"People" have been declaring the death of 2D games ever since 3D games came into being. Hell, Sony even tried to outlaw 2D games of any kind on the PS1. And what's one of the most well-remembered PS1 games?

Castlevania: Symphony of the Night. A 2D game.

2D games aren't dead; they will never be dead. Even if you wiped every 2D side-scroller off the face of the Earth, there would still be puzzle games (Tetris, etc) and other simple 2D games (SpaceChem, etc). These would gain nothing by adding a third component.

Should I just drop my engine and find something meaningful to work on?

I would say that your problem probably comes from asking the wrong question.

You say you're making a "2D soft-body physics engine". That statement alone is part of the problem: you making an engine, not a game. Do you know what you intend to do with this "2D soft-body physics engine"? What is your gameplay? Why should people want to play your game?

If the only answer to that is that it has a "2D soft-body physics engine," then you need to rethink how you're approaching the task of making a game at all. You're thinking like a programmer: making something because it seems kind of interesting. You're not thinking like a game designer: making something because it could make for good gameplay. Programmer thinking can lead to good gameplay, but it's a lot less likely to.

"People" have been declaring the death of 2D games ever since 3D games came into being. Hell, Sony even tried to outlaw 2D games of any kind on the PS1. And what's one of the most well-remembered PS1 games?

Castlevania: Symphony of the Night. A 2D game.

2D games aren't dead; they will never be dead. Even if you wiped every 2D side-scroller off the face of the Earth, there would still be puzzle games (Tetris, etc) and other simple 2D games (SpaceChem, etc). These would gain nothing by adding a third dimension.

Should I just drop my engine and find something meaningful to work on?

I would say that your problem probably comes from asking the wrong question.

You say you're making a "2D soft-body physics engine". That statement alone is part of the problem: you making an engine, not a game. Do you know what you intend to do with this "2D soft-body physics engine"? What is your gameplay? Why should people want to play your game?

If the only answer to that is that it has a "2D soft-body physics engine," then you need to rethink how you're approaching the task of making a game at all. You're thinking like a programmer: making something because it seems kind of interesting. You're not thinking like a game designer: making something because it could make for good gameplay. Programmer thinking can lead to good gameplay, but it's a lot less likely to.