I just found this gem... ini ;-) In the world of self-hosted LLMs, we often worry about the context size and the limitations with RAG. This, practically remove those limitations. I can't straight up attach a 50GB db to an LLM to do context searches, but this method removes that need. Context is reduced interpreting my data request - which LLMs do very well - and converting that request into a query for the backend data. In other words"from the million widget entries we have in our db, how many of those are blue?". The LLM then creates a query and steps aside the big data flow. This is especially well suited to SQL, but it can also be used in RAG applications for web searches, etc. Well done guys!
Hey Folks! I'm excited about this application of AI, because it can lower the barrier for non-programming folks using programming interfaces, and avoid programmers always having to be in the loop to extract data from business systems. I hope you build something great with this AI application pattern!
Hello, Manikantah this side do you have 5 mins to chat about implementation of a AI chat bot we are trying to build using Google cloud services. Essentially we want to ground responses to enterprise data and also answer questions about live CRM data in the same flow. We have access to every tool in GCP but we have no idea how to implement. We have contacted google support many times but they never give a relevant response. Thank you,
Why the advanced Gemini costly? And Gemini Nano is overpriced? Is it because of genuine better generative AI search results or output?❤❤❤ From India. Thanks for the update about GEMINI API
yes it is possible. You can set up an AI agent on your BQ instances or make an API call from your database wrapped in a code base. Ground LLM on your database/table ....and that will allow Gemini to respond while grounded on your database/data. Hope this helps. This is the future..
@@spixe22the corporate AI models are aligned too much to not make any serious political opinions. Open AI & Gemini models are absolutely useless in this regard. Their answers are too balanced or biased to not make powerful entities unhappy.
Register to save your spot for Google I/O✨➡ goo.gle/io24-yt
I just found this gem... ini ;-) In the world of self-hosted LLMs, we often worry about the context size and the limitations with RAG. This, practically remove those limitations. I can't straight up attach a 50GB db to an LLM to do context searches, but this method removes that need. Context is reduced interpreting my data request - which LLMs do very well - and converting that request into a query for the backend data.
In other words"from the million widget entries we have in our db, how many of those are blue?". The LLM then creates a query and steps aside the big data flow. This is especially well suited to SQL, but it can also be used in RAG applications for web searches, etc. Well done guys!
Hey Folks! I'm excited about this application of AI, because it can lower the barrier for non-programming folks using programming interfaces, and avoid programmers always having to be in the loop to extract data from business systems. I hope you build something great with this AI application pattern!
Hello, Manikantah this side do you have 5 mins to chat about implementation of a AI chat bot we are trying to build using Google cloud services.
Essentially we want to ground responses to enterprise data and also answer questions about live CRM data in the same flow. We have access to every tool in GCP but we have no idea how to implement.
We have contacted google support many times but they never give a relevant response.
Thank you,
@@devatalamanikantah4449 @devatalamanikantah4449 you can contact us we have built the same application
Frankly I’m overwhelmed with the ideas.
I'm overwhelmed too! Gave the transcript of the video to my ai assist. They keep track of all my important ai stuff's. 😅
This is where it is going, automation remains the paradigm, but closing the human-machine interface gap even more than previous coding frameworks
I am just a beginner and I am thinking of using it for my library management system
Curious if the data needs to be described beforehand? Or does Gemini extrapolate the meaning of attributes from the data?
Why the advanced Gemini costly? And Gemini Nano is overpriced? Is it because of genuine better generative AI search results or output?❤❤❤ From India. Thanks for the update about GEMINI API
Very helpful! Thank You! 😎🤖
Is it possible Gemini respond with an answer from the database?
yes it is possible. You can set up an AI agent on your BQ instances or make an API call from your database wrapped in a code base. Ground LLM on your database/table ....and that will allow Gemini to respond while grounded on your database/data. Hope this helps. This is the future..
Great, thanks. Looks pretty good.
Very good question 👌
I am interested but I don't no English please tamil language expected
Veri good question❤
Interesting
What are your thoughts on this ?
I appreciate you
Franklly, after almost a year I don't have access to the API without any reasoning, google ways of things - years long beta
Jarvis?
Okok AI is cool and everything but can we talk about the political and economic state of the world?!
Ask AI 😅
Jaden Smith moment
@@spixe22the corporate AI models are aligned too much to not make any serious political opinions. Open AI & Gemini models are absolutely useless in this regard.
Their answers are too balanced or biased to not make powerful entities unhappy.
Which part
Watch another video about what u wanna know
Lol
Why'd you ask everyone at the end of every video that same question "do you think this counts as AI development"?? That's so cringe and annoying.
free?
Very helpful!