Yes, that is the complete prompt. The completion is: “symbol, name, … where industry=‘Packaged Foods’”
SELECT is appended to that completion to form the SQL that is submitted to the database.
In the hybrid approach I will experiment with attaching lexical information from C-Phrase to the prompt and present only the subset of the schema that is relevant to the question. That, I assume, should improve things.
Summary created by AI.
This is a discussion on the possibility of future plugins that provide access to private SQL databases using ChatGPT. Michael Minock explains that C-Phrase is planning on offering private SQL database access through their natural language interfaces, leveraging a five-step model involving VMs, HTTPS, REST APIs, and ChatGPT plugins. He seeks feedback on feasibility.
Bil French, Nunodonato, and Cliff Rosen propose alternative solutions that don’t rely heavily on plugins but suggest custom user interfaces. They express concerns about potentially sacrificing brand identity, user privacy, control over functionality, and becoming too reliant on ChatGPT.
Michael still believes in the potential of ChatGPT and envisages a significant user base sticking to OpenAI’s interface, using plugins to access private company databases. RonaldGRuckus doesn’t find plugins beneficial in their current state but sees their potential in the future.
Lastly, Bill French questions the tendency of users to use a universal interface like Google search or ChatGPT over individual, private access points. He has reservations about the trust and privacy implications. He concludes by stating that although plugins will be massively successful, he doesn’t believe they’ll apply to the use cases Michael envisions.
Summarized with AI on Nov 30 2023
AI used: gpt-4-32k