As skilled as Josh may be as a coder, even when he started LT, he's still inexperienced at actually delivering a product. The problem with being intelligent, young and still learning is that every week sees new ideas which seem like great ideas, and every week sees the ideas from the previous week seeming inadequate. Add in some inexperience, and you get the reason that software architects and project managers are roles that exist rather than just letting genius coders do what they want 100% of the time. Having someone who knows how to deliver a project is often more valuable than someone who knows how to code shit into rainbows.
Without smart, experienced management, you get a constant turnover of the new ideas, new code and new structure to replace what was there before and while that can generate an endless cycle of "improvements", it isn't necessarily progress towards completion.
Maybe Josh has adequately refined that skill set in himself and is actually is on track to getting things out the door now, or maybe he hasn't. The only thing that will confirm that is an actual delivered product in some form, and there's nothing for anyone but Josh to do about that besides wait.
Personally, I don't expect anything of LT anymore. That has long been a baseless expectation and managing your own hopes and disappointments is much easier when you expect nothing. That said, I do look forward to one day being pleasantly surprised, but I'm sure as hell no longer waiting for it.
Similarly along these lines is another game I've been following for a few years now: Starbound. They also had a team of great coders, but perhaps not so great at being efficient project managers to reign in scope for what would be considered a 1.0 release and what to save for future DLC or just game updates - you can't realistically get everything and
the kitchen sink in by 1.0. Not even AAA titles achieve this. SB's scope has been all over the place, but is slowly creeping towards 1.0.
Difference there, though, is that we've been able to play-test the game all this time with an occasional 'stable' release as well as 'nightlies' that may work and may not. We have something to hang our hats on while the devs figure out how to cross that finish line.
Over in the Star Citizen universe (which is also taking a millennium to reach 1.0), folks are given 'chunks' of the game to actual satiate their need for play. (though small consolation for the SC crowd)
In a similar boat to LT is Wollay developer of CubeWorld. Basically a one man team and boy is it taking forever for release. We did get the alpha to play with all this time though.
The big difference here, too, is that LT has one
developer, so you can basically stretch development out 10 or 20 times what a regular group would deliver.
So, what's my point here? Who knows.
I do believe that a game that's a one man show is not something you want to hold your breath for. It often becomes that one man's life's work to get exactly perfect before he lets it out the door. If a group of 5-10 take 3-5 years to get a game out.. well, you do the math.