A new prompt engineering project titled 'as-an-engineer' has surfaced on Hacker News, targeting a widespread frustration among developers: the tendency for Large Language Models to explain basic concepts to experts as if they were novices. Hosted on GitHub by user mayooot, the project offers a specific system prompt designed to override the default 'helpful assistant' persona. Instead of verbose, tutorial-style responses, the skill instructs the AI to assume the user possesses domain expertise, stripping away unnecessary context and focusing on high-level implementation details and trade-offs.

The Problem with Default Personas

For senior engineers and architects, the default behavior of most commercial LLMs is a productivity killer. When you ask for a database schema optimization or a complex algorithmic review, you don't need a definition of what a primary key is. The 'as-an-engineer' skill addresses this by shifting the interaction model from teacher-student to peer-to-peer. It encourages the model to use industry-standard shorthand, reference specific libraries and patterns without explanation, and prioritize architectural decisions over syntax tutorials.

Early Reception and Context

The project was submitted to Hacker News on September 13, 2026, but has yet to gain significant traction, currently sitting at a score of 1 with zero comments. Despite the low visibility, it taps into a growing niche of 'power user' prompts that curate the AI's output density. As AI agents become more integrated into development workflows, the ability to customize their communication style is becoming as critical as the models' reasoning capabilities themselves. This project is a simple but effective example of prompt engineering acting as a layer of abstraction for human preference.

Key Takeaways

  • The 'as-an-engineer' skill modifies system prompts to assume user expertise, reducing token waste on basic explanations.
  • It aims to replace the 'helpful assistant' bias with a 'peer collaborator' tone, focusing on trade-offs and implementation.
  • The project is currently early-stage with minimal community engagement on Hacker News.

The Bottom Line

We don't need an AI to hold our hands; we need it to keep up. This skill is a necessary step toward treating LLMs as senior engineers rather than eager interns.