Capability Threshold And Feasibility Shift For Small Software
Key takeaways
- AI assistance enabled Ashworth to deliver a polished application for an audience of about three people in only a few days, which he indicated would have been economically unjustifiable otherwise.
- Ashworth described shipping AI-generated code that the programmer has not reviewed or does not understand as terrifying and unacceptable.
- Chris Ashworth is the creator and CEO of QLab, a macOS cue-based multimedia playback tool used to automate lighting and audio for live theatre.
- Ashworth believes AI coding tools do not make someone a fundamentally better programmer but instead amplify the speed at which they produce code aligned with their existing skill level.
- Ashworth compared AI coding tools to power tools that increase productivity for trained users while posing high risk to untrained users.
Sections
Capability Threshold And Feasibility Shift For Small Software
The corpus reports a perceived step-change in coding-assistant usefulness over a multi-year evaluation window and ties it to an outcome: delivering a polished app in days for an extremely small audience that was previously not worth building. The combined effect is a feasibility shift toward producing bespoke or long-tail tools under tight time and ROI constraints, at least for this practitioner and project type.
- AI assistance enabled Ashworth to deliver a polished application for an audience of about three people in only a few days, which he indicated would have been economically unjustifiable otherwise.
- Ashworth used Claude Code to build a custom lighting design application for a niche project in a few days that he would not otherwise have had time to create.
- After two years of skepticism and repeated evaluation, Ashworth found recent AI coding capabilities to be meaningfully improved and described them as astonishing.
- Despite personal dislike and community opposition to the technology, Ashworth expects AI-assisted programming to be a career-changing moment.
Mechanism: Skill Amplification With Strict Safety And Review Conditions
The corpus frames AI coding tools as accelerants rather than skill creators, implying outcomes depend strongly on user competence. It also specifies governance conditions (understanding, ability to edit/delete, and quality control) and rejects unreviewed shipping, highlighting that operational safety norms remain central even when productivity improves.
- Ashworth described shipping AI-generated code that the programmer has not reviewed or does not understand as terrifying and unacceptable.
- Ashworth believes AI coding tools do not make someone a fundamentally better programmer but instead amplify the speed at which they produce code aligned with their existing skill level.
- Ashworth compared AI coding tools to power tools that increase productivity for trained users while posing high risk to untrained users.
- Ashworth stated that AI-assisted programming is valuable when the user can understand, direct, edit, delete, and quality-control the code it produces.
Domain Anchoring: Theatre Automation Context With An Experimentation Venue
The corpus situates the claims in live-theatre automation software and notes an associated venue used for performance, teaching, and research, which plausibly supports iterative experimentation and dissemination in that community. The reported app-building example is specific to a lighting design niche, grounding the feasibility shift in a concrete operational domain.
- AI assistance enabled Ashworth to deliver a polished application for an audience of about three people in only a few days, which he indicated would have been economically unjustifiable otherwise.
- Chris Ashworth is the creator and CEO of QLab, a macOS cue-based multimedia playback tool used to automate lighting and audio for live theatre.
- Ashworth founded the Voxel Theatre in Baltimore, which is used by QLab as a combined performance venue, teaching hub, and research lab.
- Ashworth used Claude Code to build a custom lighting design application for a niche project in a few days that he would not otherwise have had time to create.
Unknowns
- What specific quality metrics (defect rate, test coverage, maintainability, security findings) changed for the delivered app relative to Ashworth's non-AI workflow?
- Which exact model versions, tool configurations, and surrounding developer tooling (tests, linters, CI) were used when capabilities became "astonishing"?
- How generalizable is the reported productivity gain across different skill levels and roles (novice vs expert programmers, designers with limited coding experience)?
- What governance practices were actually applied (code review rigor, mandatory tests, access controls) in the example project, and how much effort did review/cleanup require?
- What is the nature and magnitude of the "community opposition" mentioned, and does it measurably affect adoption or distribution of AI-assisted tools in this arts community?