@TheStalwart Was thinking about something similar. How are you abel to differnetiate a great and an ok product when everything looks the same? Sure, eventual user adoption etc but how do you tell ex-ante?
This is inevitable. Open weight AI models cost a fraction of the US SOTA models and have crossed the rubicon in terms of ability. They are great and very capable alternatives. Pbly won't bode well for AI related earnings outlooks.
Credit spreads are super tight and default rates pretty low, but the share of distressed exchanges continues to climb.
Kick the can down the road type of approach.