Has continual learning, as Dario said, really been solved?

The goal of continual learning research is to go beyond the scope of tasks that current techniques can reach. Models need to be able to learn new things while performing tasks, so that the time horizon of tasks can extend beyond weeks or months.

Engineering improvements like context management can significantly extend task duration, but very long-horizon tasks remain difficult. Human-level continual learning may have been solved to some degree, but not all of the change is complete yet.