Execute actions and cognition that pay back the cost of said actions, and support the next generation. No intelligence can appear outside social bootstrapping, it always needs someone pay the initial costs. So the cost of execution drives a need for efficiency, which is intelligence.
Current AIs cannot comprehend on the fly, meaning if they are presented with data outside of their training, the reply generated will be a hallucination interpolated off the training data into unknown output. Yet, a person in possession of comprehension can go beyond their training, on the fly, and that is how humans learn. AI's cannot do that, which is critical.
I agree with you, current models can't work totally outside their training set. An example of AI that trained with environment and feedback/outcome learning is AlphaZero, and it totally beat us at our own game. Even so, DeepMind seems not to care to pay the costs of further development, so we see LLMs need to make themselves useful to people to survive. It's a "pay your costs or stop executing" situation.
