Assuming they updated the crawled training data, just having a bunch of examples of specifically pelicans on bicycles from other models is likely to make a difference.
But then how does the quality increase? Normally we hear that when models are trained on the output of other models the style becomes very muted and various other issues start to appear. But this probably the best pelicans on a bicycle I've ever seen, by quite some margin.
I thought a human would be a considerable step up in complexity but I asked it first for a pelican[0] and then for a rat [1] to get out of the bird world and it did a great job on both.
But just fot thrills I also asked for a "punk rocker"[2] and the result--while not perfect--is leaps and bounds above anything from the last generation.
0 -- ok, here's the first hurdle! It's giving me "something went wrong" when I try to get a share link on any of my artifacts. So for now it'll have to be a "trust me bro" and I'll try to edit this comment soon.
Right, that chart shows App usage based on the user-agent header but doesn't tell you if there is a single individual user of an app that skews the results.
I was skewing the Gemini starts with my Aider usage. Basically the only model in using with openrouter, until I recently started running qwen3-next locally.
2.5 is probably the best balance for tools like Aider.
Yeah, I forgot there’s the intermediate VM level, and user folders are shared there so that folders could be mounted to the individual containers using host paths.
My ISP one got kicked to the curb once they started returning results for anything including invalid sites. Basically to try to steer you towards their search.
> As with a lot of judge rulings, and what they're always supposed to do, they ruled on what the actual law is and not just on what sounds good.
A "lot" of judicial rulings do indeed follow that pattern. But there have been mulitple high-profile & high-stakes examples recently of just the opposite. To the point where I thought you were making a joke at first.
reply