AI for Research Efficiency, episode 0: TLDR: one request, a team of agents
Narration: an AI-generated voice (ElevenLabs).

[0:00] Start here
Best practices. Instruction files, version control, careful habits: a whole series of them. Very responsible. Too long; didn't read. Let's just vibe code.
One request, to a team of agents. Twelve and a half minutes later: eighty-six years of a river's records, analyzed, with figures for a paper.
AI for Research Efficiency. Episode zero: T L D R.

[0:33] The quick part
First, the quick part. Make a folder for your project, right-click inside it, and choose Open in Terminal.
Paste one install line: for Claude Code, Codex or Antigravity. Each has its own episode. Then open the terminal again, the same way.
Type the agent's name, sign in, and tell it you trust the folder.

[1:00] One request
Now the real part. Say who you are, and what you want at the end. And ask it to spin up a team of agents.

[1:12] The team at work
It works in auto mode: a second model checks each step before it runs, so this run never stopped to ask.
First, the lead agent set the project up: a Python environment, eighty-six years of gauge records, an instruction file it reads every time it starts here, and a first saved version.
So those best practices we skipped? It did them anyway. Show-off.
Then it split the work. Three agents at once: one on floods, one on low flows, and one on the river's whole pattern, and the rain.
As each one reported back, it pulled the results together, checked the numbers it would quote, and drew the figures.

[2:00] The result
The finding: the river runs about sixteen percent higher than before nineteen eighty-three. By the team's analysis, more rain explains nearly all of it, and its floods and low flows haven't changed detectably.
Every stage is saved, so you can go back to any of them.

[2:19] Almost anything
The same approach works on almost anything.
In this series, a team of agents reviewed fifty-seven files for a literature review. Others mapped a research field, turned a paper into a narrated video, and built a web page.
Some call it vibe coding. Say who you are, what you want at the end, and: spin up a team of agents.

[2:43] Before you start
Your run will look different: models update, and every computer is set up its own way. If something goes wrong, ask your agent to help fix it.
Before you point it at sensitive data, watch episode nine: your data, your keys, and your campus's rules.
And if you ever do want those best practices, the rest of the series is right there. No pressure.
Next: episode one, why the command line.