Is Local AI Worth it?

.NET AI Community Standup: Interview Coach Evolves: MAF + GitHub Copilot SDK

We just updated the Interview Coach! Join Bruno Capuano and Justin Yoo to explore the latest evolution of Interview Coach, a .NET sample built with Microsoft Agent Framework (MAF), now with GitHub Copilot SDK support. Discover how Copilot moves beyond the code editor to power an application where specialist agents review resumes, ask behavioral and technical questions, and deliver personalized feedback. The Copilot SDK handles model interactions and tool calls, while Agent Framework coordinates the agents and their handoffs. With shared MCP tools, a Blazor interface, and continued Microsoft Foundry support, this sample offers a practical starting point for your own agent-powered applications. 🌍 Explore the source code and continue learning with the companion walkthrough and hands-on workshop: https://aka.ms/mafinterviewcoach/repo https://aka.ms/mafinterviewcoach/blog https://aka.ms/mafinterviewcoach/workshop 🔗 Links: https://learn.microsoft.com/en-us/collections/d4gkh1tee47wo?source=docs 🎙️ Featuring: Bruno Capuano https://www.linkedin.com/in/elbruno/, Justin Yoo https://www.linkedin.com/in/justinyoo/ #dotnet #MAF #GitHubCopilotSDK #Agents #AI

Scaling the Claw: Skills, Shell, CodeAct and Background Agents

Four additions make the agent genuinely useful. Skills package domain knowledge such as valuation and risk scoring into files it discovers and loads only when a request calls for them, including skills managed centrally in Foundry that update with no redeploy. Shell access lets it reorganise a messy folder of trade confirmations, confined to one working directory. CodeAct gives it a sandbox where it writes and runs code to work out an answer rather than guessing at the arithmetic. Background agents let it research several tickers at once and pull the results back together. 📌 This event is part of a series. Learn more here: https://aka.ms/FMTA/series 🔗 Learn more by exploring the resource: https://aka.ms/AgentHarness #MicrosoftReactor #LearnConnectBuild [eventID:27519]

3 unknown Claude Code Skills #claudecode #claude #aiengineering

My Claude Code Status Line (Steal It)

💎 Download my script here: https://gui-ferreira.kit.com/c6dd52421e #claude #claudecode

Beyond AI Agents: Is This the Future of Operating Systems?

What if every application on your computer could explain how it works, discover other applications, and collaborate with them to get things done? In this session, I sit down with Maxim Khailo to explore Abject and the idea of Abject-Oriented Programming. Inspired by Alan Kay, Smalltalk, and the actor model, Max is building a system where objects communicate through messages and expose an “ask” handler that lets you ask them questions in plain language. LLMs help when needed, while ordinary messages remain deterministic. Max walks us through a live demo, including a chat that discovers how to interact with a virtual fish tank and feeds the fish. We explore shared workspaces, peer-to-peer communication, local AI, and his vision of “private software”: applications people create to solve their own problems. We also discuss why coding agents can struggle with original ideas, why engineering still matters, and why people should retain control over their software and computing. Explore Abject: Website: https://abject.world/ Source code: https://github.com/mempko/abject Max website and socials: https://mempko.com/ https://github.com/mempko https://x.com/MaximKhailo CHAPTERS 00:00 Introducing Abject and Maxim Khailo 02:20 Alan Kay, Smalltalk, and messaging 07:22 Why build Abject? 09:19 The “ask” handler 12:06 What does “Abject” mean? 15:00 Objects communicating across machines 17:00 Biology and peer-to-peer design 19:00 Live demo: Inside Abject 21:00 Asking objects how they work 23:00 Workspaces, chat, and agents 25:00 Discovery and collaboration 27:00 Shared workspaces and the fish-tank demo 29:00 Inspecting LLM calls and costs 31:00 Web browsing, skills, and MCP 32:30 Max’s vision for personal computing 36:00 The age of private software 41:13 Exploring the open-source project 42:34 Building Abject with Abject 46:14 Why software engineering still matters 48:28 Decentralization and software freedom 52:53 What comes next 54:23 Running with local LLMs offline 54:44 Closing thoughts What would you build if your applications could discover each other and work together? Let us know in the comments. Subscribe for more conversations with the people building the future of technology. #Abject #AI #OpenSource

Working With Your Data, Safely: Files, Approvals and Memory

So far the agent can talk about markets but cannot touch anything of yours, and nothing stops it doing something it should not. This episode fixes both. We give it file access so it reads a real portfolio.csv and writes its reports back to disk, then put a place_trade tool behind an explicit human approval. Approving every single read gets old fast, so we add auto-approval rules for the safe calls and one custom rule that clears small trades on its own. We finish with durable memory in two forms, the files the agent curates itself and the facts Foundry extracts automatically, and watch both survive a restart. 📌 This event is part of a series. Learn more here: https://aka.ms/FMTA/series 🔗 Learn more by exploring the resource: https://aka.ms/AgentHarness #MicrosoftReactor #LearnConnectBuild [eventID:27518]