Crawford's big point is that the "knowledge worker" has basically had the life sucked out of him by a corporate culture in which there are no objective criteria of evaluation. Nothing that he does has an easily observable and demonstrable effect, so it all comes down to rhetoric and feelings. If you feel good and a part of the team, and everyone has a warm fuzzy feeling about their company and their brand, then you have done your job well. Crawford compares this to his motorcycle shop (or most trades) where either the bike runs or it doesn't. Basically, whereas breathless futurists (or educational reformers) have said that we'll all be knowledge workers in the future, so we better go to college and prepare to think for a living, Crawford is saying that this supposed "knowledge work" that awaits us is not Google, but Dilbert and the Office, and it is soul-sapping and bad in all sorts of ways.
"There is a pride of accomplishment in the performance of whole tasks that can be held in the mind all at once, and contemplated as whole once finished. In most work that transpires in large organizations, one's work is meaningless taken by itself" p.156One way that he attacks modern work is its reliance on algorithmic or recipe knowledge. Algorithms such as these, whether used to write abstracts for professional journal articles (a mind-numbing and stupid job he had for a while) or motorcycle repair manuals (a disaster when someone who does not know about motorcycles just copies and pastes from plans they don't understand, drive Crawford crazy, and illustrate how our modern society no longer values the tacit knowledge and expertise of the expert tradesman. This I can agree with ... to a point. It is certainly true for the extreme examples he cites. But he fails to acknowledge that there are still a fair number of jobs that are a mix of "knowledge work" and trade work.
This is where the Checklist seems superior and more sophisticated to me. There is a fair amount of tacit knowledge with surgeons (and the pilots, and construction workers profiled in Gawande's book) but a checklist is also an important supplement to their own expertise. This doesn't have to be soul-deadening or frustrating, as Crawford depicts it, but can free our minds to do the amazing pattern-recognizing that our expertise allows. Rather than dismissing the algorithm as comparing humans to computers and finding them not rule-following enough, Gawande shows that there are some situations which are so complicated that they need a checklist. Crawford has a disdain for "teamwork" in the corporate setting, he much prefers the solitary puzzle solving of him vs. the motorcycle, but he doesn't acknowledge that the very fact that the motorcycle exists is due to specialization, teamwork, and yes, some recipe following.
I do agree with Crawford that some knowledge work that by separating us from the effects of our labors, corrupts morals, inhibits learning, and degrades the purpose and value that our work holds. But I do think that some of this is necessary, and we should try to do the best we can with it (we are not going to back to small businesses making cars, TV's, furniture, appliances, etc). Also, there are a lot of interesting professions which are somewhere in the middle of the shop class vs. mathematical physics (a convenient straw man throughout the book is his dad, who offers pure equations and formulas, when the world of a 1983 VW carburetor has dirty nuts and bolts). Doctors need trade knowledge, but they also need to utilize the science of a knowledge worker. Teachers need to have trade knowledge of their students and what makes them seem happy, but also the science of memory and learning. If we could acknowledge that many professions are both trade and professional, instead of glorifying one at the expense of the other, I think we would be much better off.