The vast majority of the time, if you build something that scratches your own itch, there will be others out there that will also find it useful. The challenge is to find a big enough itch.

I always find this to be terrible advice. Most programmers have similar itches, if you scratch your own itch, you are going into a lot of competition with many other developers. Why not solve a problem for an industry that gets much less attention? Maybe success rates would be much higher if we stopped all building the same things.

I never said anything about building tools for programmers. I'm actually a major proponent of building apps and tools that scratch non-technical itches, or more specifically, intersecting your development skills with your hobbies or interests.

unfortunately for those of us who don't have any non-programming/tech interests that doesn't leave much left. :)

(uber, taskrabbit notwithstanding..)

The subject matter of your software doesn't really matter. Building software is what is fun, using the latest tech is fun. Who the end user is, won't affect your day-to-day very much.

It seems that the companies coming out of YC now are scratching a very diversified set of itches, for example YourMechanic.

you could start with a replacement for grove.io.

heh. can someone recommend an alternative that also supports connecting via standard IRC protocol?

Try https://irccloud.com, just set up a channel on any public IRC server.