2025 Software Development Wishlist
Posted on:
Here's some things I hope for in the new year for Software development space
The Continuation of JS Bloat Reduction
Initiatives like e18e spun up in 2024 to start actually cleaning up the JS ecosystem. I'm okay with a node module for anything you wish, but when we have libraries that still have polyfills or libraries that have modern replacements / built into the language or runtime for years now, there is a point that the node_modules folder can start to downsize.
GitHub pushes on more non-AI tooling
Very often I see a VS Code release notes where a huge chunk is about what new thing GitHub Copilot can do, and although it is a big revenue push for GitHub, there is a lot of tooling that I feel GitHub is putting on the back burner. Whether it is improvements in their CI/CD with Actions, or improvements to their Issues & Discussions experiences, to a further push on software supply chain security, I hope that GitHub continues to work on these, is vocal about such, and make them widely available and not just for enterprise customers.
AI Hype Train slows down
GenAI has been riding an infinite amount of VC funding and popularity, but I hope that we soon reach a hump where people start realizing that AI does not actually need to be everywhere. As well, I hope that more people become skeptical of the tech that exists now and be proactive and smart about using these tools and become LLM literate to the point that a developer will put in due dilligance when reading LLM output rather than just trusting it from the get-go.
Human Creation & Curation in Software Projects
Alongside the last section, as more GenAI Slop is made, I feel the human touch of projects is going away. Whether it is readmes, discussions, documentation, all of these places are mainly not code, and you can let an LLM write these things for you, but I feel many projects who decide to pay real people to do these jobs are going to be better than just outsourcing this work to GenAI. Also, alongside that, pay people for high quality work, not downsize thinking GenAI can replace a human being, cause it just can't in many aspects.