Failing to Understand the Exponential, Again by Julian Schrittwieser "They see two consecutive model releases and don’t notice much difference in their conversations, and they conclude that AI is plateauing and scaling is over." www.julian.ac/blog/2025/09...

IOS 26 Liquid Glass going for that blank caps vibe. Great stuff /s

Azure #AppService now supports multiagent AI systems with orchestration, MCP tools, OpenAPI integration, and .NET Aspire observability—all on the Premium v4 tier. No rewrites needed, just powerful coordination across agents. Read the blog: msft.it/63328sYrij

Talked with the Claude Code team on how they build Claude Code. It feels I get a peek into the future, and I get why Dario said 6 months ago that 90% of code will be written by AI. This team works SO differently than any eng team I saw. Will share in-depth soon. One example:

Big #dotnet news, the STS (Standard Term Support) releases are now supported for 24 months! devblogs.microsoft.com/dotnet/dotne...

Azure AI Foundry weekly show + discord community

In this #MergeConflict Podcast, we explore what spec-driven development is and a new toolkit from GitHub that streamlines the approach when using AI. Listen in: msft.it/63325si375

Getting LLMs to write prompts for themselves is one of those things that I didn't really trust a year ago - but the latest generation of models (Claude 4, GPT-5 in particular) have impressed me with their promoting abilities simonwillison.net/2025/Sep/14/...

TIL: If you have privacy.resistFingerprinting true in Firefox, this will mess up Bluesky image posting :| . Your images will end up looking like this

In today's #VSLive session replay, we learn to harness the power of the cloud to scale AI workloads and discover how to bring AI workloads to your existing applications. Don't miss this opportunity to elevate your skills and knowledge in the world of AI and .NET! msft.it/63322s5uls

I'm running a 9-part livestream series in October about Python & AI, covering: 🔍 LLMs, Vector Embeddings & RAG 🖼️ Multimodal AI (text + images) 🤖 Agents and MCP 🛡️ AI Safety, Quality & Evaluation 🔗 Register @ aka.ms/PythonAI/ser... (Code-along for free with GitHub Models and GitHub Codespaces)

CLI agents + semantic search = powerful document analysis. We tested SemTools with Claude Code on 1000 Arxiv papers. Unix tools + semantic search creates capable knowledge workers. 🔍 Parse & search commands 📊 More accurate answers ⚡ CLI beats custom RAG www.llamaindex.ai/blog/semtoo...

”The rise of event-driven and AI-assisted GitOps will change how we reason about change”

Vscode native support for OAI compatible endpoints coming to stable branch this week

I made an agent to help me triage stale issues, using Langgraph + AgentInbox from @langchain.bsky.social, plus the GitHub API + Azure OpenAI gpt-5-mini. Quick demo vid @ www.youtube.com/watch?v=x-4k... Code @ github.com/pamelafox/is...

TIL: You can tell vscode to store pasted images in markdown to be stored into specific folder, without any extension: "markdown.copyFiles.destination": { "**/*": "markdown-images/${fileName}" } For example in workspace settings .vscode/settings.json

We (GitHub Next) put out a new project today - GitHub Agentic Workflows! This is natural language programming for Continuous AI on GitHub. eg "build the code, run the tests, read the coverage report, write tests", running daily, delivering test improvements in PRs githubnext.com/projects/age...

vscode + Github Copilot updates to try out: chat.todoListTool.enabled = new planning tool to track progress with task lists. chat.checkpoints.enabled = you can revert changes and restore different states of your chats/edits. chat.tools.terminal.autoApprove = more control for approved tools