Jas: AI-Paired Engineering as a Revival of N-Version Programming

A researcher demonstrates that AI-paired software engineering, combined with executable specifications and parallel implementations as safeguards, enabled a single developer to port a vector illustration application across five platforms (Rust, Swift, OCaml, Python, browser) in 120 hours. This approach revives N-version programming, a 1980s technique previously abandoned due to cost, making it economically viable by leveraging AI assistance.

This case study documents a significant shift in software development economics. The researcher successfully completed work that conventionally requires multiple developers over several months, using AI assistance paired with rigorous safeguards—an executable YAML specification and differential testing across implementations. The methodology addresses the fundamental challenge in AI-assisted coding: verification and correctness. By forcing the AI to target a precise specification and comparing outputs across multiple language implementations, the approach creates built-in redundancy that catches errors that single-implementation AI coding typically misses. This revival of N-version programming, originally proposed in the 1980s but abandoned because maintaining multiple codebases was prohibitively expensive, becomes economically rational when AI handles implementation complexity. The 23,000-line shared specification with per-port codebases ranging up to 95,000 lines suggests AI excels at translating specification into language-specific idioms while humans remain essential for specification clarity. For the developer ecosystem, this implies AI's highest value lies not in replacing developers but in enabling single developers to handle scope that previously required teams. The honest acknowledgment of limitations in this case study strengthens its credibility, suggesting the methodology has genuine constraints rather than being presented as a silver bullet. The work signals a potential productivity inflection point for software engineering, particularly for cross-platform development where specification clarity is achievable.

• →AI-paired engineering combined with executable specifications and parallel implementations creates a practical, verifiable approach to multi-platform development.
• →N-version programming becomes economically feasible when AI handles implementation complexity, reviving a 1980s technique abandoned on cost grounds.
• →A single developer delivered five production-quality platform ports in 120 hours using this methodology, suggesting significant productivity gains in targeted scenarios.
• →Precise specifications and differential testing across implementations serve as critical safeguards against AI coding errors and hallucinations.
• →The approach reveals AI's strength in translating specifications into language-specific implementations rather than in eliminating the need for human specification design.