Most programming tasks are exactly like that. Is this agent able to complete this task? Is this agent able to optimize a kernel beyond previous attempts?
Of course some are subjective and that's where progress is harder, like "Is this website pretty?". But for tasks that can be objectively measured, LLMs will go beyond human level, just like with Chess and Go.
Of course some are subjective and that's where progress is harder, like "Is this website pretty?". But for tasks that can be objectively measured, LLMs will go beyond human level, just like with Chess and Go.
That's why RL is so important when training LLMs.