Skip to content

deansher/awesome-ai-software-engineering

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 

Repository files navigation

Awesome AI Software Engineering

Practical resources for achieving 10x development velocity with AI coding assistants

The Pocket AI Singularity

This repository provides resources to help development teams achieve dramatic productivity gains (3x to 10x) through effective collaboration with AI coding assistants. We call this approach the "Pocket AI Singularity" – exponential acceleration that you control.

The Core Insight

Today's AI coding agents (Claude Code, Cursor, Aider, etc.) are so fast and capable that we can finally afford all the investments in code and system agility that we've always wanted:

  • Thorough, automated tests – Unit and integration tests that actually get written and maintained
  • Current documentation – Design docs and specs that stay up-to-date
  • Helpful tooling – Custom tools for every recurring task
  • Continuous refactoring – Clean code that stays clean

The breakthrough: AI agents benefit even more from these investments than humans do. When you direct your AI colleague to create great docs, tests, and tools – and then use them – you gain rapidly compounding returns. This creates a virtuous cycle of exponential productivity gains.

Why Most Teams Are Stuck at 10-30% Gains

While AI models have drastically improved throughout 2025, most teams haven't adapted their workflows to leverage these capabilities. Common obstacles include:

  • Treating AI as a fancy autocomplete rather than a capable colleague
  • Not investing in the documentation and tests that make AI most effective
  • Missing the compounding returns from systematic agility investments
  • Lack of team alignment on tools and practices

The Path to 10x

This repository provides practical resources to help your team:

  1. Choose and configure the right AI tools – Current recommendations and team alignment strategies
  2. Develop effective prompts – Core prompts that shape your AI colleague's behavior
  3. Establish productive workflows – Step-by-step practices for AI-assisted development
  4. Build context effectively – Help your AI colleague understand the full picture
  5. Create living documentation – Plans and specs that guide both humans and AI
  6. Leverage compounding returns – Systematically invest in code agility

What You'll Find Here

  • 📝 deans-coding-agent-prompt.md – A comprehensive prompt for AI coding assistants that emphasizes incremental development, thorough testing, and continuous improvement
  • 🎯 Workflow templates – Structured approaches for common development tasks (coming soon)
  • 🔧 Tool configurations – Optimal settings for various AI coding assistants (coming soon)
  • 📚 Case studies – Real examples of 10x productivity gains (coming soon)
  • 🤝 Team practices – How to align your team around AI-assisted development (coming soon)

Key Principles

1. Treat AI as a Capable Colleague

Your AI assistant excels at:

  • Writing large volumes of clean, idiomatic code
  • Following patterns consistently
  • Patient, methodical testing
  • Comprehensive refactoring

You excel at:

  • Strategic thinking and architecture decisions
  • Understanding business context
  • Spotting conceptual errors
  • Making judgment calls

2. Invest in Agility at AI Speed

With AI, the ROI on agility investments has fundamentally changed. What used to take days now takes hours or minutes. This means:

  • Write the test first (TDD is finally practical for everything)
  • Document as you go (specs write themselves)
  • Refactor continuously (it's essentially free)
  • Automate everything (custom tools are trivial to create)

3. Build Context Thoughtfully

Help your AI colleague succeed by:

  • Providing comprehensive context upfront
  • Creating and maintaining plan documents
  • Pointing to relevant existing code and documentation
  • Discussing the approach before diving into implementation

4. Embrace the Compound Effect

Each investment in documentation, testing, and tooling:

  • Makes the next task easier
  • Reduces errors and rework
  • Accelerates future development
  • Improves AI effectiveness

Getting Started

  1. Read the core prompt: Start with deans-coding-agent-prompt.md to understand the collaborative approach
  2. Choose your tools: Select an AI coding assistant (currently recommending Cursor with Claude Opus 4.1)
  3. Align your team: Agree on consistent tools and practices
  4. Start small: Pick a well-scoped project to practice these techniques
  5. Measure and iterate: Track your productivity gains and refine your approach

Current Tool Recommendations (September 2025)

Primary Setup:

  • Tool: Cursor
  • Model: Claude Opus 4.1 in "max" mode
  • Cost: ~$3,000/month for full-time use
  • ROI: 5-10x productivity multiplier

Alternatives Worth Considering:

  • Claude Code
  • Aider
  • OpenHands
  • Models: Claude Sonnet 3.5, GPT-4

Contributing

This is a living repository. Contributions are welcome! Please share:

  • Effective prompts and workflows
  • Success stories and case studies
  • Tool configurations and tips
  • Team practices that work

About

This approach was developed through extensive hands-on experience with AI-assisted development across 2024-2025, including production deployments at Admired Leadership and intensive experimentation with various AI coding tools.

The goal is simple: help development teams achieve the 10x productivity gains that are possible today, while maintaining or improving code quality, test coverage, and system maintainability.


"The future is already here – it's just not evenly distributed." – William Gibson

Let's distribute it.

About

Resources and practices for effective AI-assisted software development

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published