For years, software economics pushed users toward general products. A tool needed enough customers to justify design, engineering, hosting and support. The result was software that served a category well enough rather than an individual perfectly.
Code generation, hosted infrastructure and reusable APIs are changing that threshold. A useful internal tool can now be created for a team of five. A personal dashboard can exist for one person. A tiny automation can be worth building even if nobody intends to sell it.
From product choice to product composition
This does not mean everyone becomes a software engineer. It means more people can specify behavior and combine existing components without waiting for a market to produce exactly the right application.
The important shift is ownership of workflow. Instead of adapting a process to the software, people can increasingly adapt the software to the process.
Mass-market applications will remain where network effects and reliability matter. Around them, a long tail of personal software is likely to grow.