When We Stop Using the Same Software
Different islands. A common thread.
A speculative essay on personal tools, shared software, security maintenance, and why we should extend what works before building another app.
Concept by Shreyam Adhikari. Writing: AI-assisted; presentation: AI-built. This speculative essay imagines a timeline from 2020 to 2050. Future dates describe the imagined world, rather than an adoption forecast.
Personal tools can fit individual needs, while browsers, professional creative tools, and dependable text editors continue to provide common ground.
Shared software gives people common places to learn. Familiar platform conventions help that knowledge travel: Ctrl+P often prints in Windows applications, with Command+P on a Mac.
Shared software also carries a social contract: people choose common places to communicate, learn, and help each other. Personal tools can improve an individual's workflow while making shared activity harder.
Compatible formats help people exchange editable work and leave a tool without starting over. Maintained foundations help them repair it. OpenTimelineIO provides a concrete example of editorial interchange, within its supported scope.
Security is also a maintenance question. In 2026, Windows Paint, Ubuntu's libarchive, and libcurl had published advisories or fixes. The curl advisory credits an AI-assisted report. Shared code can spread one flaw to many apps, while shared advisories and patches help only when affected products update.
We do not need another application merely because making one is easier. Before building one, ask whether shared software already meets the need. Configure it, extend it for a meaningful gap, and build a new tool when the remaining need warrants its maintenance responsibility.
Existing software makers could offer agents inside modular applications, letting people ask for a focused editing workspace or a tested reader for a custom format. Personalised experiences could share a maintained base and an extension system with supported interfaces, permissions, compatibility checks, and an update path.
AI builders could carry established controls, clear saving behaviour, suitable documented formats, and maintained components into new tools. Different islands need connecting threads. Just because we can build something, should we?
Topics
- personal-software
- AI
- shared-conventions
- interoperability
- software-maintenance
- software-security
Sources
- Nielsen Norman Group: Jakob’s Law of Internet User Experience
- Nielsen Norman Group: Consistency and Standards
- Microsoft: Windows keyboard tips and tricks
- Apple: Mac keyboard shortcuts
- OpenTimelineIO: an interchange format for editorial timeline information
- Ink and Switch: Malleable Software
- Microsoft: Windows Paint security advisory
- Ubuntu: libarchive security notice
- curl: HTTP/2 upload security advisory