| ▲ | anuramat 3 hours ago | |||||||||||||||||||||||||||||||
"some model I don't get to use is much better at benchmarks" pick one or more: comically huge model, test time scaling at 10e12W, benchmark overfit | ||||||||||||||||||||||||||||||||
| ▲ | estearum 3 hours ago | parent [-] | |||||||||||||||||||||||||||||||
So... you're not excited because it might take a few months before we can use it or something? I don't get your comment. | ||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||