I think in this context he's talking about the 'beginner response' where strength gains are largely due to increased neuromuscular efficiency in UNTRAINED athletes.
I believe he says, without reading my own copy to make sure:
1) 5-10 lbs per 3-4 weeks in the presses is the normal rate of progression for novices for the first several months of training--until they have an intermediate level of strength.
2) weights will then increase in smaller increments, such as 1lb
3) then more complex programming is used
If you've been training for two years you have probably already used up your super-fast awesome beginner strength gains.
I see. That is unfortunate I suppose, but makes more sense. So I guess the question for me is (I think?) whether I should drop in weight for a while to really focus on proper lifting form. Reading through his descriptions, I seem to already have happened on close to correct form just through experimentation, but I am missing a few things like the glute squeeze during bench press. I think I'll probably spend a month or so nailing down the right form for everything and then read through Mark's book to figure out what my long-term program should look like.
I've just read the Four-Hour Body by Timothy Ferriss. It seems on the face of it like ridiculously valuable material, if true - like what the completed version of Michael Vassar's proposed reboot of dietary science would look like at the finish point if dieting turned out to be more susceptible to Munchkinism than in my wildest dreams. Ferriss also talks the rationalist talk quite well in this book, much more so than in Four-Hour Workweek; he cites the experiments and occasionally says things like "I spent a lot of money on this and I expected it to work and it didn't work at all" or "I tried this and it seemed to work and I have no idea why it worked and I think it was probably a placebo effect."
Does the LessWrong hivemind have an opinion about 4HB? Has anyone tried it and found that it doesn't work, or that it does work, or that it works but not as well as Ferriss thinks it should work?