Hacker News Reader

It depends! If a startup is entering with a large model to face other larger models, it must be better at least in 1 meaningful dimension.

500B params performing worse than other OSS of the same size is pretty meaningless if no one will use it.

View on HN

Look at meta, while they went from open to closed, they got better as time went on

I mean, the obvious way it's better than the Chinese open-weight models is literally the fact that it's not Chinese and is therefore less likely to be banned or restricted. Chinese models cannot be used on certain government systems already in the US, and regulators are actively considering expanding these restrictions more generally (such as adding it to the Entity List).

Reply on news.ycombinator.com