And here we are, agonising over a generic CRM system or investing in a telco industry one. Interoperable context. Databricks customer here, Brain whirring around.
Great content, as always! Just a quick off-topic question: My OKX wallet holds some USDT, and I have the seed phrase. (behave today finger ski upon boy assault summer exhaust beauty stereo over). What's the best way to send them to Binance?
Hi, great video. One question: how can we be sure the vector space we develop for RAG has any relation to the vector space of the LLM? Don’t the tokens have to be encoded using the same process to make a coherent “space”? So without ‘back translation “ to strings, this is of no help for fine tuning or generalized local integration into the LLM, yes?
Ok, so this was a silly question. I clearly need to learn more about how to leverage LLMS to design KGs based on existing relational data!!! Any suggestions from anyone on this path would be greatly appreciated!
I thoroughly enjoyed the episode and would like to express my gratitude to Maddy for agreeing to be a guest. The information presented was highly informative.
IT literature and systems are built and branded around "least privilege," so how do you verbally reconcile the idea of data access liberation with the Security and Sysadmin?
Data as a Product - One primitive way we have done this is by developing a BOT on top of our data that comes with a bunch of predefined questions where users come and ask those questions - the answers to that rendered by the BOT via an API call. But I really loved this idea of developing a capability, sort of an NLP on top of your data to be able to ask natural questions
That's so true Karen. It seems so obvious but there is no standardization for names. In some countries there is a concept of First Name and Last Name, but Last Name isn't a mandatory field whereas in other countries Last Name is mandatory and First Name is optional. This causes a lot of problems.
