HN
Today

Software for One

This piece argues that the concept of "personal software" has finally arrived, enabled by the dramatic advancements in AI and agentic coding. The author demonstrates how LLMs drastically reduce the time and cost to develop bespoke applications for individual needs, making custom tools feasible for anyone. It suggests this paradigm shift will lead to a new era where highly personalized software becomes the expected norm, fulfilling long-held desires for a "HyperCard for the modern age."

6
Score
0
Comments
#5
Highest Rank
18h
on Front Page
First Seen
Aug 1, 2:00 AM
Last Seen
Aug 1, 7:00 PM
Rank Over Time
2756782181011121415161625242628

The Lowdown

The article posits that the vision of "personal software," where individuals create custom applications for their own specific needs, has materialized in 2026 due to AI advancements. Referencing Robin Sloan's 2020 concept of a "home-cooked app," the author details how LLMs and modern development stacks have made this a practical and efficient endeavor.

The author provides compelling examples of personalized software built for his household:

  • A dynamic sleep app that manages complex conditional logic for his son's sleep schedule, accessible by multiple caregivers.
  • A fitness app that customizes smoothie ingredients and calorie targets based on daily running mileage, integrating data from various sources like Strava and Oura.
  • A marathon training plan generated from personal race history and heart rate zones, rather than generic age-based formulas.
  • A "Duolingo for jazz" that quizzes him on chord voicings relevant to his piano lessons.
  • A medical records tool that flagged gaps in his health data prior to a specialist visit.

Key learnings from this rapid development cycle include a dramatic reduction in building costs (e.g., a fitness app in a weekend), surprisingly easy maintenance with AI assistance, and the power of aggregating personal data with LLMs for unique insights. The author notes that "ephemeral" apps are now viable, and that AI unlocks small, niche markets that wouldn't attract commercial developers. While current tools still require some technical skill, the author anticipates a future where building such personal software will be as simple as describing a need to an AI agent, fulfilling Sloan's wish for a "modern, flexible HyperCard for iOS."