Harnessing OpenAI to help millions
summary
Shows concrete tooling and results: uses Azure OpenAI (GA Dec 2022), DALL·E and text-davinci-003, trains on Common Crawl with DeepSpeed/ZeRO, and runs on Azure with 99.5% SLA and AAD security. Demonstrates a chat-summary app wired to Azure Cognitive Services that replaces prompt-only designs with semantic kernel skills, planner and in-memory memory to cut hallucinations (War and Peace prompt used as a failure case) and exports summaries/action items used by Teams Premium — it works in the demo.
AI-written summary · about
chapters (10)
- 0:00Introduction and background
- 1:05AI evolution and generative models
- 6:22Azure OpenAI service overview
- 13:23Semantic Kernel introduction
- 16:59Demo: Stopping hallucinations
- 27:57Conclusion and future of coding
- 29:00Q&A: Ethical AI and risks
- 31:31Q&A: Multilingual support
- 33:17Q&A: Future of programming jobs
- 36:08Q&A: OpenAI model openness
related talks
48mGetting started with generative AI using Azure OpenAI Service
20mUnlocking Azure OpenAI: A 20-Minute Introduction
59mLearn Live: Introduction to Azure OpenAI Service
31mAzure AI Innovations at Microsoft
01h 10mSolutions prototyping using Azure AI Studio
44mSupercharge your data with Azure AI Search and OpenAI
01h 01mUnlocking the Potential of AI
48mTake an Azure OpenAI chat application from PoC to enterprise-ready
