🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Excited to represent the AutoGen team at this year's Research Forum! Sign up for my talk below! This is joint work with many incredible colleagues especially Eric Zhu Jack Gerrits Victor Dibia Adam Fourney Hussein Mozannar Saleema Amershi Ahmed Awadallah Rafah Hosn bigmiao and many

This is how you create a browser use agent in AutoGen. See our README (github.com/microsoft/auto…). We had browser use this time last year, it's wild that it took this long to take off. Follow us to see what we are cooking next: AutoGen

The recent VSCode blog on Agent Mode (Preview) reveals its architecture design. It's a single agent with tools and a model. Read: code.visualstudio.com/blogs/2025/02/…

Debugging your multi-agent workflows using checkpoint and rollback.

Q: ChatGPT why do you hate "Attention is all you need"? A: Because it's been cited so many times lol

Almost three months has passed since we officially released AutoGen v0.4 and 6 months since the preview release. We have made many new friends: - 1012 PRs since the v0.4 preview ♥️ - 35 Contributors 👯 👯‍♀️ - 1060 Issues closed 🙌 AutoGen welcomes your feedback and

The Belief State Transformer edwardshu.com/bst-website/ is at ICLR this week. The BST objective efficiently creates compact belief states: summaries of the past sufficient for all future predictions. See the short talk: microsoft.com/en-us/research… and mgostIH for further discussion.

Tools form the environment for agents, and workbench makes it easier to create stateful environments. It’s a simple but effective abstraction.

Introducing Phi-4-reasoning, adding reasoning models to the Phi family of SLMs. The model is trained with both supervised finetuning (using a carefully curated dataset of reasoning demonstration) and Reinforcement Learning. 📌Competitive results on reasoning benchmarks with

RL works!

Building apps is the best way to build a framework.

✨ Human-in-the-loop, by design. Magentic-UI introduces four core HCI features to make human–agent collaboration more transparent, controllable, and adaptive: 🧑‍🤝‍🧑 Co-Planning 🤝 Co-Tasking 🛡️ Action Guards 🧠 Plan Learning Let’s dive into each 👇

Respect