MIT Media Lab’s principle: if you’re building something you don’t know how to build, go to the thing you know the least about and do the simplest experiment to learn the most about it.
The reason: if you build everything you know first and save the unknown for last, what you learn from the unknown will force you to rebuild everything. People naturally do what they know and avoid the unknown, which is maximally inefficient.
Applied to self-discovery: don’t start experiments with what you already believe. Start with something completely contradictory—something you’re sure can’t be true. That’s where the learning is. If you’re sure talking less is the answer, first experiment with talking more. The nuances you discover will teach you far more than confirming what you already think.