Using Bing Search to ground LLM responses We often get the question to build an assistant based on the content of a website. These assistants often get implemented in one of two ways: Turn-based chat assistant: user can ask a question and …

ICYMI: Using Bing Search to ground LLM responses

#ai #llm #bing #rag

blog.baeke.info/2025/01/19/u...

GitHub - gbaeke/smolagents Contribute to gbaeke/smolagents development by creating an account on GitHub.

With all the fuss around OpenAI Operator, I added Browser Use to one of my demo repos. Browser Use is similar to OpenAI Operator in that it can go to a website and interact with it in the background and perform actions.

Check it out below 👇

github.com/gbaeke/smola...

#ai #openai #operator #agents

Even though I did not like it, I finished season two of Silo yesterday.

What I liked: the style, cinematography and Rebecca Ferguson.

What I did not like: the continuous switching between the two silos and the overall story line which was kind of bland and boring

Holding my breath for S3! 😉

Using Bing Search to ground LLM responses We often get the question to build an assistant based on the content of a website. These assistants often get implemented in one of two ways: Turn-based chat assistant: user can ask a question and …

Using Bing Search to ground LLM responses: a custom solution versus an out-of-the-box Azure AI Agent approach ⬇️

blog.baeke.info/2025/01/19/u...

#al #llm #bing #grounding

I often use, or try to use, the port forwarding capabilities of VS Code. I am not sure why but for me, the tunnels it creates are just not reliable. I get much better results from ngrok. On top of that, ngrok provides a nice UI to see all the calls and inspect outputs.

#devtunnel #ngrok

I decided to drink this beer in Cluj Napoca. Of course… Heineken

#beer

🚨 New blog post: creating an AI agent with the Azure AI Agent SDK

#ai #azure #agents

blog.baeke.info/2024/12/24/c...

It's also slow but at least there was this mystery about the outside, the cleaners, etc...

I needed a simple UI for an agent created with the Azure Agent Service. I decided to try Lovable and created a chat ui with authentication and chat history support in minutes.

#ai #dev #lovable

a man in a suit and tie is sitting at a desk with his hand on his face . ALT: a man in a suit and tie is sitting at a desk with his hand on his face .

Season 2 of Silo is unbelievably boring. Total snoozefest.

Using WebRTC with the OpenAI Realtime API In October 2024, OpenAI introduced the Realtime API. It enables developers to integrate low-latency, multimodal conversational experiences into their applications. It supports both text and audio i…

I wrote about using WebRTC to work with the OpenAI Realtime API. It allows you to add advanced voice capabilities, similar to advanced voice in ChatGPT, to your own apps.

#openai #webrtc #realtime #ai

blog.baeke.info/2024/12/19/u...

GitHub - gbaeke/realtime-webrtc: OpenAI realtime audio with WebRTC OpenAI realtime audio with WebRTC. Contribute to gbaeke/realtime-webrtc development by creating an account on GitHub.

Now that the OpenAI realtime API supports WebRTC, I created a quick sample that also includes function calling.

Check it out below ⬇️

#openai #realtime #webrtc

github.com/gbaeke/realt...

a cartoon character says " that 's all just excited about these deals " in a parking lot ALT: a cartoon character says " that 's all just excited about these deals " in a parking lot

It's so weird... you go buy a car and you know more about the car than the salesperson. How is that even possible?

Using the Azure AI Inference Service If you are a generative AI developer that works with different LLMs, it can be cumbersome to make sure your code works with your LLM of choice. You might start with Azure OpenAI models and the Open…

I wrote about the Azure AI Inferencing Service and how to use it with GitHub models or from Azure AI Foundry

blog.baeke.info/2024/12/11/u...

#azure #ai #foundry #inferencing

I had a lot of fun watching this. We are building lots of applications that use genAI and LLMs in particular. When discussing the use of genAI in apps, this session should be required viewing for all key stakeholders. 😉

I predict you will get sick of the term "agents" pretty soon. If you aren't already! 😉

Yuval begrijpt toch niet al te goed hoe AI en specifiek GenAI en LLMs werken. Ik begrijp dat hij over potentiële toekomstige evoluties spreekt maar die staan heel ver af van de huidige realiteit. Boeiend interview trouwens…

I will die on this hill: private endpoints in Azure were a mistake.

It’s confusing for most people and gets worse as more services are added.

Everything should have been internal by default with the option of making it public with configurable WAF.

#azure #networking #rant

woody and buzz lightyear from toy story are looking up at something ALT: woody and buzz lightyear from toy story are looking up at something

There's a VP of Conversational AI at Microsoft? That somehow feels odd to me but then again, why not? 😉

#ai #chatbots #assistants

How Azure AI Search powers RAG in ChatGPT and global scale apps | Microsoft Community Hub You can enable your apps with the same search that enables retrieval-augmented generation (RAG) capabilities when you build Custom GPTs or attach files in...

How Azure AI Search powers RAG in ChatGPT and global scaleapps by Zachary Cavanell techcommunity.microsoft.com/discussions/MicrosoftMec...

The billing for users without a Copilot license was not clear to me. Now it is!

✨ Agents in SharePoint now generally available!

• SharePoint site is an agent
• Create specialized agents
• Share and collaborate in Teams
• Built on security 🔒
• Pay-as-you-go option

📃 See more → techcommunity.microsoft.com/blog/Microso...

#SharePoint #Copilot #Microsoft365 #MicrosoftTeams

Introducing Local emulator for Azure Service Bus | Microsoft Community Hub Azure Service Bus is a fully managed enterprise message broker offering queues and publish-subscribe topics. It decouples applications and services,...

This is somewhat drowned out by all the AI stuff @ Ignite. However, a local Azure Service Bus emulator is super useful.

#azure #servicebus #emulator

techcommunity.microsoft.com/blog/messagi...

Check out this great example of how to use a Graph Connector in a Copilot declarative agent. ⬇️

#copilot #microsoft #ai

I expect a lot of talks about agents at Ignite. A lot of this is already known:

➡️ Agents in SharePoint
➡️ Agents created from Copilot Chat or Studio
➡️ Agents created with VS Code
➡️ Autonomous agents in Copilot Studio
➡️ Lower-level agents built with Semantic Kernel

And more!

#ai #agents #microsoft

I deployed API Management Basic v2 with Azure Verified Modules and it configured two units by default. Why is that set to two?

#bicep #apim #avm

a red button that has the word easy on it ALT: a red button that has the word easy on it

I had a discussion about Azure Container Apps versus Azure Functions today. It's not the first and will not be the last.

Unless it's for lightweight event-driven workloads, we pretty much default to Azure Container Apps and in some cases Kubernetes.

Easy choice but I guess it depends on prefs...

Demon YouTube video by Sad Night Dynamite - Topic

I was in an Urban Outfitters shop last week. For my daughter, not me! 😉

Great music though. Picked up a few good tracks with Shazam. This one for instance.

#music

music.youtube.com/watch?v=sEg7...

Data Analysis for Copilot - Visual Studio Marketplace Extension for Visual Studio Code - This tool extends the LLM's capabilities by allowing it to run Python code in a sandboxed Python environment (Pyodide) for a wide range of computational tasks and da...

Looking good this... data analysis directly from GitHub Copilot!

#ai #copilot #vscode

marketplace.visualstudio.com/items?itemNa...

💡 Learn how to implement ALM in Power Platform with Azure DevOps!

Join Luise Freese as she shows how to setup your pipelines!

🔗 Watch the demo here → https://youtu.be/Lp4DSla91-I

#Microsoft365Dev #PowerPlatform #Azure #DevOps