Tiny, light-weight drones are very suitable for indoor drone applications. They are inherently safe for humans, can navigate even in narrow spaces, while still capturing the information necessary for the application such as warehouse stock tracking or search-and-rescue. For such applications, the tiny drones will have to fly completely by themselves. This is very challenging, since they are very restricted in terms of sensing, processing power, and memory. I will present the research we performed at TU Delft to achieve fully autonomous flight of swarms of tiny drones. First, I will delve into the intelligence of single drones, such as the DelFly Explorer, a flapping wing Micro Air Vehicle that with its 20 grams is still the lightest drone in the world able to fly around and avoid obstacle completely by itself. After that, I will discuss how we achieved to let the drones sense where neighboring drones are. Sensing other robots in a local neighborhood is a very common and fundamental assumption of much of the theoretical work in the area of swarm algorithms, but is very hard to achieve on small flying robots. Finally, I will describe the remaining major challenges on the road towards swarms of small flying robots.