It’s pretty interesting.

I see a ton of potential for testing. RPA can quickly get annoying because even simple change can break automation. The LLM ability to “reason” could really bridge the gap.

Coupled with agents who help turn specification/stories into a testing plan, I could really see automated end to end testing becoming far cheaper than it is nowadays in the near future.

That’s a very good piece of news for system reliability.