How to learn to code in the age of AI
Video by Tina Huang
Key takeaways
- You still need coding basics to read and judge what AI writes
- Architecture explains how the parts of an app fit together
- Git and GitHub keep your work safe and let you undo mistakes
- Security and privacy matter from your very first project
Tina Huang, a former data scientist with a huge following among people learning tech skills, answers the question every beginner asks now: if AI can write code, what should I actually learn? Her answer in this 16-minute video is practical and calm. You still need foundations, but which ones matter has changed.
What it covers
She starts by showing what coding looks like today, with AI assistants doing much of the typing. Then she lays out five lessons she thinks every new learner should focus on: the basics of coding (so you can read and judge what the AI writes), software architecture (how the pieces of an app fit together), version control with Git and GitHub, security and privacy, and containerization with microservices. She finishes with tips to speed up your learning, including a study-plan prompt you can use with an AI assistant.
Who it's for
People who want to learn real coding skills, not just use app builders, and who feel unsure where to start now that AI tools exist. It's also useful for career changers deciding whether learning to code is still worth it.
Good to know
The video is sponsored by Warp, an AI-powered terminal, and links to Tina's own courses. The learning roadmap stands on its own without them.
Tip: treat the five lessons as a checklist. You don't need to master them all at once, but knowing they exist helps you understand why the AI's code sometimes breaks, and what to ask for when it does. Revisit the list every few months as your projects grow.
Tools mentioned
GitHub is a developer platform for hosting code repositories, reviewing and collaborating on code, and automating software workflows. It serves individual developers, open source communities, and businesses of all sizes, from startups to large enterprises. The platform includes GitHub Copilot, an AI coding agent that helps with tasks from an open issue through reviewing and merging a pull request. GitHub also offers CI/CD automation through Actions, cloud development environments via Codespaces, and tools for code quality, vulnerability detection, and secret scanning. With plans ranging from a free tier for individuals to enterprise offerings with advanced security and governance, GitHub supports public and private repository hosting, pull request workflows, and integration with a wide range of developer tools.