[righthand]

Agreed there really isn’t any metrics that indicate this is true. Considering many models are still too complex to run locally. Llms are getting better for the corporations that sell access to them. Not necessarily for the people that use them.