A hypothesis only upgrades to theory if evidence supports said hypothesis. This still has uncertainty, which is why theories tend to be debunked over time. That's what he's saying.

Hypotheses are taken to be true given the weight of evidence. [b]This is not proof[/b]. It's as close to proof as we can replicate. Take an imaginary hypothesis about stars. I can hypothesise that "All stars must have property X". I can look at every star, and see whether or not it has property X. If every star has property X, I can say with pretty damn good certainty that my hypothesis is true - but I'll never [i]know[/i] if it's correct. In science "proven" is used as shorthand for "very probably true". Having said that, mathematical proof is quite different, and is indeed empirical and universal, provided mathematical axioms are correct.