They’re not talking about the squiggly line. The word return/retunr gets highlighted in slightly different colors based on whether the highlighter is picking it up as a keyword or a variable name, but in the scheme the colors are similar.
But their argument is based on the underlying assumption that it's the responsibility of the syntax highlighter to tell you when you've made a typo.
I am presenting an argument against this - and stating it's the responsibility of _diagnostics_ to tell you when you've made a typo, not syntax highlighting.
I would add that no highlighting would be applied to this mistyped word as it would not be recognised as a variable or a keyword. This may depend on the way syntax highlighting is handled. A common solution like treesitter would sus out that no variable with the name retunr exists, probably causing the mythical base colour to be applied, heh. That is arguably a big highlight in a colourscheme that rarely uses its base colour.
That's not what syntax highlighting is for though.
That's falls under diagnostics. A big red squiggly line. The honest test would be to add that in both cases.