Open lobinhojr opened 8 months ago
I don't think SK will get popular in any case.
The major problem is that SK is a Azure-strong-binding framework. They even don't have plan to support AWS at all. After checking the resume of principle product manager of SK, I figure out the reason. The whole team used to work for Azure machine learning. I don't think they wanna do something out of the box of Azure. This will become a critical failure reason. Let's see
Take a look at this comment from the product manager and you will understand what's political correctness in Microsoft. They don't wanna touch AWS thing but they wanna the community to contribute.
I take a screenshot in case he will remove it. :p
These comparisons are way out of context.
We cannot compare a stack that solves everything within the same framework with another framework that uses pluggable pieces of Azure services. For example: Azure's OpenAi service supports .md, .txt, .html and .pdf types in its semantic searches. The semantic kernel would only integrate with these services. (which makes the comparison a bit
I suggest taking a look at: