The New Computing

The Moment

The Moment

In 1960 J. C. R. Licklider observed that “instructions directed to computers specify courses; instructions directed to human beings specify goals,” and imagined how machines might someday work from goals.⁽¹⁾J. C. R. Licklider, “Man-Computer Symbiosis,” IRE Transactions on Human Factors in Electronics HFE-1 (March 1960): 4–11, §5.4, “The Language Problem.” In 1972 John Kemeny wrote that “we at Dartmouth envisaged the possibility of millions of people writing their own computer programs.”⁽²⁾John G. Kemeny, Man and the Computer (New York: Scribner, 1972), 30. Kemeny co-created BASIC and the Dartmouth Time-Sharing System to make it so. Grace Hopper spent her career pushing programming languages toward English while being told it was impossible.⁽³⁾Of her 1952 compiler she recalled: “Nobody believed that. I had a running compiler and nobody would touch it. They told me computers could only do arithmetic.” Philip Schieber, “The Wit and Wisdom of Grace Hopper,” OCLC Newsletter 167 (March/April 1987). For most of a century the dream remained out of reach. Accessing computing required speaking the computer’s language, which remains difficult and dehumanizing. Many people couldn’t or wouldn’t.

Computing educators have been waiting for this moment. We have a lot of work ahead: assignments, courses, degree programs, and pedagogies. This is the work that finally realizes the dream of computing for everyone. Generative AI is the enabling technology. But we’re technology educators. We know what to do next. I’m working on this at every opportunity—through assignments, courses, a new degree, and assessment; through writing and talks; and by building community—independently and with colleagues. This is the most meaningful work I’ve ever done.

Assignments

Assignments

CS 124 is introductory computer science at Illinois: one of the largest CS1 courses anywhere, 80% non-majors, more than 16,000 students since 2017, and free online to anyone who wants to learn. We’re one of the first large CS1 courses to embrace AI, and we’re both teaching students to use it and integrating it into our existing courseware. In Fall 2025 we kept the old project and permitted AI, and it failed: the information students needed for the project to be graded fairly was enough for an agent to do it for them.

So starting in Spring 2026, every student designs and builds their own app with a coding agent that writes and runs the code while you direct it in plain English. Directing one well is a skill, and it’s what I mean by conversational programming. Nearly 400 students built an app that first semester. Between them and their agents, more than three million messages went back and forth. They built some amazing things: a simplified phone launcher for seniors, with medication reminders; a tracker for multi-day materials-science experiments; and an app that tags quotes by emotion and floats them on a 3D sphere. Because no two projects are alike, we grade effort rather than output, using AI to read each student’s transcripts and estimate the time they spent working with their agent.

Classical programming hasn’t gone anywhere: students still write code by hand every week on proctored, AI-free quizzes in our Computer-Based Testing Facility. Both skills, taught side by side. We’re also building AI into the courseware itself: an MCP server that publishes the course to whatever agent a student is already running, personalized quiz review built into our online materials, and search and chat across everything we publish. Everything we’re doing is written up for educators.

We’re still working through the inherent tension as a new skill replaces an old one. And universities are still determining how to provide equitable access to AI tools. But we can’t leave students unprepared in the meantime. CS 124 students pay about $20 a month for their own coding agent, the first cost we’ve ever attached to the course, and less than a textbook costs. It’s a modest price to pay for a generational technology. Inference costs are falling, and soon agents will run on student machines.

Courses

Courses

Conversational Programming launched in Fall 2026 for anyone who wants to build—no prerequisites, no prior experience, no code. Students create technology entirely by directing a coding agent, learning to work like architects: exploring problems, sketching solutions, critiquing what emerges, and validating with real users. The course uses design pedagogy and continuous interaction between human participants. It’s a studio, not a lecture course. Nobody knows how to teach this skill yet, including me. The course’s content and pedagogy are both experimental: we’re teaching a new skill in new ways, exploring the possibilities of personalized pedagogical agents. We’re sharing everything as it happens so computing educators can follow along.

Using and Understanding AI is for students who will use AI throughout their lives but never build it: the philosophy major, the pre-med, the future policymaker. No prerequisites, and definitely no code. Students create websites, music, and data analyses with AI while exploring how it works and what it means for society. For their final projects in Spring 2026 they made an aquascaping tracker, a platform for running a sorority, and a meal planner—see for yourself. It’s a course about AI, built with AI, and taught using AI: I built the entire platform through conversation with Claude Code without writing a line of code, and every one of the 200-plus build sessions is published. Everything is open to educators, including a course kit you can fork.

A New Degree

A New Degree

Beyond courses, I’m leading the design of Applied Computing, a proposed four-year undergraduate degree at Illinois for students who want to use AI to build for the domains they care about. Its model is architecture school: studio-based design from day one; a required concentration outside computing; breadth in studio art, moral reasoning, literature, and advanced writing; and a capstone thesis with a public defense. Graduates will be fluent enough in technology to build, deep enough in a domain to know what’s worth making, and ready to take responsibility for their creations. It’s a computing degree with room, finally, for all the parts of a creator’s education that technical training used to crowd out. If our students can dream it, they can build it. We have to help them dream well.

Assessment

Assessment

AI is breaking how we assess students and showing us new ways to measure learning. I’m building the new tools and helping computing educators fix the old assignments.

