Comment by cgio
4 hours ago
If anything, that makes Jev look even less impressive. They may just be hiding their similar quality outputs behind good PR.
4 hours ago
If anything, that makes Jev look even less impressive. They may just be hiding their similar quality outputs behind good PR.
Laya is nowhere close to Jev in the 3rd party benchmarks I've seen, rather Laya is among the lowest scoring of the scores of Jev-alikes
example ranking shared in comment on this HN post
https://benchmarkheaven.com/jev-models
That is not related to the ad hominem above, or bear to the claim of innovation. Just comparing them implies they are similar. Implementation quality, training data etc. are all optimisations. I have not seen a compelling argument Jev is different, maybe one that it’s better. I don’t have huge insights in the domain, so open to learning or changing opinion.