I read some stuff about GPT-3, and what I noticed was that although it was doing some amazing things, there were a few tests that it did very badly on. I think

I read some stuff about GPT-3, and what I noticed was that although it was doing some amazing things, there were a few tests that it did very badly on. I think like 30%-ish? It would be interesting to read an article that focused on that, for contrast, and to give insight into what is still lacking.