This and that are different, languages like Go and Erlang use m:n mapping to map their m lighweight threads/processes onto n kernel threads (where n is usually the number of physical cores on the machine) to benefit from hardware concurrency.

But the presentation is about concurrency and parallelism not being the same thing, and more concurrency not necessarily leading to more parallelism.

> Don't make excuses for your language.

There's no excuse made.

> Just solve the problem!

There's no problem to solve. If you're asked to count from 1 to 100, you can't parallelize it: it's a purely sequential task, each step has a hard dependency on the previous step. So there is no possible parallelism.

But it is possible to do something else at the same time, or to have one process handling the counting and sending the current count to an other process printing it. That's concurrency: you've got two things running at the same time.