Agreed 1000%. And the rate ate which OpenAI is spitting out new product tests like this I think they know very well they're hitting the limits of what the underlying model can do