i think there will be less “advancements” than in the 1990s-2010s, like Moore’s Law will no longer apply. But there will always be new innovations.
The “next big thing” won’t be obvious like cryptocurrency, VR, or machine learning. Unless it requires something most people don’t have (e.g. maybe they’ll be more ML advancements from companies simply running huge neural networks, like GPT3 and copilot). Because otherwise, someone would’ve already done it. You’re right that the massive, obvious low-hanging fruit is already taken.
But there are still plenty of small innovations and things to improve. See: lots of the small libraries posted to HN, various new HTML/CSS/JS features, new packages on npm/gradle/rust, etc.
Honestly, while on the surface it seems like the internet hasn’t changed much since 2015, behind the scenes i believe that developer tools and libraries have gotten significantly better. JavaScript used to be a huge pain, and sometimes it still is, but tools like ES-modules and Vite are making it a lot easier. Kotlin, Swift, Rust, and Go have evolved (Go just got generics), their tooling has gotten less buggy, and more companies are adopting them.
And new things will become trends. See: TikTok. Old techniques and technologies will become popular again, as new technologies make them better/easier to implement, or just because people get bored and attention shifts.