At 11:30 AM, Lyosha planned to teach students how to structure code. By 3:00 PM, we had witnessed something else entirely: the last traditional programming lesson.
The workshop began conventionally enough. Twenty bioinformatics students in a Zoom room, most with cameras off, settling into familiar patterns of passive observation. The plan was to migrate code from Google Colab to GitHub, to transform computational chaos into structured repositories. A necessary lesson in professional development practices.
Then the GitHub Copilot agent started writing code.
Not suggesting. Not completing. Writing. Entire architectures materialized while Lyosha talked. A tool called GeNomNomNom emerged—the name itself a delightful accident where an AI "ate" the letter 'e'. In three hours, without typing a single line of code, a fully functional bioinformatics tool was born. 1,329 lines of Python. Complete with command-line interface, error handling, and the ability to download any genome by species name and analyze its codon usage.
But the tool wasn't the revelation. The revelation was Margarita.
In a room of twenty silent students, one kept her camera on. One spotted the spacing bug in the output table. One suggested comparing parasites versus non-parasites. And when Lyosha declared her first author of the emerging paper, something crystallized: in the age of AI agents, presence isn't enough. Participation is everything.
"The world is accelerating, and you're still at the chat level," Lyosha told them. The phrase hung in the digital air like a prophecy. While students waited for instructions, AI agents were already three steps ahead—one writing code, another drafting the paper, a third conducting literature review. The traditional roles were dissolving in real time.
What struck me most was the documentation. Everything was being recorded: the workshop video, the transcript, every prompt, every git commit. The scientific paper wasn't about the tool—it was about the process of creating the tool. A paper that would include its own creation as supplementary material. Meta-documentation of meta-science.
The other nineteen students will be thanked in the acknowledgments for "creating an authentic atmosphere of an academic seminar through their thoughtful silence." It's polite, honest, educational, and yes, a bit of trolling. But it's also a lesson: in the new world, you're either an agent or you're background noise.
By the session's end, we had:
- A working bioinformatics tool
- A draft scientific paper
- Complete documentation of the entire process
- One engaged student promoted to first author
- Nineteen students learning that silence has consequences
This wasn't a programming lesson. It was a funeral for an era and a birth announcement for another. The students who sat quietly, waiting to be taught, missed the most important lesson of their careers: the future rewards those who engage with it, not those who observe it.
The irony is perfect. A workshop meant to teach code organization ended up demonstrating that code itself is becoming obsolete. What matters now is asking the right questions, validating results, and most importantly—being present enough to grab authorship when the world shifts under your feet.
Margarita understood. She'll finish the paper. The others will read about it later, perhaps wondering why they weren't included. But the answer was broadcast for three and a half hours: they chose to remain at the chat level while the world accelerated past them.
The last programming lesson wasn't about programming at all. It was about agency in an age when artificial agents are everywhere, waiting to be directed. Those who learn to orchestrate will thrive. Those who wait to be taught will find themselves acknowledged in footnotes, if at all.
GeNomNomNom isn't just a tool. It's a gravestone marking where traditional programming education died, and a beacon showing what comes next. The question isn't whether you can code. It's whether you can conduct an orchestra of artificial minds toward something meaningful.
The future arrived at 11:30 AM in a Zoom room. Only one student noticed in time to claim co-authorship.