In this meeting, the State Changers discussed issues and solutions for a location-based food ordering app being developed by Justin. The main challenge was dealing with semi-structured data: pulling relevant item and ingredient information from API descriptions, which were inconsistent and varied in length and detail.
They discussed using Xano to automate some of the mechanics such as looping through and breaking down data, but recognized that it was not enough for the judgment-based task of determining ingredients. For this, they suggested using OpenAI's generative machine learning capabilities to turn the semi-structured data into something more structured. This process involves formulating prompt queries to OpenAI's AI, and using the results to populate ingredient lists. A proposed workflow was then suggested, incorporating both Xano for its automated processing abilities, and OpenAI for its ability to intelligently extract and structure data. The discussion concluded with Justin electing to look further into using this combination of tools for his app. No other tools (WeWeb, FlutterFlow, Zapier, Make, Integromat, Outseta, Retool, Bubble, Adalo, AppGyver, AppSheet, Comnoco, Fastgen, Firebase, Google, OAuth, Stripe, Twilio, Airtable, DraftBit, Javascript, Typescript, React, Vue.js, JSX, HTML, CSS, lambda, serverless, ScriptTag, AI21) were mentioned in this meeting.
(Source: Office Hours 1/12 )
