diff --git a/assets/uploads/dan-meyers-9frtwo5i3gc-unsplash.jpg b/assets/uploads/dan-meyers-9frtwo5i3gc-unsplash.jpg
new file mode 100644
index 00000000..243dbd34
Binary files /dev/null and b/assets/uploads/dan-meyers-9frtwo5i3gc-unsplash.jpg differ
diff --git a/blog/five-differentiators-make-juji-the-most-unique-conversational-ai-platform.md b/blog/five-differentiators-make-juji-the-most-unique-conversational-ai-platform.md
new file mode 100644
index 00000000..80468dd4
--- /dev/null
+++ b/blog/five-differentiators-make-juji-the-most-unique-conversational-ai-platform.md
@@ -0,0 +1,30 @@
+---
+layout: blog/article.njk
+featured: false
+title: Five Differentiators Make Juji the Most Unique Conversational AI Platform
+date: 2021-09-01T04:54:23.854Z
+author: Huahai Yang
+category: Viewpoints
+image: /assets/uploads/dan-meyers-9frtwo5i3gc-unsplash.jpg
+---
+it seems to me that they are trying everything they can read from papers that never stop to think of their own ideas
+6:42
+slot, entity, rules, transformers, embedding, all disparate techniques that they just provide, the users are supposed to figure it out what to do with these
+6:43
+their problem is that they do not have a coherent, well thought-out system that package everything together into a user friendly package, i don’t think anyone can be successful using theirs to create something useful
+6:44
+it’s just a bunch of academic ideas, now, you have to configure your self, and set up all kinds of parameters, 0.3 here , 0.5 there, nobody can do these successfully.
+6:45
+i think your experience with CLIPS should give your authority on why these kinds of systems do not work
+6:46
+they tried this and that, and deprecated features all the time, we deprecated nothing, everything we build are well thought out and tie together
+6:49
+that’s the distinction I can see. we can implements all kind of ideas too, but they fit together, and users do not have to understand any jargon to be able to use it. i.e. we are a business platform, rasa is an open source tool kit. they just provide a box of tools, it’s the user’s responsibility to figure out how to use them.
+6:51
+their so called “policy” are just different technology, rule policy are rules, TED policy is just their transformer ML stuff, and they deprecated a lot of others that they tried and failed, I am sure these current ones will be deprecated someday. They are just providing tools, and they may not work.
+6:54
+Our “policy” is higher concept, not about technology, but about conversation context, choosing which topics to try, it’s about semantics, not about which technology to use. All technology are integrated in our framework as the same thing: topics, rules and attributes. It’s coherent and consistent.
+6:59
+“symbolics as the bones, data driven as the flesh”, we have our principled approach, they do not, We have solutions, they have tools.
+7:03
+i think “no code” is a good sales pitch, as it seems to be fashionable these days.
\ No newline at end of file