For 50 years, programming had one rule: be exact. Miss a semicolon — error. Misspell a variable — error. The computer did exactly what you told it, and if you didn't speak its precise language, it didn't understand you at all.
That era is ending. LLMs introduced something fundamentally different: intent-based interaction. You describe what you want in natural language. The model figures out how to achieve it.
The Shift to Fluid Programming
In traditional programming, there's usually one 'right' way to do something. With LLM-assisted building, multiple implementations can achieve the same result.
I call this fluid programming — it flows around obstacles instead of requiring you to remove them first. MIT's CSAIL confirmed this shift with frameworks like LILO (synthesizes and documents code), Ada (develops libraries of useful plans), and LGA (helps robots understand environments).
The key insight: natural language isn't just a nice interface — it's a rich source of context that helps AI build better abstractions.
The Vibe Coding Revolution
In developer circles, this shift is called vibe coding — describe what you want, AI handles implementation details.
Before: Idea, learn language, write code, debug syntax errors, fix logic errors, ship.
Now: Idea, describe it, AI generates code, test, refine description, ship.
The iteration loop is tighter and the barrier to starting is lower. But the flexibility comes with trade-offs:
- Less predictability — same prompt yields different code
- Hallucination risk
- Context dependency
- Maintenance questions
For mission-critical systems, these matter. For automation workflows, internal tools, and MVPs, the benefits usually outweigh the risks.
What This Means for Non-Technical Builders
The gap between 'I know what I want' and 'I can build it' is closing. The new skill isn't Python or JavaScript — it's describing what you want clearly enough for AI to build it.
This means breaking problems into clear steps, specifying edge cases, and iterating on description rather than code. Fluid programming shines for:
- Automation workflows
- Data transformation
- Rapid prototyping
- Internal tools
The human role shifts from syntax expert to intent articulator. The skill ceiling is still high, but the skill floor dropped dramatically.
Key Takeaways
- ✓ Programming is shifting from memorizing syntax to articulating intent — describe what you want, not how to code it
- ✓ Fluid programming excels for automation workflows, data transformation, rapid prototyping, and internal tools
- ✓ The skill floor dropped dramatically, but clear thinking and problem decomposition matter more than ever
Conclusion
You don't need to code. You need to know what you want.
The operators who learn to articulate clearly will build faster than those still waiting for the right developer. If you can describe your business processes, you can automate them. The barrier between idea and implementation has never been thinner.
Founder of d2b — building private AI automation and Gen-AI solutions for businesses across Europe.
