Video models can certainly predict what stirring cream into coffee does.
Most LLMs these days are limited by context, and within context they are super human at reasoning tasks/per unit time.
I agree with the author in one sense in that "intellectual" tasks are a small part of our evolutionary history and I wouldn't be surprised if we were replaced by silicon in that regard. Even individual bees at a small number of neurons can ingest enormous amounts of data and plan over a year.
Most LLMs these days are limited by context, and within context they are super human at reasoning tasks/per unit time.
I agree with the author in one sense in that "intellectual" tasks are a small part of our evolutionary history and I wouldn't be surprised if we were replaced by silicon in that regard. Even individual bees at a small number of neurons can ingest enormous amounts of data and plan over a year.