Solo development was always a story about trade-offs: one person can't be a great programmer, artist, composer, and designer simultaneously, so solo devs historically picked their strongest skill and either learned the rest badly or scoped around the gap entirely. That trade-off hasn't disappeared, but its shape has changed a lot.
What's genuinely easier now
The floor for the skills you're weakest at has risen a lot. A solo programmer who can't draw can now get usable art into their game without either learning illustration from scratch or hiring someone they can't afford yet. A solo artist who's never written a line of code can get working game logic without a computer science background. This doesn't make anyone an expert in their weak area, it makes "not an expert" no longer a hard blocker to shipping something.
What hasn't changed at all
Scope discipline. Knowing when to cut a feature. The judgment call about what makes your specific game worth playing instead of a competent but forgettable clone of something else. AI tools make execution faster in every individual skill lane, but the project-level decisions, what to build, what to cut, when it's done, are exactly as hard as they always were, and arguably more important now that execution speed alone isn't a differentiator anymore.
The new trap that didn't exist before
Faster execution makes it easier to keep adding scope instead of finishing, because every new feature feels cheap to try. The classic solo-dev failure mode, a project that never ships because it never stops growing, is if anything easier to fall into now, not harder, because the cost of saying yes to one more idea has dropped. Finishing is still a discipline, not a byproduct of better tools.
The honest framing
AI assistance moves the bottleneck for a solo developer from "can I technically build this" to "do I actually know what I'm building and when to stop." That's a better bottleneck to have, technical skill was never the fun part for most people, but it's still a bottleneck, and pretending otherwise is how a solo project quietly turns into a five-year unfinished demo.