The cost of slop; Dartmouth provost undone; publisher AI controversy: Newsletter 5 November 2026
Newsletter 175. What is workslop doing to your organization? Plus, how AI is undermining professional services, the Andy Awards shortlist, three people to follow and three books to read.
Program note. This newsletter was published on day later than usual to avoid being swamped in the discussion of the US midterm elections.
The cost of slop
AI writes quickly and with low effort. The slop that it creates is bogging your company down. Writing slop faster is not an improvement in productivity, it is the opposite. This is a hidden sclerosis in every company and leaders must address it.
I don’t have statistics to prove my case. But I have lots of anecdotal evidence.
Anecdote 1. Amazon has a unique writing culture. Every meeting intended to make a decision starts with a six-page memo. Powerpoint is banned. Those present read the memo together, and then discuss based on what they have read and understood. The act of writing the memo forces the person convening the meeting to make their case.
Or at least it used to.
In his Substack, Steve Hyunh describes what happened at a recent engineering meeting there where a product manager used AI to create his six-pager.
Thirty minutes into a recent review meeting at Amazon, [a] principal product manager began a discussion on a six-page document. Two minutes into his presentation, it was blindingly obvious to everybody in the room that he was reading the document for the first time. . . .
The . . . PM from above may have produced a six-pager, but because he had skipped the critical thinking step, he didn’t have a credible plan. I don’t know how the situation ended, but there are only two ways it could have gone. Either the plan got reworked, and a room full of senior engineers lost time to a meeting that should never have been scheduled, or it didn’t get reworked, and that team executed a plan nobody had thought through.
Amazon’s unique writing culture made decision-focused meetings productive. AI slop is undermining it.
Anecdote 2. According to an article in Fortune, Shopify’s CEO Tobias Lütke has had to backtrack on his recommendation to accomplish more with AI, because his workers are hurling “slop grenades” at each other.
He told employees last year that using AI is “a baseline expectation,” and they should first prove they “cannot get what they want done using AI” before asking for more resources. He’s now saying Shopify employees are producing unexamined emails and code with AI and not taking responsibility for the sloppy output.
“We call those ‘slop grenades’ that people toss at each other,” he said during an interview on The Knowledge Project podcast on Tuesday. “That’s definitely a bad thing.”
Lütke said it’s easy to let AI “go nuts,” but abusing it creates more headaches for the people receiving and reviewing the work. For instance, he suggested AI use is supposed to help synthesize points in an email rather than turning it into “a big missive” that wastes time.
What Lütke calls “a bad thing,” I call disaster. His people may are saving time writing at the expense of wasting the time of those who have to read what they “wrote.” Writing is the lifeblood of your business, and slop is cholesterol.
Anecdote 3. Microsoft employees are generating documents with no actual significance or value. From The Guardian:
At Microsoft, which rolled out AI widely internally, executive vice-president Ryan Roslansky recently publicly lamented the prevalence of AI slop in his day-to-day work. After receiving “what was clearly an AI-generated doc” from a co-worker, he used his AI to summarize it and came away with no new thinking about the idea, he said. He warned his followers against falling into this potential “doom loop”, where AI does the reading and writing and leaves behind an “ocean of sameness”.
“Workslop” stresses your employees, wastes time, frustrates managers, and makes everything harder. Imagine that you are attempting to cross a brook by stepping on what appear to be steppingstones — but half of them topple over or slide away when you step on them. You can’t tell what stones you can depend on which only appear to be helpful but in fact are useless. That’s what it’s like inside a company that allows its employees to go wild with AI-generated slop documents and emails.
Part of my life is dedicated to helping companies avoid this trap. It requires putting in place a collection of policies that and training that help people write better, with more meaning and effectiveness, in the materials they create at work. I’d like to tell you in one essay exactly what to do, but this problem is bigger than one post can cover.
But if you manage a group, a department, or a company, start with a few commonsense ideas:
- Reward action, not production. Now that AI can produce slop documents effortlessly, metrics like documents produced, reports created, or any other production-based measurement are pointless. Reward action — decisions made, clients onboarded, products launched. Documents and other communications that help with these metrics are a means to an end. Slop documents don’t count.
- Train on the difference between AI preparation and AI production. AI is fantastic tool for conducting research, compiling data, and suggesting organization for documents. It’s also pretty good at finding flaws in drafts. But writing and the thinking that it requires aren’t tasks that should be outsourced to AI. Training that supports these preparation goals will make your staff more productive, but such training should emphasize the difference between preparation and AI slop generation.
- Invest in better writing practices. There are plenty of good books and training programs on better writing. My own Writing Without Bullshit is one of them, but any of them are better than just punting the task to AI. Emphasize effective practices like front-loading meaning, avoiding jargon and passive voice, chunking with bullets and headings, and brevity. Better writers will develop facility with communication and avoid the temptation to “let AI do it.”
- Highlight effective communication. Hold useful writing up to the light — praise the people who created it. Ask them how they used AI to prepare, rather than to write.
- Ask potential slop authors, “What does this mean?” Any reader to whom a document is directed should have the right to ask for clarity. This especially applies to senior managers attempting to make sense of potential slop from their staffs. “This is slop” is the wrong message. “I couldn’t make sense of this, can you please clarify” sends a better message: that writers who create meaningless slop will be forced to spend time clarifying it. They’ll get the message soon enough.
News for writers and others who think
In Wired, Miles Klee asks if the commercial internet ought to be allowed to sink into a swamp.
Jory Des Jardins reveals how AI is perturbing the professional services industry — how can you charge by the hour for work done by bots?
danah boyd suggests that the problem with teens is not their social media use, it’s all the other relief valves and social spaces they’ve been deprived of.
The provost of Dartmouth, Santiago Schnell, resigned after admitting he used AI to create some of what he wrote, including a column in the Washington Post.
Adam Morgan wrote about how authors are frustrated that their publishers are using AI to generate marketing materials, even as they insist authors had better not use the same sorts of tools.
Gotham Ghostwriters announced the shortlist for the Andy Awards, which recognize books written with the collaboration of a ghostwriter.
Four of insightful commentators ( John Warner, Jane Rosenzweig, Jessica Weber Metzenroth, PhD and me) will be addressing “Writing Without Slop” in a special webinar at 1pm ET on Friday November 20. Sign up here.
Three people to follow
Luiza Jarovsky, PhD , AI governance expert
Barbara Caraballo , who’ll manage your book launch
Frank X. Shaw , head of communications for Microsoft
Three books to read
Bad Writing in the Modernist Novel by Nathan Wainstein (Johns Hopkins University Press, 2026). Is it avant-garde, or does it just suck?
The Collision: What AI Does to Us by Eric C So (W. W. Norton, 2026). Is AI teaching you to think better, or not to think at all?
Hyperscale: The Ambition and Excess of Big Tech’s Data Empires by Paris Marx (Riverhead, 2026. Are data centers boiling the planet?