Will an LLM improve its own ability along some important metric well beyond the best trained LLMs before 2026?

Will there by a major breakthrough in LLM continual learning before 2027?

Will there be any simple text-based task that most humans can solve, but top LLMs can't? By the end of 2026

Will an LLM be able to solve the Self-Referential Aptitude Test before 2027?

Will there be any major breakthrough in LLM continual learning before 2029?

Will there be any major breakthrough in LLM continual learning before 2028?

Will there be any major breakthrough in LLM continual learning before 2030?

Will LLMs be used for academic peer review by 2030?

Will prompting of generalist models be the dominant method of interaction with LLMs on 1 January 2028?

Will the most advanced LLM stop being from a US-based company any time before 2030?