This blog post comes off as way too harsh, to say that it "doesn't seem like a very serious paper" is aggressive for no reason.
The criticism that they don't compare to Lc0 is valid. But this is a pre-print, the entire point is to gather feedback before publication. And indeed in a later revision they do compare to Lc0.
The claim that the authors concluded that their model is sometimes better than Stockfish is wrong. The authors never claimed that and the screenshot in the blog shows that the authors say that stockfish is better. So it looks like the blog author didn't even read their own screenshot carefully.
HN.zip
The criticism that they don't compare to Lc0 is valid. But this is a pre-print, the entire point is to gather feedback before publication. And indeed in a later revision they do compare to Lc0.
The claim that the authors concluded that their model is sometimes better than Stockfish is wrong. The authors never claimed that and the screenshot in the blog shows that the authors say that stockfish is better. So it looks like the blog author didn't even read their own screenshot carefully.