The future of AI and generative code Q&A
summary
Panellists describe concrete features: NL-to-App turns an English prompt and an Excel file into a Dataverse table and a responsive app, shaving roughly 80 clicks from setup; Power Apps and Power Automate Copilots run on Azure OpenAI (public preview in North America) with some fine‑tuned Codex variants for Power FX. They measure acceptance/retention and thumbs up/down feedback, warn about hallucinations, note token limits (4K vs 32K), require massive GPU VMs today, and plan distillation and provenance checks.
AI-written summary · about
chapters (23)
- 0:00Introduction
- 2:15OpenAI models
- 6:55Questions
- 13:03Changing the role of the developer
- 14:56IP rights
- 16:25Image generation
- 17:03Open vs finetuned models
- 18:13Metaprompt
- 18:58How to run models on edge devices
- 23:40When would you like to look at using lowcode or Power Platform
- 26:02Plugins and Skills
- 26:55GitHub Copilot
- 27:51When do you get access
- 29:17Measuring AI quality
- 30:32Measuring user satisfaction
- 31:14Measuring MPS
- 31:53New service or new product
- 32:44ML for other roles
- 33:55ML for data science
- 34:55How can we leverage copilot
- 35:21How to start using copilot
- 38:30One more question
- 39:22Outro
related talks
21mRevolutionizing software creation: Harnessing the power of GenAI
45mPanel Discussion on AI for Code Generation
29mThe AI Revolution in Coding: Tools, Trends, and Tomorrow
01h 10mPractical Applications of Generative AI: How to Sprinkle a Little AI in Your App
01h 10mPractical Applications of Generative AI: How to Sprinkle a Little AI in Your App
56mPANEL: Unleashing the Future: The Impact of Gen AI on Development Productivity
20mGenAI Application Development
49mGenerative AI: Concepts, Tools, and Applications for Data Professionals