Owen Barford @SQLOB1

🗓️ Weekly recap (Sep 12-Sep 19) for Copilot CLI 🚀 2 releases 🛠️ 119 improvements 119 features & enhancements in this release Top features: • Vim mode is now available to everyone for modal editing in the composer, showing current mode as you type 🧑‍💻 • Added /settings options to opt-in to context management tools for agents and subagents • Added /config command to open a sidebar configuration screen in the CLI • Added /sandbox command for managing network host allow/deny rules without replacing configured upstream proxy • Session and memory import commands now support the semantic JSONL interchange format • Added copilot instruction list and copilot LSP list commands, replacing older plugin-kind flags • Added --json output to copilot plugin list and marketplace list/browse commands • Added enable and disable commands to copilot plugin, copilot MCP, and copilot skill, replacing separate enable/disable flags • GPT-6 Astra support has been added • Managed sandbox sessions can now be disabled for the rest of the session from an approved bypass prompt 🚫 Enhancements: • /vim command or editorMode setting enables Vim modal editing with mode shown inline • Added 'concise' transcriptView mode to group tool activity into expandable work summaries • Streamer mode masks internal model names on /model, footer and diagnostics without restarting model initialization 🎭 • --add-dir now rejects non-directory and inaccessible paths uniformly before session start • Terminal color themes are now respected even with incomplete palettes • End and Ctrl+E keys move cursor to true end of wrapped lines for precise editing • Modified key chords (Ctrl+X, Alt+X, capital X) no longer trigger /tasks letter shortcuts to reduce conflicts • Scheduled prompts now show errors when they fail at runtime for clearer diagnostics • CLI automatically retries image prompts without images when providers reject image-limit requests • Authenticated model lists refresh automatically after startup auth hydration • MCP servers no longer fail loading when connected to a running IDE, improving plugin stability • copilot init now removes .github directory on exit if no instructions file was written • MCP servers maintain correct workspace source labels after folder trust changes • /permissions picker marks Allow all when --allow-all-tools, --allow-all-paths, and --allow-all-urls are all set ⚙️ • Resumed sessions now show completed reasoning as Thought instead of Thinking… • copilot --help now correctly describes --no-* options and --no-auto-update behavior • Restore keyboard input and screen output correctly when returning from external editor on macOS and Linux • Model lists refresh after signing in, switching accounts, or signing out Bug fixes: • Fixed Windows one-command sandbox bypass to approve policy-blocked writes without disabling whole session 🛠️ • Fixed failed new-session handoffs so unused clients stop polling and current session data is preserved • Prevent terminal from leaving mouse tracking mode active after session or CLI exit • Fixed exported session --share=~/notes.md to write to home directory instead of creating a folder named '~' • Fixed in-process IDE bridge rejection by config validation that caused plugin reload failures and session restrictions • Fixed scheduled prompt runtime errors showing better diagnostics • Fixed thinking shape sent to Cla… (truncated in original data but assumed fixed issue) • MCP turns continue correctly even if tool list refresh fails after tool change • Fixed restored keyboard input and screen issues after returning from external editors on macOS and Linux Misc: • Updated CLI to keep --no-auto-login and --no-sandbox options supported but hidden in help as before • Improved CLI documentation for --no-auto-update indicating it runs bundled version • Improved help text for copilot --help on --no-* option descriptions • Added semantic JSONL interchange format support for session and memory import tooling github.com/github/copilot… #GitHubCopilotCLI

FFmpeg 9.0.2 is out with bug fixes! 9.0-series highlights: animated WebP decoding, ProRes RAW VideoToolbox & APV Vulkan acceleration, Playdate encoding, new Vulkan/CUDA filters, ONNX Runtime GPU support, Dolby Vision and AMF improvements. Download: ffmpeg.org/download.html#…

Weekend learners, this one's for you. 💪 What if your report consumers could do more than browse reports? In this walkthrough, Marthe demonstrates how to build a Microsoft Fabric Data Agent. 🎥 Watch Now: youtube.com/watch?v=mFoSVc… #microsoftfabric #powerbi #guyinacube

Q: What is a trace? A: The complete record of a user session, from the first query to the final response. It includes the tool calls and intermediate steps. hamel.dev/blog/posts/eva…

I sat down with Cat a week before Hot Chips for a couple of hours to talk about life, liberty, and the pursuit of Dylan Patel. The article came out really well, congrats!

CAT @cattttang

20 hours ago

I sat down with @dylan522p and spoke with over a dozen people about his information machine. "I don't think anyone who has a traditional equity research background would even try to figure out where $3T in [the AI labs' combined] ARR comes from." Slack private messaging limits,

34 41 589 124K 731

Yet another step closer to completion. I just published the Cursor page on the #SQLServerFast Execution Plan Reference. Plus, as usual, a bunch of updates to existing pages. See the change log for the full list! sqlserverfast.com/epr/cursor/ sqlserverfast.com/epr/change-log/

Local transcription on your PC! 👌

LM Studio @lmstudio

4 days ago

Introducing local realtime voice transcription. Speak freely to your agents, without sending your voice data to the cloud. Available for Mac, Windows - and now also Linux!

