The Friday Binding — What stays human at work
Five reads on AI at work, the Pope's warning, and why most rollouts fail for reasons that have nothing to do with the technology.
Substack Coach Community posted a Note this week that I restacked eagerly. The point:
On most platforms you fight to stand out from everyone in your space. On Substack, writers working the same territory can grow each other, because they share an audience. Think alignment instead of competition, and the place opens up.
That has been my whole week. Wednesday’s essay pulled three writers into my space to work a single problem. This Binding does the same thing. Five writers, four of whom I had never read before this week, all working through the same question from different doors.
The question: once AI does the routine layer of knowledge work for pennies, what is the human layer actually for? And how do you run an organization that knows the difference?
1. Hollis Robbins, Anecdotal
Eric Patterson sent me this one, and it set the frame for everything below it. Robbins notices that no one ever pinned down the “higher” in “higher education.” Higher than what? She argues that large language models are forcing the question, because a model that has read everything can see when a word is being used against its own meaning.
One image from the piece lodged in my head. AI is a high jumper. Whatever it clears becomes the new floor. Breadth, summary, the general-education layer, all of it now sits at floor level and costs almost nothing. What remains above the bar is the human work: depth, judgment, the sustained relationship between an expert and a hard problem. Her conclusion is that if “higher” does not come to mean deeper, it stops meaning anything at all.
Swap “higher ed” for your org chart and the essay reads the same.
2. Stefan Bauschard, Education Disrupted
Bauschard reads Pope Leo XIV’s Magnifica Humanitas as a non-religious reader, and the first thing he flags is the form. A 1.4-billion-member institution that thinks in centuries produced a book-length treatise on AI within three years of ChatGPT. Most companies have produced a three-page acceptable-use policy.
His core point is the one every leader running a rollout needs to absorb. You cannot adopt the tool first and decide your values later. Adopt first, and the tool decides for you. The encyclical’s claim that AI cannot be treated as morally neutral carries the whole argument. Bauschard also disagrees with the Pope where he thinks the text overreaches, which is exactly why I trust the reading.
3. Al Dea, Work In Progress
Dea names the thing most adoption decks skip past. The gap is a management gap. He pulls Gallup’s 2026 global workplace data: employees whose managers actively back AI use are far more likely to say it changed how they work, and only about one in five workers worldwide strongly agree their manager supports the team’s use of it. Louder mandates and newer models do not budge it.
Four skills close it. Getting curious about your people. Helping them think about the work itself, beyond just doing it. Making a place where trying something feels safe. Learning out loud, where the team can watch you do it. For anyone steering a team through a rollout right now, this is the most directly usable piece in the set.
4. Rachel Sparkes, Love Your Work!
Sparkes comes in strong. Atlassian expanded its Chief People Officer into Chief People AND AI Enablement Officer. The AI work now sits with the person who owns culture and change, not the person who owns the tech stack.
Her argument runs through line managers. They are the culture carriers, and if they fear the tool, the investment turns into shelfware. One detail lands hard: Atlassian’s “strategic collaborator” tier means using advanced AI features at least forty times in a single week, and one researcher who reached it produced four times the output of the people stuck on the bottom rung. The role has a title now. That alone tells you where this is going.
5. Sharon Gai
Gai spent weeks inside IT governance, consumer goods, and automotive organizations, and came back with a blunt diagnosis. Ownership is missing. Few firms have a Chief AI Officer, and the ones that do often sit far from the actual work, so departments quietly spin up their own tools and agents and central governance collapses into fiefdoms.
She is just as hard on the dominant training model: buy a course library, assign it, call it literacy. And she refuses to look away from the cost. She points to Cloudflare cutting more than a fifth of its staff while internal AI use surged, the CEO describing the eliminated roles as “measurers.” The fear in those rooms is real. Pretending it is not is its own kind of failure.
What the five share
Read them together, and the pattern is hard to miss. Every failure here is organizational. The models work. What breaks sits upstream of the model: managers who haven’t bought in, stakeholders pointing at different goals while using the same words, tools bolted onto workflows that went unexamined. RAND puts the AI-project failure rate above eighty percent. MIT’s NANDA group found most enterprise pilots show no measurable change to the bottom line. The cause keeps tracing back to alignment, not architecture.
That is the part I work on. Before a team deploys anything, the words on the rollout deck carry different meanings for the executive signing the check, the manager held to the number, and the worker afraid of being measured out of a job. “Productivity.” “Augmentation.” “Efficiency.” Even “AI enablement” itself. Pin those meanings in the room first. Get the stakeholders aimed at the same outcome. Then buy the tool. Skip that step, and you join the eighty percent.
Robbins and the Pope are working the same seam from above. Once the routine layer is cheap, value migrates to whatever stays human: depth, judgment, dignity at work. An organization that cannot tell those layers apart will automate the wrong one and call it progress, and wonder why there is no progress.
This week on What Time Binds
I spent the week on the word sitting at the center of all of it: dignity.
The Word Is Dignity read Pope Leo XIV’s encyclical and the apology no pope had made before it. The companion, What Do You Mean? — Dignity, walked the floors the encyclical left standing: contribution, consent, and the procedural floor the Vatican stage never reached. It did that alongside three writers working the same problem.
Both circle the question this Binding keeps returning to. What does your organization mean when it says AI will “help” people? Pin it before the tool answers for you.
Subscribe to What Time Binds for the monthly Pinned Terms strip and next week’s Monday essay. If one of these five reads is new to you, restack it and send the writer a reader.
The Friday Binding is a weekly ritual inside What Time Binds. Five reads worth passing forward, each framed through the meaning repair lens. Every Friday morning, free.
— Jerry Washington, Ed.D. • what-time-binds.com