Thank a lot for sharing this port. I really like that you found a simple use case so that we can follow.
I'm interested in MCP starter kit! I would love to see if you could share your github link for this type of projects - I understand better when I run it and do some tweaks :D
Questions:
- I have been hesitant using MCP because it is built by anthropic and I saw some alternatives coming up. I'd prefer if these companies sit together and actually agree on a "protocol". What is your take on this? Would you advice your teams to start using this protocol?
- What is your take on knowledge graphs? In your example, you are only building the context from a single user but we know that it is common to get suggestions from people who have similar interests. Would you use knowledge graph to make these connections or ML models? When why? Maybe this is itself a post but I'm very confused about this data source and I appreciate if you share some thoughts.
Very impressive stuff. It was pretty helpful. Yes, in the detailed steps, I briefly skimmed. But, I think I am a bit better off after reading this one, than I was better reading this. Thanks for doing the hardwork.
I would have also loved the example being a hotel search and booking assistant over a shopping assistant ;). I am a loyal booking.com user and while making a booking a few weeks ago, totally felt the need and that the timing is here, that we should be getting to a point where I can search in long description - "I am a looking for highly kids friendly hotel, X* or above, that is also sea facing and is rated 8.5+ but also anywhere in the range of <Xkm> from <City> and ensure they are sorted by price and if there are any unique offers or stand outs, you can call them out to prioritize".. ;)
In Booking, there are people working on this (one of them is a dear (ex-)colleague Natalie Halimi). I don't know when it will be available in different countries.
Thank a lot for sharing this port. I really like that you found a simple use case so that we can follow.
I'm interested in MCP starter kit! I would love to see if you could share your github link for this type of projects - I understand better when I run it and do some tweaks :D
Questions:
- I have been hesitant using MCP because it is built by anthropic and I saw some alternatives coming up. I'd prefer if these companies sit together and actually agree on a "protocol". What is your take on this? Would you advice your teams to start using this protocol?
- What is your take on knowledge graphs? In your example, you are only building the context from a single user but we know that it is common to get suggestions from people who have similar interests. Would you use knowledge graph to make these connections or ML models? When why? Maybe this is itself a post but I'm very confused about this data source and I appreciate if you share some thoughts.
Very impressive stuff. It was pretty helpful. Yes, in the detailed steps, I briefly skimmed. But, I think I am a bit better off after reading this one, than I was better reading this. Thanks for doing the hardwork.
I would have also loved the example being a hotel search and booking assistant over a shopping assistant ;). I am a loyal booking.com user and while making a booking a few weeks ago, totally felt the need and that the timing is here, that we should be getting to a point where I can search in long description - "I am a looking for highly kids friendly hotel, X* or above, that is also sea facing and is rated 8.5+ but also anywhere in the range of <Xkm> from <City> and ensure they are sorted by price and if there are any unique offers or stand outs, you can call them out to prioritize".. ;)
In Booking, there are people working on this (one of them is a dear (ex-)colleague Natalie Halimi). I don't know when it will be available in different countries.
https://openai.com/index/booking-com/
Yes, I surely expect this to be worked upon. I just can't wait ;)