You're not alone in that opinion. For some reason the British cycling media seems to think that he's Merckx MkII.
Knox looks to be the far better rider but gets a fraction of the focus.
I think the thing is that Tao was picked up earlier and had stronger results at 20-21, largely in the notoriously difficult-to-rate US scene admittedly at first, due to his time with Hagens Berman. Those races are always tough to judge because they're out of the way for the European-based péloton and oftentimes therefore come at a point in the form cycle which varies wildly, so Tao was producing some very competitive rides against some big names, but what form those big names had is variable. Obviously he also had some strong performances in U23 and espoir races in Europe, but the North American races meant there was a more visible body of work for Tao before turning pro than Knox had. Remember Tao has also been 2nd in the Tour of the Alpes and 5th in the Tour de Pologne this year. Knox had fewer results prior to turning pro, sure, but going to Deceuninck is a very different proposition to going to Ineos. Tao will be learning in a very disciplined, ranked environment where he will only get freedom to race for himself in select races, whereas DQS are a largely stage-wins and one-day-racing based organisation and, especially with their main stage racer (look, before July, nobody thought Alaphilippe was a GC leader) leaving at the end of the season, Knox has nobody who is going to stand in front of him in the queue for freedom in this kind of race.
I also wonder if Knox might be more of a Steven Kruijswijk; there were a lot of Dutch supertalents coming through at his time, with Gesink particularly notable with his almost podiuming the Vuelta and so on, and Mollema also. But the one that then nearly won a GT was Kruijswijk, who hadn't looked especially standout as an espoir - however what his big strength was was his recovery; most espoir races are short stage races so he hadn't been able to showcase this strength until already a pro. If you look at Knox's Vuelta, even taking the breakaway in the Cubilla stage out of it, he was around the 30-40 position in the first week mountain stages and around the 10-20 mark in the later ones.
Time will tell if this is an anomaly - remember, Phil Deignan got a 17 minute bonus in the Ávila stage in week 3 and finished 9th in the 2009 Vuelta, which similarly to this one had a very strong top 6 or so, then some domestiques and coattail-hangers, but he was 26 then, rather than 23 like Knox is.