1. Elon Musk has admitted many times since that point that he was wrong (something Trevor never does) about factory automation.
2. The solar tiles exist.
3. Turns out making super good AI takes more time and Elon is always too optimistic, that is still their goal however that they are working towards.
4. That's still their goal and they're working towards it.
You can't call someone a fraud just because they haven't reached a goal yet and it's taking longer than expected.
I could potentially see such a thing working for, say, epilepsy, where we might be able to "mute" or counter a stimuli that causes seizure, but, at least for the forseeable future, I don't see "brain chips" working in contexts other than concrete causal links, stimuli, reaction/response mechanisms - I think one hint is that there's nothing we can do to "medicate" autism. Medications currently used are for "byproducts" of ASD, such as ADHD, anxiety, not the disorder itself. Indeed, the closest (and we're not particularly close) we have come to thinking of how to manage ASD is gene therapy, which wouldn't be manageable by a "brain chip".