You should first ask if you really even want to make something. As developers, we like building stuff and solving problems, but sometimes you truly are just “forcing it”. The biggest problem for me isn’t deciding what to build, it’s finding the time to make something. When my time is so constrained, I find myself prioritizing higher pursuits in my life than just hacking on projects.<p>Personally, I only make something when a problem becomes so annoying or time consuming, that I’m forced to make it. This is why you should always do things the hard way or manual way first before automating it and only then can you say “Wow, this could make money”.<p>Contrary to what others have mentioned, I advise against keeping a notebook on you at all times. You end up obsessing over things and overdeveloping the idea and much of it is wasted effort. If a concept for a project/idea you randomly think of is truly good, you’ll remember it, otherwise it wasn’t really that important. The exception to this rule is if if the scope of what you want to do or the problem is very large and requires tons of research, such as developing a new programming language or OS, etc, and not just an app or farming ads.<p>Overall, I would say to keep your process organic and invest more time in learning new things than you do on gambling on what may be half baked ideas. When I’m not sure what to build and have time to, I just buy a book on a subject I’m interested in, that has some overlap with the idea and but more so with what I’m interested in as an individual—-hell, maybe consider writing a book. Otherwise, be very experimental and conduct small hypothesis tests, starting with your largest to smallest, to make conclusions and _do_ keep a notebook of them. This is simply a matter of taking inventory of what assumptions/conjectures you have and either rejecting them or failing to. What I usually do is take an assumption and do the exact opposite; if, say, you wanted to make a new chat app, then write an essay on why making a new chat app is a bad idea and publish it. You’ll be surprised how much software out there is “broken by design”.