One of the clearest benefits of AI is not creating everything from scratch.
It is going back to old scripts, tools, and unfinished projects, understanding what they were supposed to do, and improving them with a different perspective.
This only works when the project is reasonably organized and the context is explained properly. AI needs more than a folder full of files. It needs to understand the objective, the existing logic, the limitations, and what the project was missing.
When that context is available, the value can be very real.
A good example is an old web project I built around the solar system.
At the time, large language models did not exist. The original exercise (Codecademy – “Sun, Earth, and Code) was simple: create a sun, add the Earth, and make it rotate around it.
I pushed it much further than the initial exercise at the time...
I added the rest of the solar system and tried to turn the basic animation into something more complete. But eventually, the features I wanted to add required more time and development knowledge than I was willing to invest. I am not a developer. I simply like taking a basic idea and seeing how far I can push it.
The project stayed unfinished for years.
Today, I was finally able to reopen it and add some of the features I originally had in mind: a navigation menu, zoom capabilities, and a more complete visualization of the solar system.
AI did not invent the project. It did not replace the work I had already done. It helped me finish the old structure, correct parts of it, adapt the code, and continue from a foundation that already existed.
Old scripts and unfinished tools often still contain useful ideas. The problem is that their author may no longer remember every decision, the technology may have moved on, or the effort required to restart the project may feel disproportionate.
AI lowers that restart cost.
It can help document undocumented code, identify weak logic, modernize parts of a project, propose a cleaner structure, or implement features that were previously out of reach. It also provides something that was often missing when these projects were created: a second technical perspective available on demand.
This does not mean that every old project should be revived. Some should remain archived. Some are based on outdated assumptions. Some are simply not worth maintaining.
But dismissing everything because it is old is also a mistake.
When the objective is still relevant and the context is clear, AI can turn an abandoned experiment into something useful, finished, or at least much closer to the original vision.
For people who build scripts and small tools (like me !) without being full-time developers, that changes a lot.
AI is no longer a tool that can simply be ignored. Not because it should be used everywhere, but because in the right context, it gives old work a second chance.
Project link
See it in action here : https://apps.adhomelabs.ch
0 Comments