Poll: What will a better LLM do for me?

A poll on Hacker News asks whether users are working on problems that current LLMs cannot solve but that a next-generation first-tier LLM might, with early responses indicating most users are not gated on better models. The poll offers options such as using LLMs for research and ad hoc problem-solving, having a specific gated problem, or using coding agents that already mostly solve their needs.

Is anyone here working on a problem for which current generation LLMs are inadequate, but that could possibly be solved by the next release of a first tier LLM? No, I use LLMs for research and ad hoc problem-solving, so improvements are better but I'm not working on a specific unsolved problem that a better LLM would solve. 1 point Yes, I have a specific problem, which I have described in a comment, the solution for which is gated on better LLMs. 1 point Other, please describe. 1 point No, but I use coding agents so my problem is mostly solved but I welcome all incremental gains.