Thanks for the article @flis, some questions:-
Will the project parse discussions / information on public content blockchains like Steem?
How will machine-learning, or pattern-recognition of large datasets be involved? Or are information "reduced" to TML which then doesn't require such a practice?