The whole concept of roguelikes is so ancient and many developers think it should remain like that. I don't agree with that. That's why it's easy to surpass them. And I'm not even doing anything special, just thinking that there can be something else than copying 1:1 old roguelikes like everyone seems to do.
I agree the roguelike genre needs a few shots in the arm. In some cases it's happening as many games become popular by being less roguelike than roguelike-lites and borrowing only a few roguish features. Popularizing the RL labeling (even when incorrectly applied) has attracted some new roguelike devs interested in making what's old new again.
I'd disagree that it's easy to surpass long running projects like Nethack. It may be easy to come up with better designs, like this Info Line Idea (now with visible history), or Gods that employ Machine Learning, etc; However, at some point one has to factor in all the dev hours it takes to get something as complete as one of the big old games out of the dungeon and released. Therein lies the "difficulty"; Even more so for One Man Army development teams.
The big old roguelike projects have had the benefit of having many devs and/or starting out modifying a working game (one way to get devs), or at the least borrowing heavily from an existing working design (but forever being a slave to past design decisions). There's a reason so many devs start out (re)making a crusty "classic" roguelike, and there's nothing wrong with that. It seems impossible for most devs to avoid coming up with fresh ideas while working on any old project, whether they can make them a reality is another story.
For instance,
here is the W.I.P. tile sheet for the Roguelike that Derek Yu was developing when he came up with the idea for Spelunky and made that instead. I personally wouldn't say that Spelunky surpassed Nethack or ADoM as a roguelike since it's so far from the original genre and is much closer to
Notch's Infinite Mario than a RL, IMHO. It's the extreme example of adding new things to a roguelike (so much that it's not a RL anymore). If only more devs didn't pander to the Fakebook generation with such dumbed down gameplay we might get more innovations in the hardcore RL genre.
It's hard to compare any solo project to a decades old existing game. 1 Man Army successes seem to be at the cost of neutering much roguelikeness. Personally, I'd avoid spreading FUD about ambitious solo projects. They rarely materialize because of the time and effort required to climb that mountain, not due to any fault of the developer (except for being ambitious). Props to the "starving artist" devs that actually try to pull it off. We can use all the dedicated souls we can get.
P.S. I like shy Info Lines that autohide but aren't so timid they disappear before I'm done checking them out.