Popularity is a poor judge of quality. But if you meant that the author used more C/C++ than Forth itself, you do have a point. I wouldn't respect a system programming language that can not be used to implement itself (bootstrap). That's exactly the reason why I didn't start to publicize my own. But if you want to know, there is a language named 8th that you can use to make multi-platform applications. In my book, that would be proof enough that it can be done.
Now, can it be done productively ? Is the 10x claim verifiable ? That depends on your skill. Maybe it's not for everyone's taste. If you prefer to let the compiler do the compiling for you, maybe not. But that being said, Forth makes you more productive because it allows you to change the problem. Modern environments are bloated and waste so many resources. Libraries are huge, software is huge.
One user only require 10% of the functionality, yet you pay for all this fat. When you do real Forth, you do away with the unnecessary... to the extreme. You also change the compiler and the syntax, when it doesn't map to the problem you are solving. With Forth, you can change the problem. Other programming languages can't do that.
Prove that. Record some videos, show me some code and metrics that'll be evidence that I'm getting 10x more productive.
Look at the languages in the repository:
C 50.6%
C++ 47.6%
Forth 1.5% <-----
HTML 0.1%
Makefile 0.1%
JavaScript 0.1%
How does that build credibility?