Doing the Work in Public

Doing the Work in Public

Professor Bear Does His Assignments, video 1 of the series: an introduction.

A professor does the same assignments he gives his students, in public, with the same process log they keep. The point is not the answer. It is how the work actually gets done. This film covers the idea, who does what when an AI does the AI work and a person conducts it, the project the three courses share, and how the rest of the series will work.

Made for INFO 7375: Computational Skepticism for AI, Branding and AI, and Prompt Engineering for Generative AI. Narrated by Liam, in for Bear.

Chapters 0:00 Opening: the series and this film 0:46 Part 1: The Idea 1:15 Part 2: Who Does What 2:11 Part 3: The Project 3:12 Part 4: How To Follow 3:38 Recap 3:49 Your Turn

Your turn Pick one thing you made with AI help. List what the AI did, what you decided, and one thing you doubted and checked. Then check each line against a file or a source.

Tools shown as themselves were captured from their public pages on 2026-10-07. Nothing here is a product endorsement.

@NikBearBrown

▶ The full series playlist — Brutalist https://www.youtube.com/playlist?list=PLG9H-C6rp5RU

What this video covers

• 0:00 — Opening: the series and this film
• 0:46 — Part 1: The Idea
• 1:15 — Part 2: Who Does What
• 2:11 — Part 3: The Project
• 3:12 — Part 4: How To Follow
• 3:38 — Recap
• 3:49 — Your Turn

Transcript

A lightly formatted transcript of the video, with timestamps so you can jump to any section on YouTube.

Opening: the series and this film

[0:00] Professor Bear does his assignments, doing the work in public. Learn why a professor does the assignments he gives his students in public and how to follow along. Hello, this is Liam in for Bear. The journal covers three INFO 73-75 courses, computational skepticism, prompt engineering, and branding in AI. In all three, the assignments are done in public with the same process logs students keep. The point is not the answer. It is how the work actually gets done. Five terms, a frictional log is a dated record of what was tried, what went wrong, and who did what. Conducting AI means the person directs and supervises the AI's work. Irreducibly human is what stays with the person, the decisions and the understanding. Skepticism is not believing something because it looks polished. An AI slop is output that looks finished but was never checked.

Part 1: The Idea

[0:46] Part one, the idea. A student hands in an assignment. The teacher sees the finished thing. What the teacher cannot see is how it was made or whether the student learned anything. So, Professor Bear does the same assignments in the same courses and keeps the same log. Every try, every wrong turn, every fix goes in dated. The log is the evidence. Here is the real log, public on GitHub. His stated purpose is to let a teacher judge the learning and not only the thing that was handed in.

Part 2: Who Does What

[1:15] Part two, who does what? AI is good at producing. It collects, drafts, sorts, and matches fast. And whatever it produces looks finished whether or not it is right. That is why the human work matters. Conducting AI means the person directs the work, chooses the goal, hears what is off, and decides what goes out. What stays with the person is decided task by task. The AI can draft, compute, retrieve, test, and challenge. The person owns the understanding, the justification, and the decision. Skepticism holds it together. Something is not true because it is fluent, polished, or confident. It is true when something has been checked. Every film in this series shows the split with three lines from the real log, what the A I did, what the human decided, and what doubt caught. Two real ones, a tool was specified for the wrong application and Professor Bear stopped it. And an audit found a filter was keeping postings it should not have. Both are in the log.

Part 3: The Project

[2:12] Part three, the project. The project is one tool built live in three courses. It reads company job boards for roles that teach, advocate, or make learning materials. Each course asks it a different question. Each course stresses something different. Skepticism asks, is the output right? Prompt engineering asks how we work with AI, which is conducting it. And branding asks how we present something so people can understand it. Alongside it, Professor Bear tests the tools students and teachers actually use. Figma is a design tool and its own page calls it free tools for the classroom. Google's research tool, formerly called Notebook LM, is now Gemini Notebook. Its page describes a research partner grounded in information you trust. The test is the same for every tool. Use it on a real education problem honestly. Film what happened, including what failed, then the viewer decides. Tools that fail the trial get a film, too. If someone thinks one was dropped wrongly, they can say so in the comments and he can revisit it.

Part 4: How To Follow

[3:12] Part four, how to follow. After this introduction, each film is a dated diary entry. It says what exists that day, what ran, what did not, and what changed. Failures go in the entry where they happened. Nothing is rewritten afterward. It will not look coherent at first. Professor Bear says he gets better at making films by making films using the same AI as last week, so the improvement is not the tool. It is how the person works with it. Let's recap

Recap

[3:39] With Claude. A professor does the work he assigns and keeps the log. The AI produces, the person conducts, polish is not proof, check it. The series is a diary and the person improves with

Your Turn

[3:49] Practice. Pick one thing you made with AI. Help list what the AI did, what you decided, and one thing you doubted and checked. Then check each line against a file or a source. Doing the work in public at Nick Bear Brown.