Kent Beck. Foreword
What I notice in practice was that the more change I anticipated, the harder it got to make changes. My incorrect speculations interfered with the changes I actually ended up making.
I first experimented by ignoring any changes that seemed like they would happen longer than six month in the future. My designs were simpler, I started making progress sooner, and I stressed less about the unknowable future. I shortened the time horizon to three month. More better. One month. More. A week. A day. Oh hell, what happens if I don't add any design elements not demanded by the current code and tests? Still more better.