Show HN: Why write code if the LLM can just do the thing? (web app experiment)
Research#llm👥 Community|Analyzed: Jan 3, 2026 06:18•
Published: Nov 1, 2025 17:45
•1 min read
•Hacker NewsAnalysis
The article describes an experiment using an LLM to build a contact manager web app without writing code. The LLM handles database interaction, UI generation, and logic based on natural language input and feedback. While functional, the system suffers from significant performance issues (slow response times and high cost) and lacks UI consistency. The core takeaway is that the technology is promising but needs substantial improvements in speed and efficiency before it becomes practical.
Key Takeaways
- •LLMs can potentially replace code generation for simple applications.
- •Current performance (speed and cost) is a major bottleneck.
- •UI consistency is a challenge.
- •Significant improvements in inference speed are needed for practical use.
Reference / Citation
View Original"The capability exists; performance is the problem. When inference gets 10x faster, maybe the question shifts from "how do we generate better code?" to "why generate code at all?""