I’ve published a new pair of articles on Ask Good Questions about a question I think is becoming increasingly important for developers who sell software.
AI can now create in minutes things that once required substantial programming time and specialized knowledge. That does not simply make us more productive. In some cases, it can remove the reason somebody would have purchased a commercial product or component at all.
So if the code itself becomes cheap, what are software companies really selling?
The Guide looks at that broader question and makes the case that the value is increasingly moving toward judgment, architecture, reliability, accumulated testing, integration, support, and accountability.
The companion Field Note looks at our own experience with Markdown editing. We actually began with an AI-generated Markdown editor that worked quite well for what we needed at the time. But after months of real product development in PageSnip, ProHelp Studio, and related work, the difference between that prototype and something we were willing to depend upon had become enormous.
When we eventually brought that work back to the component level, an AI-assisted architectural review recommended something interesting: do not improve the original editor. Rip it out and replace it.
That experience became a good illustration of the difference between software that works and software that has been turned into a product.
Guide:
https://askgoodquestions.dev/guides/when-code-becomes-cheap-what-are-software-companies-selling
Field Note:
https://askgoodquestions.dev/field-notes/ai-could-have-built-us-a-markdown-editor-in-an-afternoon
Charles