LLMs have] a very limited understanding of logic, and don't understand the physical world, do not have persistent memory, cannot reason in any reasonable definition of the term and cannot plan hierarchically.
We could still get to that next stage, with several groups (including Meta) working on Artificial general intelligence.
Once we have techniques to learn "world models" by just watching the world go by, combine this with planning techniques, and perhaps short-term memory systems, then we might have a path towards cat-level intelligence.
It's more pressing in the case of X, given that it's already, as noted, using its Grok AI tools to generate news headlines in the app.
[
add
]
[
|
|
...
]