Yeah, that seems like a strange sentence coming from a language designer who should know better. Just on the surface there is the differences of statically typed VS dynamic, class-based VS prototype-based, the implicit global state in JS, JS having first class functions and so on.

Longer quote seems weird as well

> To really get the performance improvements you see, it helps dramatically to have a statically-typed language. For dynamically-typed languages, like Python, it's really, really hard. And often, what people end up doing is adding annotations to the language so that you get languages like TypeScript, which is essentially JavaScript with type annotations. And it's really funny because JavaScript is essentially Java with the type declarations removed. So TypeScript is essentially Java with permuted syntax.

He seems to be saying that TypeScript leads to faster code execution as TypeScript has types, but fails to understand that TypeScript still compiles to JS and the browsers run JS without types, so should have the same performance as just writing JS directly (probably less performance, as TSC adds bunch of wrappers and other unnecessary stuff you would avoid if writing vanilla JS)