Essay writing for me begins, and nearly ends, with Annie Dillard. That’s personal: she happens to be the writer whose essays I read when I started reading essays, and they ended up moving me.

She writes mostly about nature, observing it through the humble lens of the enthusiastic non-expert. Pilgrim at Tinker Creek seems like the story of a year in the wilds, though she’s only describing the adventures of a few backyard species in suburban Virginia. Her attention to detail and her authorial grace turn these limited horizons into a land of discovery. It shows there’s life nearly everywhere, that appreciating it is a choice you can make: an attitude, a strategy, for living on Earth.

Her essay, “Aim For the Chopping Block,” still haunts me with its images of an unnamed Washington State island, and its metaphor, relating wood-chopping and writing—but more broadly any sort of exhausting endeavor—transmutes personal experience into useful knowledge.

I was reflecting on this in the wake of Paul Graham’s latest essay, “You Weren’t Meant to Have a Boss.” His writing has been moving at times; but consistently, he has taken an authoritative, condescending tone. He has often made up for it by offering to programming a dignity and glory that it rarely receives. But more and more he slips from a cocky-but-compelling voice into one that’s simply smug.

This essay starts with a dis on a crowd of coders he finds that don’t meet his approval in some unspecified way. “They are not founders,” he ends up concluding. He uses lions in the wild as a metaphor, but rather than painting a prose picture of those lions for us, he only says they “seem about ten times more alive.” In this way, and others, he stretches his personal experience, using it as a bullying-stick to put weight behind him without building an argument.

He also dwells on a claim of the form, “All the anthropology I’ve read says X.” I has hoping the footnote at the end of that paragraph would provided a citation, but alas, it’s only a clarification. He goes on: “Tribes of hunter-gatherers have more freedom. The leaders have a little more power than other members of the tribe, but they don't generally tell them what to do and when the way a boss can.” Quite frankly, I’m not convinced he’s done the relevant anthropological research; I suspect he doesn’t have a deep understanding of the social dynamics of “hunter-gatherers” in general, even if one allows that hunter-gathers’ various lifestyles could be generalized so easily.

I have a bias against big companies myself, for different reasons, mostly to do with my personal habits and wants, or political feelings about their power. I don’t agree with Graham’s argument against them.

He says that humans weren’t “meant” to work in large groups, and that the freedom of the individual is inversely proportional to the size of the organization. I don’t think either of these claims are quite right. One reason people make large organizations is to increase organizational efficiency. In principle, large companies should have groups that can share work and move in a coordinated way, without letting inter-firm competition block this coordination. It doesn’t always work that way, but it can. There can be a large team spirit, with whole groups building infrastructure for lots of other groups. And, Graham doesn’t pin down his ‘freedom’ or account for the fact that it might trade off against other virtues. What about the freedom to move to a different team within a big company? What about the chance to be highly productive by meshing with lots of other competent people, which might require sacrificing some personal freedom to muck about with code? Of course, the situation can be much worse, but Graham is wrong to conclude that individual freedom is always inversely related to organization size.

Some big organizations work better than others. I found this to be true when I worked at Amazon in the early 2000s: there were loads of sharp people there who were trying to make the development process better and who were eager to fix problems when they found them. I learned a tremendous amount at that time: about aspects of programming, debugging, and software architecture, about running an enormous website that has critical demands on uptime and performance. It honed my software practice. And when I stopped learning, I left and did other things. In fact, Amazon at that time was keen to retain some of the qualities that Graham calls for: from Bezos downward, there was a drive to reward individual initiative and empower individuals, which Graham says large organizations can’t do. Because it was done well, and because there was a tremendous enthusiasm about the company and about the web at that time, it worked.

I also think it’s false, what Graham says, that large organizations are strictly tree-structured. The official org chart is a tree, yes, but the actual structure is much different. People grow a much more organic set of relationships, and this can be interesting, dynamic and lively. My official boss was someone who checked up on me weekly to see that I was working, and he could stand up for me when something went awry, but I normally worked more closely with, and took orders from, people in another branch of the organization.

Paul Graham’s charms as an essayist should be taken with a grain of salt; he has a fairly broad personal experience, but he makes more of it than is necessarily warranted. He makes specious claims: ones that sound right but aren’t carefully backed up. Paul Graham was my guru for about a month in 2003, but no longer. The next time I want to read a strong and compelling essay, I’ll go back to the likes of Annie Dillard.