Those alternatives are old tech that has way more limitations than a neural implant.
Just because there is old tech that SORT OF does this, doesn’t mean it can’t be improved. That’s the same attitude behind “not needing more than 4mb of RAM” back in the day. You can’t stop progress all because YOU are fine with the current state of the tech.
Think we might need either a new type of image AI or maybe expanded latent space with just a shit load of new parameters. Anything to improve comprehension and prompt attention.