Lesson? If a project depends on teacher training, it will likely fail. Hard to think of a greater indictment of a profession.
Here's the analogy I would make: let's say you invent a novel method of preventing malaria, which is not only novel and inexpensive, it is self-explanatory. If you can distribute it to a population, they can figure out how to successfully administer it, and also, it is fun to do so! So you don't have to figure out how to train people to use the prevention method, and you don't have to mount an educational campaign explaining to people why it is necessary.
That's the theory at least, but when you do a pilot, you discover that, in fact, how to use it is not as obvious as you'd hoped. So now your cost structure is not just "distribute this thing" but "distribute this thing and figure out how to teach every village doctor and nurse in the country how to train the population to use it." It may still be possible and worth it, but it will sure be a lot more difficult, and you have to reconsider some basic premises, like, "if I can teach everyone in the country to do something to prevent malaria, maybe it would be cheaper to teach them to use mosquito netting instead of this goofy new thing I invented."
Getting back to Kusasa, there is also the particular complication that training teachers to use technology creates an incentive for them to take a better job in the tech sector. This exists in the developed world, but is even stronger in the developing world, types the former teacher turned technology consultant into his blog.