The point being that gist translation is not good enough to sell something or do justice to important source writing of any language.
The myth continues to be perpetuated that 'good enough' translation works for foreign readers, undermining the role of human translation, which is actually doing very well as a massively growing industry worth several billion dollars, but could be doing a lot better if people understood the implications of language issues more widely.
And, to be honest, there isn't much to understand and it all boils down to respecting the original writer and target reader enough to do both parties justice - one wants to give specific information for a particular goal, the other wants to receive specific information for their particular goal.
Other things to note
- Using English as a bridge language (between Russian and Japanese, to take the original example) multiplies the chances of errors being made, further polluting MT translations
- Statistical analysis still consistently strips text of any implied meaning