The chart is clear: after the dramatic collapse of storage costs, processor costs and bandwidth costs, it is now the turn of software itself to plummet. In other words, coding will soon cost as little as sending a text message in the 2000s.
We are entering an era where building an app is becoming as simple as posting a TikTok video. Tools like Replit, v0 or Figma Make turn any wild idea into a viable product for 200 a month. Previously, you needed a team of engineers and months of struggle; now, a good prompt and a cup of coffee are enough.
And here comes a total reversal: content becomes application, and applications become content. Tomorrow's YouTuber won't make videos anymore -- they'll publish apps. Traditional developers, meanwhile, risk becoming the TV stars who never understood YouTube.
In short, the cost of software is collapsing, but creativity is exploding. We are moving from a world where code was a wall to climb to a world where it becomes a toy.
The question is: how far will this new era of "software accessible to everyone" push the boundaries between creativity and development?
Chart source: SK Ventures
Welcome to the new game. This isn't just a trend -- it's a whole new way of building a product with AI.
Traditionally, building a product means:
Waterfall: you plan everything, then build step by step. Predictable... but rigid as soon as requirements change.
Agile: you move in iterations. You deliver value continuously, learning along the way.
With AI, it's different. You start with an idea, explore what the model can do, then shape a product from that exploration.
The third line in the image illustrates it well: you start with a car, you get a cat, then a duck. The first version is often strange. But it's the one that reveals the model's true potential.
Two strategies guide this exploration
RAG (Retrieval-Augmented Generation) is like a brain that reads a file before answering you. The AI queries your databases, documents or emails to stay up to date. Ideal for fast-moving environments (sales, support, competitive intelligence...).
CAG (Cache-Augmented Generation) is like a brain that has already memorized the right reflexes. You teach it once, it responds instantly. Perfect for stable contexts (internal processes, FAQs, business procedures...).
The result: Fewer errors. Faster decisions. An AI that truly understands your business.
RAG and CAG are the bridge between collective intelligence and artificial intelligence. And that is what true augmented decision-making looks like.
Thanks to Steve Nouri who wrote a great article on the subject: RAG vs. CAG.