> do you think Fermat had an error

Yes, with million-to-one odds. Though I'd say "error or equivalent shortcoming, for the general case".

> that perhaps is more straightforward than Wiles's?

I don't recall a mathematical definition of "straightforward", but yes. Ignoring minor improvements, I'd guess there's a much better proof...though that might require a century of new developments in related parts of mathematics, before we'll have the necessary tools to write it down.