GPT 4: A New Kind of Intelligence

The problems are so deep and I think they reflect just the speed at which these new capabilities are being developed we haven't had time to kind of chew the fat here. Just scaling up appears to actually make it better in terms of you know doing what we actually want. Once you fine tune it on human feedback so you kind of align it even more there's no performance drop actually on average but now it's less likely to do bad things.