https://3speak.tv/watch?v=taskmaster4450/sgvhjkxm
We are seeing the cost of inference computing dropping very rapidly. We have projections where ChatGPT 4.0 quality could be attained on a smartphone by the end of 2026.

In this video I discuss how this will localize AI and how it could change things. This opens up the potential for many agents to run locally while avoiding cloud processing.