# Thoughts on old software and building new

- Date: 2026.05.28
- Locale: en
- Page: https://cv.rocorgi.wang/writings/thoughts-old-software-building-new
- Author: Long Wang (汪龙, rocorgi)

AI-native interfaces shift competition from GUI-era tool mastery to collaborative intelligence, taste, and a new bridge layer for creativity.

AI is reshaping the competitive landscape of productivity and creative tools.

Before the AI era, the biggest challenge for new entrants was competing against dominant incumbents. These companies owned comprehensive feature sets, mature workflows, massive user bases, and deeply ingrained user habits. The entire software industry was fundamentally built around the idea that humans learn tools and operate tools.

But AI-native entry products like Anthropic Claude, OpenAI Codex, and Anysphere Cursor are becoming a new kind of work interface. Users are starting to rely less on complex software operations, and instead collaborate with AI by describing goals and expressing intent.

The traditional barriers built by software companies — editor complexity, feature completeness, template ecosystems, and deeply rooted workflow dependencies — were all fundamentally products of a GUI-era design philosophy centered around humans operating software directly. But when humans begin collaborating directly with agents, the importance of software itself starts to decrease, while the importance of underlying capabilities increases.

Traditional productivity tools were fundamentally about improving efficiency. A great product was defined by being faster, cheaper, and more standardized. But AI is rapidly driving execution costs toward zero. Generation, layout, production, and editing are all becoming low-cost commodities.

What will matter more in the future is no longer how fast something can be made, but how creative, tasteful, and emotionally resonant it is. The truly valuable capabilities will become aesthetic judgment, creative reasoning, storytelling, and a deep understanding of culture and emotion.

If we continue following the old logic — building a smarter, more automated, more efficient version of Canva or Figma — we are still fundamentally building better tools. The paradigm itself remains unchanged.

From the user side, the next generation will naturally grow up in a world where AI is always present. They will no longer learn software first and create afterward. Their natural workflow will instead be continuous conversation, exploration, generation, and iteration with AI.

In that world, understanding products primarily through the lens of “tool supply” — adding more features, more buttons, and more complex editors — will increasingly stop making sense.

The real question for next-generation products is no longer: “What tools can we provide?” But rather: “How can we become collaborative intelligence within the user’s creative process?”

The opportunity is that any player now has the chance to redefine a new “bridge layer” between AI-native interfaces and their own products — a layer that helps users unlock creativity and amplify imagination.

[Original post](https://x.com/wangdouglas/status/2059807328468348977)