Early this evening, I asked Bard if was updated to PaLM 2, and it said it was. I then asked it to write some Python programs, giving it more or less the same pr

Early this evening, I asked Bard if was updated to PaLM 2, and it said it was. I then asked it to write some Python programs, giving it more or less the same prompts I've given GPT4. Bard doesn't seem to be any better than it was a couple weeks ago in the cases I tried, and nowhere near as capable at GPT4. And it goes off the rails quickly. After even a short dialog (~5 statements), it becomes less and less able to stay on track and make coherent corrections to the code.