Using coding agents I am able to go beyond just writing code fast. I can write a lot more different type of code - that I was before AI. There are new powers that are afforded to me and I am able to get things that I always aspired to but couldn't get around to doing them. Definitely as well as I would have liked in some cases.

Even for the most motivated engineers the should haves and outside competence area items remain largely unachievable - in the context of their day-to-day work.

Should Haves

The absence of these items eat away at the Developer eXperience on daily basis - yet one never gets time to do these things. I have listed a few of them but there is a lot of software that goes into making software. Coding agents make these far more affordable and accessible to the developers. These supporting software contribute significantly towards improving the quality of the main product they are making.

Outside Competence Area

One can now implement a utility, less critical internal administration application - in a different tech stack, library, language. Or, one can even prototype specialized open-source ML models to real problems on the data.

The ability to generate code, apply on a real project, use it, review the code - is a great way to learn than doing off-band. Learning new things on a working software, on one's own project, is far easier and hence should be pursued - wherever appropriate.

If there are no UI designers on the project - a developer can improve the UI to far greater extent - than was possible before.

AI for software engineering can be also seen from a very different lens than offering raw productivity. It offers much improved developer experience and a many oppotunities for professional growth. The developers who are motivated will make the most of AI in these area - like any other technology.