Conversational assessment is an oral exam that scales. A student explains a concept to an AI interviewer in chat while a second, hidden AI grader holds the rubric and grades the transcript. The interviewer never receives the answer key, so it has nothing to leak—619 adversarial checks, zero leaks. Only ideas the student offers unprompted earn credit. On a blind re-grade of ten transcripts, two AI graders agreed with each other about as often as two human graders did; a full reliability study comes next. It runs proctored in our testing facility, costs about a dime per conversation, and is free for educators during the beta. Try one yourself, no login required. We’ve completed an integration with PrairieLearn and have more planned.

I’m co-leading AI Can Do Your Homework. Now What?, a SIGCSE Virtual 2026 working group: one of us from each of nine institutions in four countries. We’re building a taxonomy of how computing educators are adapting their assessments—or deliberately preserving them—now that agents can complete most of them. Our July 2026 workshop drew more than sixty instructors; the workshop report is public. A community survey follows in October 2026.

Writing

Writing

I write about this frequently. Move 37 Coding asks why an alien intelligence should reproduce human coding patterns, and Teaching and Doing declares that there has never been a better time to be an educator who builds. Epistemic Rents points out that coders and teachers both profit from scarcity, and I’m rooting for its end. Healing Agents admits that maybe I never actually loved to code. With Or Without AI? offers advice to students and makes a demand of their educators, while Another Skill argues that agentic development can be learned, so it can be taught. Tech v. Tech imagines educational technology without most of the educational technology companies. Welcoming Thoughts speaks to new students at a moment when nobody knows what computing will look like when they graduate.

Four studies are underway: whether grading process rather than product works, whether conversational assessment measures what we think it does, what it costs to build a course by conversation, and where student help-seeking went when agents arrived. The Atlantic quoted me on conversational programming in May 2026.

Talks

Talks

At FISEE 2026, an invitation-only workshop on AI in software engineering education taking place September 27–30, 2026, I’m presenting Nobody Knows How to Teach Conversational Programming: Today’s Most Powerful Computing Skill. The position: figuring out how to teach conversational programming is the most important open question in computing education, and answering it starts with defining what it means to be good at it. I’ve made some version of this argument to faculty at Illinois, the University of Sydney, Georgia Tech, and Constructor University, Bremen; to educators gathered at GitHub HQ; to the Illinois CS Summer Teaching Workshop, the GenAI in CS Education Consortium with Kevin Lin and Cory Gwin, and Replit’s Faculty Fellows; and, first of all, to my own students. Slides for all of them are on the talks page.

Community

Community

Education is a team sport, and we need a movement to respond to this moment. In March 2026 Kristin Stephens-Martinez and I founded Computing Educators, a members-only forum of nearly 300 educators from K–12 through research universities, where much of my thinking gets argued out with colleagues who push back. I co-founded the Illinois CS Summer Teaching Workshop in 2022; was named a Faculty AI Fellow in Illinois’s Summer AI Studio, mentoring colleagues across campus as they bring AI into their courses; and advise the Illinois Forum on Human Flourishing in a Digital Age.

None of this happens without institutional support, and I’m grateful for mine. Illinois has backed this work with AI resources, a testing facility, and the freedom to try things, and has recognized it with the Rose Award for Teaching Excellence and the Scott H. Fisher Computer Science Teaching Award. In 2026 Replit named me a Faculty Fellow. Every university needs to enable its own educators the same way.

Practicing What I Profess

Practicing What I Profess

I don’t just teach this. I live it. Since December 2025, every Claude Code session on my machine has been captured by a recorder I built for the purpose: 1,066 sessions across 74 projects, about 1,050 hours of active conversation on 176 different days. Every site I’ve launched since was built the same way. Numbers as of September 12, 2026.

Get Involved

Get Involved

If you’re a student wondering whether there’s a place for you in this: there is, and you don’t need to know how to code to start. At Illinois, Using and Understanding AI and Conversational Programming have no prerequisites. Anywhere else, the CS 124 materials are free online, and you don’t need a coding agent to begin. Start there, then write to me about what you built.

If you’re an educator wondering whether you could do this: now you can. The barrier is no longer the ability to write code. It’s knowing what you want your students to learn. If you teach computing, join Computing Educators, where these conversations happen every day. Follow Conversational Programming as it unfolds, ask for the Using and Understanding AI course kit, or try a conversational assessment yourself. Or contact me directly. I’d love to hear what you’re trying.

Words for the Moment

Words for the Moment

A friend read this to me. Eliot is writing about his own twenty years and about mastering a craft just as its purpose changes. That’s where I am.

So here I am, in the middle way, having had twenty years—
Twenty years largely wasted, the years of l’entre deux guerres—
Trying to learn to use words, and every attempt
Is a wholly new start, and a different kind of failure
Because one has only learnt to get the better of words
For the thing one no longer has to say, or the way in which
One is no longer disposed to say it. And so each venture
Is a new beginning, a raid on the inarticulate
With shabby equipment always deteriorating
In the general mess of imprecision of feeling,
Undisciplined squads of emotion. And what there is to conquer
By strength and submission, has already been discovered
Once or twice, or several times, by men whom one cannot hope
To emulate—but there is no competition—
There is only the fight to recover what has been lost
And found and lost again and again: and now, under conditions
That seem unpropitious. But perhaps neither gain nor loss.
For us, there is only the trying. The rest is not our business. East Coker, from Four Quartets, T. S. Eliot