Such a weird time, when the gap in the performance of GPT 3.5 and 4 is huge, but the time between their releases is so short. Some of the critique that was apt for 3.5 sounds a bit out of touch when it comes to 4.