myofficehours.ai
This is issue 1, from 11 August. Read the latest issue →

Issue 1 · 11 August 2026

Office Hours — Issue 1

11 August 2026 · A weekly read on AI for faculty

The price of admission just dropped to zero

For three years the interesting question about AI was how capable it would get. Over the past two weeks the interesting question quietly became who has it, and the answer is turning into everyone.

OpenAI removed the message caps on free ChatGPT accounts, launched a free tier for verified academic researchers, and signed another campus-wide agreement. Anthropic cut the price of its flagship model roughly in half and made its assistant free for teachers. Google switched on its classroom assistant for students of every age. None of these were capability announcements. They were distribution announcements, and they land right before the fall term.

What this means for your syllabus

The practical consequence is that the informal calculation many of us have been making, that some students have the good version of these tools and most do not, has stopped being true. Students arriving in your classroom in three weeks will have a capable assistant with no usage limit, on their phone, for free. Any course policy built on the quiet assumption that AI is an occasional and slightly inconvenient option now needs to assume it is always available and costs nothing. That is not an argument for banning it or for surrendering to it. It is an argument for saying, in writing, what you actually want students to do with it in your course, because the ambiguity that was tolerable last year will not survive this one.

The counter-current worth noticing

It would be easy to read the past two weeks as pure acceleration, so it is worth flagging the opposite signal. Google shipped three new models that emphasize efficiency and cost rather than raw capability, and conspicuously did not update its flagship, which has now gone six months without a refresh amid reports the team has struggled to hit internal targets. Read alongside the price cuts, this suggests the frontier may be flattening while the competition moves to distribution and cost. If so, that is good news for anyone tired of rebuilding a course around whichever tool was best that month. The tools you learn this fall are likely to still be roughly the tools you have next spring.

Where AI is starting to check us

One development deserves attention across every discipline. Researchers have begun pointing AI agents at the published literature to look for errors, and they are finding them, including a case where a chemist discovered that a reference database seventy-five years old was wrong after a model flagged an inconsistency. A separate team tried to reproduce the core claims of one hundred sixty-eight papers from a major machine-learning conference and succeeded, partially, in thirty-four of them. Both results point the same direction. An automated layer is forming over the scholarly record, one that will catch real errors and will also generate confident false alarms, and the norms for how we respond to it do not exist yet.

One thing to actually try this week

If you do one thing, make it a course notebook. Upload your syllabus, your readings, and your slides into a single grounded workspace such as NotebookLM or a Claude Project, and then ask questions of that material rather than of the open internet. It takes about fifteen minutes to set up, it keeps the model anchored to sources you chose, and it is the single change most faculty report as the point where these tools stopped feeling like a novelty and started saving real time. The dashboard has step-by-step guides for both under Custom workflows.

What someone who studies this thinks

Mollick has been making this argument for a year and this week is the evidence for it: capability that used to cost twenty dollars a month and require clever prompting is now free to more or less everyone. His point is that our institutions were designed around intelligence being scarce and expensive, and almost none of that design has been revisited. The assessment you wrote for a world where good AI was a paid perk is now being handed to a room where every single student has it.

— Ethan Mollick, Wharton School, University of Pennsylvania · read the piece

Where the experts actually disagree

Should faculty teach with these tools, or teach around them?

Ethan Mollick (enthusiast) — Use them deliberately, deciding task by task whether the model is a visible collaborator you learn from or a wizard you simply delegate to. The skill worth teaching is that judgement. Their argument

John Warner (skeptic) — Writing is the worst possible use case, because the struggle with language is the thinking, and outsourcing the draft removes the very thing the assignment exists to develop. Their argument

Robert Talbert (pragmatist) — Neither, yet. He moved assessment in-class and on paper as deliberate emergency triage, and says plainly that it is not his ideal design but it is honest about what he can currently verify. Their argument

Nobody credible is arguing this is settled. If your department is being told there is a consensus about teaching with AI, that is a claim about your department, not about the field.

The rest of the week, briefly

One line each, ordered by how much it should change what you do. The full account of any of them is on the dashboard.

Act on this

Watch

6 more stories ran this week and are waiting on the dashboard.

Three worth your time

121 resources went into the library this week. These three are the ones to open first.

[Napkin AI: Turn Text into Diagrams and Visuals](https://www.napkin.ai/) — Images and figures · 15 min · Napkin AI Paste in existing text, such as a paragraph from a lecture or a paper, and Napkin auto-generates an editable diagram, flowchart, or infographic that exports as PNG, SVG, or PowerPoint. You will be able to turn a wall of bullet points into a clean slide visual without prompting or design skill.

[Ten Simple Rules for Using AI in Grant Writing](https://medicine.stanford.edu/news/stories/2025/07/10-rules-for-ai-in-grant-writing.html) — grants · 10 min · Stanford Medicine Ten concrete rules for using AI responsibly in grant proposals, including checking funder-specific AI policies, never pasting unpublished data into public chatbots, and verifying every AI-suggested citation before submission. You will be able to use AI as an editing and brainstorming aid without risking intellectual property leaks or fabricated references.

[AI-Resilient Assignments](https://ctl.wustl.edu/resources/ai-resistant-assignments/) — Teaching and course design · 15 min · WashU Center for Teaching and Learning Six concrete strategies for making assignments harder to complete with AI alone, such as requiring authentic real-world tasks, oral exams, or references to specific in-class discussions the model cannot access. You will be able to pick at least one strategy and apply it to an existing assignment this week.

Who we read this week

This issue drew on OpenAI, Anthropic, Robert Talbert, MIT Sloan Teaching & Learning Technologies, Ethan Mollick, John Warner, Derek Newton, European Commission AI Office. The full watchlist, with what each source is good for and where they stand on AI in education, is on the dashboard under "By voice".

The whole library lives on the dashboard, sorted by what you are trying to get done and by the tools you already have. Each task runs from a twenty-minute start to something you could spend a weekend on.

Office Hours is assembled automatically: a daily sweep for new tutorials and a weekly edition on Monday mornings. Every link is checked before it ships. Reply with anything broken, missing, or worth adding.

← Every issue