WTH has happened to limits and basic usability?
I have been a pro member for about 7 months, and in the time, I have seen my limits go from "I can snap out some projects this week" to "WTF, I literally started 5 minutes ago".
I have been using sonnet 4.6 on medium for most of my tasks and it seems no matter what I do, my usage is gone instantly. All I mainly do it clean up my HA instance and write automations. I have had it do a couple of other things, but nothing crazy.
This week I switched to Sonnet 5 to see if it really made a difference. Token usage is much higher, but 4.6 was constantly wrong and getting stuck in loops. Although Sonnet 5 used 2 full sessions of usage getting stuck fixing my cameras on HA because it literally tried the same thing 6 times, even though its rules say try twice then stop. It has ignored guides I have given it and found ideas from 2017 and used those instead, ect.
First 4 months were great, but not.... its not very useful. Surprisingly, Gemini and ChatGPT free have done more correct work than my paid subscription which is making me reconsider paying for it.
Anyone else feeling this?