So I'm no expert, I only worked in this area briefly and not as the business owner. And don't work in consulting anymore because I didn't appreciate the kind of work it was, so I'm biased. And I didn't say it was a "bad" kind of client/project, I said it was dangerous.

But other kinds of clients include:

* An existing successful business which needs more development or a new project

* An enterprise which needs things for internal use

* Universities, or non-profits

* Government contracting

If you are going to be developing for a software startup, being aware of the risk may help you manage it. Do _you_ think they have a possibly successful business, does it make sense to you? Of course, maybe you can't judge that, it's not like you're in the business of judging startup potential and it's not like that's easy. But does the founder seem to know what they are doing? Again, you may not be able to judge, but if you get a bad feeling it may not be worth it -- or if you get a bad feeling while in the middle, you may want to find a way to fire your client. (Even though it's hard to give up the money).

Will they be the kind of client that wants to tell you exactly what to implement and have you do exactly what they say? Or will it be a partnership where they identify requirements but trust you to work out some details? Either can be risky, but it turns out "we implemented exactly what you told us to implement" isn't always enough to stop them from blaming you for their failure, and I think in some cases can paradoxically make it worse.