Microsoft
First, OpenAI limits you to certain number of queries and you can't query when there is a high demand. They also limit you in certain ways in you're running the free version. You're also using and older LLM with 3.5 instead of 4, which copilot uses.
Secondly, OpenAI is hosted in Azure, OpenAI isn't running their own servers in their basement. It's all running from Azure
Third, CoPilot is alot more resource intense than the ChatGPT Text Generation. Something like 60-100x more resources per query if not more.
So the core issue is that people don't understand the resources needed to run this, we're in a new era of commuting. It's like when Crysis came out and no graphics card could run it for years in max. This is gonna take time and Microsoft is making huge investments to speed this up. General availability has always ment that it's out of beta and that there's additional requirements isn't anything out of the ordinary.
Because Microsoft would have made astronomical more money if they could release it for everyone, the 300+ seats is something they are force to do to keep demand down, even releasing it now isn't in the best intresse due to it being in the middle of a current budget or after a new budget has been set by most companies.
This isn't a cost most companies can just throw in, it takes alot for a company to get something like this approved outside of their setbudget.
