What he meant is that lazy programmers will make the effort in the short term to make their work more accurate (as in meeting the requirements), more bug-free, more robust and more scalable in the long term. So over time they end up doing less work and far less drudge work. So he was talking about a rather creative form of laziness.
Some of the useful techniques I learnt were like writing scripts to automate tedious steps. Writing macros to render the code clearer. Writing small DSLs (mostly a *nix concept) to raise the levels of abstraction to better align with the domain.
Good programmers will minimize the time they spend over the lifetime of a project. This is why you don't optimize prematurely, to avoid wasting time when you already have a solution that is good enough. However, the same programmer will then be diligent about writing modular, easy-to-read code, so that editing and refactoring will be as simple as possible.
Relevant - http://paulgraham.com/goodart.html
They don't just accept how algoritms from life are. They think how they could improve all the time, that is more than a normal programmer will think just while in the work in optimization