Yes, new frameworks will have a harder time getting uptake.
Worse, with LLM's easily generating boilerplate, there's less pressure to make old framework code concise or clear, and the superior usability of a new framework won't be a big draw.
But coding is a primary application/profit center, and you can be sure they'll reduce the latency between release and model support, and they'll start to emphasize/suggest new frameworks/paradigms as a distinguishing feature.
My concern is about gaming the system, like SEO. If LLM coding is the gatekeeper, they'll be corrupted by companies seeking access/exposure. Developer uptake used to be a reasonable measure of quality, but in this new world it might only reflect exposure.
Worse, with LLM's easily generating boilerplate, there's less pressure to make old framework code concise or clear, and the superior usability of a new framework won't be a big draw.
But coding is a primary application/profit center, and you can be sure they'll reduce the latency between release and model support, and they'll start to emphasize/suggest new frameworks/paradigms as a distinguishing feature.
My concern is about gaming the system, like SEO. If LLM coding is the gatekeeper, they'll be corrupted by companies seeking access/exposure. Developer uptake used to be a reasonable measure of quality, but in this new world it might only reflect exposure.