I talk about 'agile'. A bunch of useful ideas stolen from scrum and kanban and lean and so on that actually help put developers in a position to solve the right problems in the right way, and get 'management' off their back.
I'm also a fan of daily planning meetings to facilitate collaboration inside teams. I like retrospectives as a way to identify ways for teams to improve how they work. I like short fixed-duration iterations which deliver working code. I like pairing on problems. I like continuous integration and deployment and automated tests written alongside the code. Those are good practices that we, as a small team of developers, can use to help one another do good work, and ship working software.
And those ideas are all also stolen from things like scrum and XP and TDD and devops, which are also all 'agile' practices.
If someone tells me they don't like agile I am confused because... I don't know why you wouldn't want to work on teams that follow those kinds of practices.
I do get that plenty of people have been subjected to environments that use things with the same names as these things in ways that aren't remotely agile. But throwing out the idea that daily standups are a good idea because you've been subjected to a bad implementation of them is like saying you never want to use a programming language that has strings because PHP's string coercion sucks.
I find these practices helpful, and better than not following them. I've been lucky to get to enjoy working in environments that use them, rather than have them badly forced upon me.
Then I took my role, and tried to implement the first thing they taught us at the training... but the management said "no". So I tried a different thing... but the management said "no" again. I asked what's the point of sending me to an Agile training if they override all my decisions, and the answer was, more or less, "we are doing Agile this way". My options were limited to do "Agile" their way, or to quit my role and be replaced by someone who will do "Agile" their way.
So now we are doing Agile the way everyone hates and everyone does, with me in the role of the bad guy. Except our standups are 5 minutes long, because... that's the only part that remained firmly under my control. Otherwise, everything is the opposite of what the Agile trainings and books said. The management made a big announcement where they congratulated themselves for successfully transforming the company to Agile.
From my perspective, this answers the question why "true Agile has never been tried". The people who make the decisions, they want the buzzwords and some of the rituals. They do not really want to give up any of their managing power.
To try the Agile as intended... you would probably have to start a cooperative. Or be a very enlightened founder who can resist the temptation to use their power.
I don't know how to make people stop lying about what they're actually doing (I wish I did), but I don't think it's reasonable to expect any kind of process to work when you do the opposite of what that process actually tells you to do. (And I'm quite prepared to entertain the possibility that an organisation that actually committed to following RUP or Six Sigma or whatever and followed through on it might be better than half-assed non-Agile "Agile". But I haven't found the HN-endorsed "lol no process" to be effective in practice either)