We build
products with AI.
We teach you how.

A small team, seven products shipped.
Have us build it for you, or follow the courses
and build it yourself.

What each one does

Products · 07
dailyup.xyz 01 / 07

What it takes to work with AI

Why · 02

Tools change. Saying what you need, checking the result and learning new methods are the skills we keep practising.

  1. 01

    Make it real, and make it hold

    AI can write the code, call the APIs and set up the project. You still have to state what you need, check whether the result matches it, and work out why it broke when it does.

  2. 02

    Learn the next thing

    Models and frameworks change, and projects run into problems you haven’t met before. Breaking a problem down, finding sources and checking the method is what keeps you moving.

Live · 02 courses

Two live courses

One practises learning with AI, the other builds a product with Codex. Five live sessions each, with assignments, Q&A and replays.

01 Open 5 live sessions · 2h each

The AI Era: From Asking to Learning

Learn AI, and learn through it

We start with what AI can and can’t do, then take a topic you want to learn and practise finding primary sources, understanding the material and putting it to work. Along the way the sources, notes and output become a knowledge base you can reuse.

  • Understand what AI changed: from asking questions to finishing tasks with Skills
  • Find and verify primary sources, and set up channels that keep good information coming
  • Break a learning goal down, use AI for understanding, practice and feedback, then check you learned it
  • Learn against a real problem and put what you learn into an actual piece of work
  • Build a second brain where sources and notes stay searchable and reusable
Course fee ¥1,980 / 5 sessions
02 Open 5 live sessions · 2h each

AI Coding: Shipping Ideas Reliably

Practise on a real project and take an idea to a working app

Codex is the main tool in the demos. On a real project you practise clarifying requirements, choosing an approach, building and acceptance — covering both the lightweight and the complex project path. The methods apply to Claude, DeepSeek Harness, Workbuddy and other agent tools too.

  • Clarify requirements: turn “build me a website” into a spec you can check item by item
  • Choose a stack and a visual direction that fit the requirement
  • Run tests, debug, and check completion item by item
  • Practise both workflows: the lightweight project and the complex one
  • Save the spec, build and acceptance steps as Skills for the next project
Course fee ¥1,980 / 5 sessions

I built DailyUP Studio.

Seven products, all built with AI coding, all still in use.
The courses and the notes here come from building them.
If any of it helps, or if you’re building the same things, message me on WeChat.

Scan to add me on WeChat Click to flatten into a QR code