A New Chapter for fast.ai: How To Solve It With Code
Eight years ago, Rachel Thomas and I launched fast.ai with a mission to democratize artificial intelligence. We believed AI would become one of the most significant technologies in history, and, if widely distributed, felt like it could empower people all around the world to create anything they could imagine. But if only a small elite understood it, we worried it could lead to inequality.
Today, we’re excited to announce that fast.ai is joining Answer.AI, marking a new phase in making AI accessible to everyone. And, we’re announcing a new beta course, How To Solve It With Code—a new kind of “AI-first” educational experience. If you join this course, you’ll be part of the first cohort to experiment with this new approach.
The AI Landscape Has Changed
When we started fast.ai, making AI accessible meant teaching people to train their own models. That worked—fast.ai’s courses have helped create many of today’s top AI practitioners. But the landscape has changed dramatically. Just as the iPhone made the internet accessible without understanding TCP/IP, large language models (LLMs) like ChatGPT, Claude, and Gemini have made AI accessible without understanding model training.
Yet something interesting happens when people start using AI for coding. At first, it feels magical—like Keanu Reeves in The Matrix suddenly knowing kung fu. You can create an app in 15 minutes that would have taken weeks before. But then reality hits: you need to make changes, add features, fix bugs. The magic starts to fade, and many hit a wall they can’t get past.
A New Type of Course
At fast.ai we have not released a new course in over two years. That’s because we wanted to build something far more bold and different to any course created before—something designed from the ground up to use modern AI techniques deeply integrated into the whole learning process.
This is where our new course, “How to Solve It with Code,” comes in. It’s not just another course—it’s a completely new way of thinking about how humans and AI can work together. We call it Dialog Engineering.
Here’s how it works: Instead of asking AI to generate hundreds of lines of code at once, you work together in small steps. You might write a line or two, then have the AI suggest the next piece. This creates a powerful feedback loop where each step makes both you and the AI smarter. The code you create is cleaner, more maintainable, and more powerful than what either human or AI could create alone.
Introducing solveit
To support this approach, we’ve built a new platform called solveit. It’s the first tool specifically designed for Dialog Engineering, embedding AI deeply into the development process. We’ve been using it internally at Answer.AI for months, achieving extraordinary results with a small team (as you can see from our many breakthroughs this year.
This isn’t just theory—we’re using these exact techniques to build Answer.AI itself. When people ask how we accomplish so much with such a small team, this is the answer. It’s not about being genius programmers; it’s about having the right process for combining human and AI capabilities.
Beyond Just Coding
While many think of AI coding tools as just for building apps, Dialog Engineering can solve all kinds of problems. Whether you’re analyzing thousands of PDFs for data journalism, automating weekly reports, or building a complex web application, the approach is the same: break down the problem, solve it step by step with AI as your partner, and build something you truly understand.
Join the First Cohort
The course starts November 26th, led by a world-class team including myself, Johno Whitaker, and Audrey Roy Greenfeld. You’ll learn:
- How to solve real problems using code and AI together
- The principles of Dialog Engineering
- How to build solutions that you can maintain and extend
- Techniques used by top developers to work interactively with code
This first cohort will form a special community, just as our first fast.ai cohort did years ago. Together, we’ll explore and expand the possibilities of human-AI collaboration.
Is This For You?
This course is ideal if you’ve used AI for coding but want to go beyond simple prompts. You don’t need to be an experienced programmer, but you should have basic familiarity with coding concepts like loops and variables.
However, this isn’t for you if you believe AI can already replace programmers entirely. Our goal isn’t to have AI write all our code—it’s to create a powerful partnership between human understanding and AI capabilities.
We’re at the beginning of a new era in programming, where the question isn’t whether to use AI, but how to use it effectively. Join us in shaping this future, where humans and AI work together to solve problems in ways neither could achieve alone.
…oh and by the way…
We wrote this blog post using dialog engineering using solveit! First, my friend Hamel interviewed me to generate the material for this post, and then we used that transcript as the context to start a dialog where solveit, Hamel, and I iterated to write the post together. Here’s the video we used: