Give me a BRIEF example of what you mean
Assuming the volume dose is correct, and that the trainee can actually make these increases...
Week 1: 335 @ 9 = e1RM 398
Week 2: 355x5 @ 8.5 = e1RM 410
Week 3: 365x5 @ 9 = e1RM 434
Week 4: 375x5 @ 9.5 = e1RM 441
Week 5: 385x3 @ 8 = e1RM 447
or
Week 5: 405x3 @ 9 = e1RM 455
Is the latter a bad idea?
Right. So look at it in terms of percentage. You'd be going from sets of 5 @ ~85% to triples. In the first case, the percentage of estimated 1RM is not much higher than that, whereas in the latter case it's about 89%.
Would going from a nearly maximal set of 5 to a relatively hard triple, as opposed to an easy triple, result in less gains during the subsequent weeks?
IF you can time things right, I like the switch from a 5RM to triples to be perceived as a short "break". As we state in the book, it's better to switch to triples a little early than to switch to triples because you failed a set of 5.