Should Anyone Still Learn to Code If AI Will Generate 100% of Code Next Year?

Anthropic CEO Dario Amodei recently made a stunning prediction: within just 3-6 months, AI will be generating 90% of all code written, potentially expanding to all code within a year. 

Software is evolving from a scarce resource to an abundant one. What would your association build if coding limitations vanished overnight? Assuming Amodei knows his stuff (spoiler: we think he does), this may be a reality for your organization very soon. It's time to dream big. And then, when you're done doing that: dream bigger. Projects that seemed technically unfeasible last year may not only be possible soon—they might be easier than you think.

The Hyper-Acceleration of AI in Software Development

Traditional technology development has followed Moore's Law for decades. Named after Gordon Moore, co-founder of Intel, this principle observed that the number of components on a microchip doubled approximately every two years, leading to a predictable growth in computing power and decrease in cost. This steady progression allowed industries to plan for technological advancement on a relatively stable timeline.

AI development, however, has shattered this pattern. Instead of doubling every two years, AI capabilities are doubling approximately every six months. This means that in the time it would take traditional computing technology to improve by 2×, AI improves by roughly 16×. 

AI Learning Hub: A Real Example of AI-Accelerated Development

A perfect example of thinking bigger with AI-assisted code comes from the Sidecar AI Learning Hub. When faced with outdated educational videos, our team needed a way to extract video content from PowerPoint presentations, generate transcripts, and create text versions quickly.

The traditional approach would have required days of complex programming. Instead, we simply asked an AI (Claude 3.7) to generate the necessary Python code. In seconds—not days—the AI produced a script that:

• Unpacked PowerPoint slides
• Extracted embedded videos
• Generated transcripts
• Added the transcripts to slide notes
• Removed the original videos
• Saved the modified presentations

This task would have taken an experienced developer several days. With AI, it took minutes from idea to working solution.

And why stop at simple transcription? We're also building programs to generate:

• AI-generated natural-sounding voice narration
• AI avatars for visual presentations
• Automated content updates that keep pace with rapidly evolving AI capabilities
• Different flavored versions of content customized for different industries and associations (Think of an AI Learning Hub for CPAs or HR professionals)

What would have been financially impossible with traditional development is now completely feasible. This is the power of AI-generated code—not just faster development, but entirely new possibilities that previously would have been dismissed as too complex or expensive.

Interested in creating an AI Learning Hub for your members? More info here. 

Should People Still Learn to Code?

With AI poised to generate nearly all code in the near future, a question naturally arises: should people still bother learning to code?

Short answer: Yes, but for different reasons than before.

Understanding computer science fundamentals remains crucial because:

• Building block knowledge enables better AI collaboration - Understanding how systems work at their core helps identify what's possible (and what isn't) when directing AI to generate code
• Technical literacy reveals potential issues - Someone with coding knowledge can spot security vulnerabilities, performance problems, or logical errors in AI-generated code that non-technical users might miss
• Deeper problem understanding leads to better solutions - Coding knowledge helps break down complex problems into components that AI can effectively solve
• Maintenance and adaption require technical skill - When AI-generated code needs modification or something goes wrong, understanding the code becomes essential

The programming skills that matter are evolving from writing every line of code to orchestrating AI systems that generate solutions. It's less about memorizing syntax and more about understanding systems and identifying opportunities.

We've seen this pattern before. Programming evolved from punch cards to assembly language to high-level programming languages. Now we're entering an era where developers describe what they want to build, and AI handles the implementation details.

Jevons Paradox: Abundance Creates More Demand

When resources become more abundant and less expensive, an interesting thing happens: we use more of them, not less. This economic principle, known as Jevons Paradox, perfectly explains what's about to happen with AI-powered software development.

As the cost of software development plummets, demand will skyrocket. That custom integration you dismissed as too expensive last year? That member portal redesign that seemed too complex? That data analytics project that required specialized skills? They're all back on the table now.

For associations, this means you'll likely implement more software solutions as they become accessible—not fewer. The question shifts from "Can we afford to build this?" to "Which of these suddenly affordable projects will deliver the most value to our members?"

Preparing Your Association for the Software Abundance Era

To capitalize on this shift, associations should focus on:

1. Building with flexibility in mind - Create abstraction layers that allow you to swap out underlying technologies as capabilities evolve. The AI tool that's cutting-edge today might be obsolete tomorrow.
2. Focusing on capabilities, not technologies - Instead of getting caught up in the technical details, concentrate on what you want to achieve. The exact tools and methods will change rapidly.
3. Experimenting boldly - Try approaches that would have been cost-prohibitive in the past. The risk/reward calculation has fundamentally changed when development costs a fraction of what it once did.
4. Investing in AI literacy - Ensure your team understands enough about AI capabilities to identify opportunities. You don't need to become AI experts, but you should know what's possible.

What's next?

The transformation from software scarcity to abundance is already here. While we'll still need development knowledge to navigate this landscape, the role is evolving from writing code to directing AI to build solutions we can barely imagine today.

This isn't about eliminating software development—it's about multiplying what each developer (or technically savvy non-developer) can accomplish. The constraints that once limited your technical ambitions are rapidly dissolving.

The organizations that thrive will be those brave enough to dream at a scale previously unimaginable. What would you build if technical limitations vanished overnight? Your most ambitious ideas from last year's strategic planning session deserve another look—this time, through the lens of what's newly possible.