Precisely. Languages have more or less information density ("bits per symbol"), but when you look at information rate per unit time, it all works out to pretty similar.
Kind of like how QAM256 can convey more bits per symbol than QAM16, but it needs a lower baud for a given noise level, channel bandwidth, and error rate. You can squeeze out efficiency of a channel, but you can't overcome the noisy channel coding theorem max rate.
If the cited claim was even remotely true, there would be significant differences in the "speed of thought" between speakers of natural languages also (which differ in semantic density). Such cognitive superiority would have obvious real-world consequences, but no such phenomenon appears to exist.