10 takeaways from Microsoft Build 2023
Microsoft Build is an annual conference event held by Microsoft, aimed at software engineers and web developers using Windows, Microsoft Azure, and other Microsoft technologies. The conference is a platform for Microsoft to showcase its latest platform improvements for developers, and it is heavily focused on artificial intelligence and its integration across Microsoft's products and services

- Use GPT-4
- Use prompts with detailed task context, relevant information, instructions
- Retrieve and add any relevant context or information to the prompt
- Experiment with few-shot examples that are 1) relevant to the test case, 2) diverse (if appropriate)
- Experiment with tools/plugins to offload tasks difficult for LLMs (calculator, code execution)
- If you feel confident that you maxed out prompting, consider SFT data collection + finetuning
- Use in low-stakes applications, combine with human oversight
- Source of inspiration, suggestions
- Copilots over autonomous agents
- Helper: CoPilots are extensions that you can create alongside applications (ie GitHub CoPilot)
- Full Stack AI application (Jasper)
- Agents: Help orchestrate across apps and tasks (ChatGPT plugins)
- ChatGPT will be integrated into Windows to provide AI assistance across the OS as well as act as native ChatGPT chatbot
- ChatGPT plugins with be extended to Bing search and into Windows
1. Prompt Engineering Best Practices
2. Prompt Engineering: Best prompt for Reasoning
Let's work this out in a step by step way to be sure we have the right answer.
3. Podcast Social Media Sharing Automation
The Microsoft CTO hosts a podcast and explains an automation/orchestration process he uses to generate a tweet for this podcast. I found really interesting.
https://github.com/microsoft/PodcastCoPilot
@paolo wonder if this workflow could be useful for @JamesAltucher podcast
4. Inner workings of ChatGPT
The training process consists of several stages: pretraining, supervised fine-tuning, reward modelling, and reinforcement learning.
5. ChatGPT Deployment
6. Three type of AI Business Models
7. Windows CoPiliot
8. Windows Application Kill
Small but much-requested feature in Windows: You can kill an application directly from the taskbar rather than going into the task manager.
9. Microsoft Fabric
Microsoft Fabric is an end-to-end analytics solution with full-service capabilities including data movement, data lakes, data engineering, data integration, data science, real-time analytics, and business intelligence—all backed by a shared platform providing robust data security, governance, and compliance.
10. Azure AI Studio
Azure AI Studio, is a new platform for building AI applications. A new capability within the Azure OpenAI Service that lets customers combine a model like OpenAI's ChatGPT or GPT-4 with their own data — whether text or images — and build a chat assistant
It's a competitor program to Google Vertex.
No comments.