Ask HN: Is anyone experimenting with different ways of using LLMs for coding?
A Hacker News user finds LLM coding tools like Claude Code disrupt flow state and asks if anyone is exploring alternatives beyond the prompt-response loop, suggesting a tab-completion model as a better direction.
Background
This is a discussion on Hacker News (a tech-focused social news site run by Y Combinator) where a developer asks about alternative ways to use LLMs for programming.
Key context:
- The user mentions "Claude Code" (Anthropic's terminal-based coding agent) and "Codex" (OpenAI's coding tool) — these are the current mainstream AI coding assistants that work via a chat-like prompt-response loop.
- The "bicycle for the mind" reference is from Steve Jobs's 1990 interview, where he described computers as tools that amplify human intelligence.
- "Tab model" refers to tools like GitHub Copilot that suggest code completions inline as you type (tab to accept), rather than requiring a separate chat window.
- The frustration described — constant stopping to review and re-prompt — is a common complaint among experienced developers who feel current AI tools break their flow rather than enhancing it.
- Hacker News "Ask HN" posts are community questions where users solicit advice, tools, and experiments from the broader tech audience.