Some insights/explainations about the format.
Tl;dr : he is skilled, had a incredible run but the deck has been quite mucky stacked in his favor.
It's based on the Argentinian format El Legado - that spawned the very successful format L'Eredita in Italy or Crésis in France years before.
In this version, four contestants, including the reigning champ are playing each day, feeded with a virtual €10000 personal jackpot. Three rounds of straight quizzing, first one being two-choice questions with the second choice hidden until you select it, second round being "avoid the wrong choice from this list of answers provided" then the two remaining contestants battle it out in a Grand Slam-esque chess clock battle, winner coming back next day, trying to cash in their jackpot in the final round (+having a shot at a progressive prize package in a bonus round)
Losers of first and second round do not get eliminated straight away - they get to challenge one of the other contestants to a duel in a final, four-choice question, answered by the challenged player. Whoever wins stay in the game and adds the other player's jackpot to theirs.
With all those details on mind, you could argue that Emilien's achievement is quite impressive - and indeed he's shown very impressive general knowledge skills during his run, especially as he's only 22 years-old.
But.. it's not that easy. I won't surprise many of you by pointing first that he was we was casted as the profile for their next big long run champ - tv business, better for ratings to have a long running champ, with low or none challenging profiles against him each day (and better for business as TF1 as a premium-call viewer competition with a growing jackpot paid out when the champ gets knocked out - one lucky viewer got a huge 1.75M€ payout along with Emilien leaving the show)
Then there's some equity issues.While questions are still sourced and checked - there's no guarantee of a fair and square competition. Leading to many people noticing straightforward recycled questions in the duel part or even worse - recycled exact set of answers, same order, same illustration pictures but with an question switching from "easy" to "way harder" and therefore some suspicions that there's two questions ready to be switched depending on what the production wants to happen (one example here
https://x.com/brillant_j52653/status/1850551731903467969). Similar in the chess clock part where people trying to compile data about the challenger being asked longer questions (making them lose more time on their clock) and/or champ being sometime asked ridiculously easy questions such as "What do you turn in order to read a book?". Obviously less easy to prove.