To be fair, Lightning works, and it works now on the Bitcoin mainnet.
The idea is sound, the code is written and somewhat tested, and it has been used in a very controlled fashion on the real bitcoin network already.
At this point it's just a matter of further testing and validation that the logic is sound and there isn't anything everyone is missing, and getting vendors on board and adoption started.
Still no guarantee it will happen, but it's looking more likely that it will at least make a good push, and it's no longer a vaporware promise of "in 18 months".
I'm struggling to find references, so apply salt as needed, but I'm fairly sure I've read multiple references to the fact that the LN developers themselves describe it as pre-alpha and still a few years away.
Which, to be clear, isn't necessarily vaporware - I think most people are in agreement that the concepts are sound, but that integration of the execution into the mainstream are quite a ways off.
It is still "early alpha" but bitcoin developers use that term as it was meant to be used.
"early alpha" is just that, early access to a fully working product with all the features that are needed, and it needs a LOT of testing and validation. And because of the "network" nature of it, it needs people or vendors that are willing to try it out before it can even be tested outside of controlled areas.
I agree that we are probably a few years (probably more than a "few") from Lightning being the "default" way to use bitcoin, but I fully expect by Q3 of next year there to be a few vendors which accept lightning natively, and a few "consumer friendly" wallets that have some integration.
In my opinion, I wouldn't want to call it "beta" until there are a few implementations with "user friendly" UI that abstract away the dirty details, and until there are at least enough people on the "network" that it would be feasible to actually use it.
If you want, [0] is a video from last week that shows lightning network running on mainnet. It's a bit of a "fluff video" with nothing really substantial in it, but it at least shows that it's working, and if you compile the tools he used in the video yourself (to target them at mainnet) then you could do it right now too, but without anyone to "network" with, it's kind of pointless right now...
The idea is sound, the code is written and somewhat tested, and it has been used in a very controlled fashion on the real bitcoin network already.
At this point it's just a matter of further testing and validation that the logic is sound and there isn't anything everyone is missing, and getting vendors on board and adoption started.
Still no guarantee it will happen, but it's looking more likely that it will at least make a good push, and it's no longer a vaporware promise of "in 18 months".