Sure, technology fads are like a carousel, but that matters most if you insist on staying ON the carousel and make part of your living writing about the fads. I started in this millennium, but I've done it long enough to know that learning and using a new piece of tech is only superficially satisfying. What's exciting is building something that solves a problem. If I lose track of that and get bored, nothing is more motivating than coming across a coder who bangs out code using the tech he knows to solve problems. It's amazing to see what Rich Hickey can do with a pile of Java or Zed Shaw with a pile of C, to name two examples I've encountered. No need to complain or pontificate (or at least keep it to a minimum), just build things and solve problems.
That's why I love embedded - because the domain contains a lot of problems which are constrained by hardware (which so far is still holding to Moore's law), there are always a number of new problems for software to solve.
If you're bored of programming for the web, try branching out into a field where there's no debate about NoSQL, people aren't re-discovering binary trees for the ten thousandth time, and there's no re-invention of JavaScript on the server - try programming something real. For me it's quadcopters and fixed-wing UAVs. For you it might be something else. Embedded development is somewhere where real technological advance will always push the field forward. There's always something new you can build if only your CPU gets smaller, or draws less power, or some sensor becomes available. Plus, you'll learn a lot about programming close-to-the-metal, data structures, basic algorithms, and the industrial control algorithms that run devices in your daily life - which will both expand your horizons and make you a better web programmer, to boot.