50 115 1K 65K 600

🔰 The Databricks Advanced Learning Festival runs September 16 through October 14, with learning pathways across data engineering, analytics, machine learning, generative AI, Apache Spark™, and data warehousing. Complete all modules in at least one pathway during the event window to receive 50% off any Databricks Certification and 20% off a yearly Databricks Academy Labs subscription! Register and start learning: community.databricks.com/t5/learning-ev…

@burkeholland Give the WinApp CLI and WinUI Agent Plugin a whirl with WinUI 3 and build actual native apps. github.com/microsoft/WinA… github.com/microsoft/win-… can do all that and so much more... oh and it is native

AI is changing how software gets planned, built, and run. What’s next? As AI takes on more of the work, engineering and product teams are having to rethink more than just how they build software. How do you provide the right context? How do you measure AI’s impact? How do people and agents work together? And what changes when more of the lifecycle can move autonomously? That’s what I’m looking forward to exploring at State of AI SDLC on September 22, a digital summit for engineering and product leaders building with AI. The event brings together an incredible lineup of CEOs, VPs, and technical leaders across the industry on what organizations need to do to navigate the realities of building with AI. Speakers include: - @mcannonbrookes Mike Cannon-Brookes, CEO & Co-Founder, Atlassian - @rauchg Guillermo Rauch, CEO, Vercel - Liz Fong-Jones, Technical Fellow, Honeycomb - Wayne Duso, VP of Engineering, 1Password - @ElenaVerna Elena Verna, Head of Growth, Lovable - Engineering leaders from DX and Dropbox Seats are limited and you won’t want to miss this conversation. Join me and industry leaders at State of AI SDLC on September 22. Reserve your spot → fandf.co/4iUJv7t Huge thanks to @Atlassian for collaborating with me on this post.

🆕 Recent GitHub Copilot updates you should know

The AMD generational step in @Signal_65 PINNACLE. In our testing the @AMD MI355X runs a median 2x the concurrent agents of MI300X at a service level across seven models, up to 6.5x on MiniMax-M3, and it is faster at every load the two share. @AnushElangovan

Signal65 @Signal_65

23 hours ago

Enterprises running agents on AMD Instinct today, or sizing a new Instinct fleet, now have an independent number for what the MI355X buys per node. @Signal_65 PINNACLE measured both generations on the same fully loaded 8-GPU node, the same agentic load and the same service level,

0 0 13 33K 1

What if AI could help modernize integration workloads? In this episode of #AzureFriday, @shanselman sits down with Harold Campos to show how the Logic Apps Migration Agent uses agentic refactoring to transform BizTalk and legacy workloads. Watch the full episode: msft.it/6010adzhw

Here's Edge Magazine's long-form interview with many developers on the state of the game industry. Lots of quotes have circulated, so read the full thing here: knowledge.edge-online.com/p/the-game-ind…

I am increasingly frustrated about the state of conversation about AI in the UK. It feels like we are in never ending cycles of deluded thinking. Under current trends, Britain will be defenceless against the transformative effects of AI and have no leverage over development at the frontier. We are the most exposed economy to AI in the world. We have very little data centre capacity and few credible plans to expand – and our high energy costs mean we would struggle to even if we tried. I am publishing a paper today on what the UK should do to build that leverage, let the economy capture value from AI and prepare the state to adapt, however AI develops. The three strategies are as follows: 1. Build the infrastructure that gives us leverage. Fix the grid, fix planning, make Britain a place frontier AI can run. 2. Let the economy capture the value of AI. Reform the regulatory, labour and talent barriers that stop British firms adopting AI. 3. Build a state that can adapt to a changing world. Build a tax base that survives the labour market changing, public services that use AI to deliver better, and capacity to weather the upcoming geopolitical storms. Two years ago, we published a similar report on AI - many recommendations of which ended up in the action plan. I remember feeling hopeful, but sadly the fundamental problems facing the country on AI have not shifted. I truly hope that this week of momentum culminates in action. There’s still a narrow path for the UK, but everyday that path is narrowing.

keyboard_arrow_left Previous keyboard_arrow_right Next

🚀 Upgrade your Azure Data Science Virtual Machine to Ubuntu 20.04 and stay current with the latest platform improvements! 🔹 Modern Ubuntu environment 🔹 Enhanced security updates 🔹 Better performance & compatibility Start here 👉 msft.it/6016abDB6 #Azure #Ubuntu #DataScienceVM

Finally! This is great, now I can stop dropping CLAUDEmd files which just contain "@AGENTSMD"

Thariq @trq212

a day ago

We're adding support for AGENTS.md to Claude Code. Starting today in version 2.1.277, if there is no CLAUDE.md in a folder, Claude will check for and use AGENTS.md. You can toggle this behavior in /config.

2K 3K 30K 4.6M 6K

This new Claude Code mods system looks pretty interesting - more example mods here: github.com/anthropics/cla…

Thariq @trq212

a day ago

AGENTS.md support is built off of Claude Code mods, our upcoming way to customize the Claude Code harness. This is a built-in mod, but you’ll be able to build custom versions of project instructions yourself as you’d like too.

42 26 1K 299K 101

I've not been tweeting much from AI Infra Summit as I've had my head down with the keynote and panels and such, but here's my walkabout on the show floor that we're now calling The Silicon Notebook from day 2. youtu.be/31_eeq3ciRA

GitLab 19.4 is here, bringing new agentic automation built for cost efficiency at scale.. From the terminal to any MCP client, platform owners get the controls to extend it with confidence. Here are the key changes in this release 🧵: