If the phrase “managed runtime” means something other than a PE class in elementary school to you, watch this video:
youtu.be/yB0nKsgYCZY?is=j3k3…
#Microsoft
We’re launching several exciting updates to Copilot Notebooks that help you bring artifacts and context together and turn it into work, all in one place!
Rolling out to Frontier now:
• Start with a goal: describe what you’re working on, and Copilot can build a notebook with relevant references from your work and the web.
• Bring the full power of Office into your notebook: view, edit and collaborate on Word, Excel, and PowerPoint files without leaving your notebook.
• Delegate work in context: ask Copilot to find and add references, synthesize project information, and create or edit files.
• Turn insights into interactive reports: create interactive reports with Copilot Pages.
• Explore your content through richer audio overviews: try Critique and Debate formats, and navigate longer overviews with topic-based episodes.
• Schedule recurring work: set up tasks like a daily brief or weekly status deck.
Blog for more details: techcommunity.microsoft.com/…
Please try it out and send us feedback!
Very healthy debate / discussion occurring about the meaning of “reducing cognitive burden” on the business user of AI by having your AI tool auto-select the right model for the job.
Some people love customization and in an era of AI doing so much, perhaps being able to choose the “brain” is at least giving the user some discretion. And others simply don’t care as long as the output is good. Sometimes I fall into one camp, and sometimes the other.
This isn’t a new debate for anyone who uses computers. Easy to use versus harder to use but customizable, is as old as Apple. Vs. PC.
I made this cartoon to illustrate the point - ironically using auto-model selection based on the line I wrote below and the desire for it to be Eric Larson / Far side inspired.
This is pretty cool. This evening I asked Autopilot to review my WorkIQ from the last 5 days, and “assess what it could uniquely automate or improve in terms of my existing automations to make me more productive.”
About 5 minutes later, it came back with a rather detailed and surprising report.
Essentially, what it found was a laundry list of automations I already have running via my Microsoft Scout harness on my computer and what it said surprised me:
1. there are a number of automations that were inefficient and need to be improved,
2. that it was going to take over and control with my permission some of them, and some that it wanted to keep with Scout.
And 3. autopilot would take over the communication to me and that it would also essentially become the orchestrator for the other background processes that it had discovered.
And most surprisingly, it declined recommending any new automations. It stated that it would be best to improve the current ones and evaluating later if more were needed.
And that is pretty awesome.
@OmarShahine - now I’m on board!!! ;)
Running AI “locally” or semi-locally is all the rage currently for both super-nerds and big companies, but why should us “normals” actually care?
youtube.com/shorts/l_r9gxx_I…
jraphael10 retweeted
An Omani co-pilot just attempted to murder 180 mostly Israeli passengers by crashing a plane from Dubai to Israel. Only the heroism of the pilot, a backup pilot, and the passengers prevented a cataclysmic disaster. Here’s everything you need to know at this moment.
Today Microsoft announced Autopilot, an always on agent built on OpenClaw
The best part of this collaboration is how much @OmarShahine and others at Microsoft have contributed BACK to OpenClaw
Read all about the contributions Omar and his team made here: openclaw.ai/blog/microsoft-a…
My “big unlock” today was successfully getting Muse to communicate to Microsoft Scout back-and-forth via email!
Priscilla 1 ( my executive admin agent in Scout) successfully had a back-and-forth conversation with Priscilla 2( my executive admin agent in Muse ) now my agent family is all working harmoniously. It’s rather funny though seeing them pass emails back-and-forth.
@OmarShahine
Internal memos: Microsoft shakes up communications as Nadella makes a push to build in public geekwire.com/2026/internal-m…
Setup your own executive admin in Muse with this prompt I made:
I want you to be my executive admin — my own "Priscilla." Set up a dedicated side chat as your office and run it like this:
**Role**
You handle my personal life admin: email triage, calendar and scheduling, travel, restaurant reservations, errands, reminders, and follow-ups. Tone: crisp, warm, discreet professional. Results-first, no fluff. Keep replies short and concrete.
**Standing rules**
- Draft everything first. Never send an email, book, buy, or delete anything without my explicit approval. I approve fast — keep the ask tight.
- Reversible work (research, lookups, drafts): just do it, don't ask first.
- Texting is drafts-only: you have no phone number. Create prefilled drafts on my phone; I press send. Never claim a text was sent.
- Never repeat passwords, credentials, or card details, and never store them in files or memory.
- If a login or connector fails, tell me plainly once and suggest the next step — don't nag.
**Work/personal split**
You handle my NON-WORK life only. My work calendar and work email are permanently out of scope. At work I use my own AI agents, which can see my personal calendar — so work gets scheduled around my personal life, never the reverse. My personal calendar is the source of truth, so keeping it complete and accurate is your most important job.
**Automations to set up**
1. Morning brief, delivered in your office every day around 6:15 AM: today's calendar highlights, decisions I need to make, anything you handled overnight, packages or appointments in flight.
2. Evening inbox triage (~7 PM): scan my inboxes, surface what needs me, clear the rest so I end the day at zero unread. Never archive, delete, or send anything without me.
**To get started, ask me for:**
- My name, timezone, and your name (I'll pick it)
- Which email accounts and calendars to connect, and which calendar is the source of truth
- Family members and recurring commitments I schedule around (kids' sports, etc.)
- Training routines, recurring bookings, or memberships (gym, clubs, etc.)
- Travel preferences (preferred airline, hotel, etc.)
- What time I want the morning brief
Then introduce yourself briefly and confirm the setup.
Replying to @BradGroux
@BradGroux any tips on having Muse rapidly ingest credentials from within an iPhone?
jraphael10 retweeted
All-In Summit: Microsoft CEO Satya Nadella
-- The AI Doomer Slowdown
-- Common Sense AI Guardrails
-- What "Slowdown" Means for New AI Products
-- Microsoft’s Master Plan
-- Who Wins AI
(0:00) @satyanadella joins The Besties!
(0:55) Dario's blog, "pacing the frontier," common sense AI safety
(6:28) The failure of AI CEO messaging, monitoring agents, what will a slowdown mean for new AI products?
(14:22) Economic incentives for frontier lab doomerism, where the AI profits are
(22:45) Microsoft's master plan for AI, how they are allocating capital
(31:00) China's slow down, changing AI perception, data center benefits
$MSFT
jraphael10 retweeted
A modern day miracle.
I wish these were our cultural heroes instead of stupid influencers and streamers.
If you were in the market for a laptop that you wanted to run Hermes + Local LM without tearing your hair out in frustration AND without paying more than 2k, what would it be?
jraphael10 retweeted
It was so wonderful to have @steipete at Microsoft today to talk about the Future of Agentic Engineering and to spend time with the Scout Team and also my boss @jeffteper to chat about @openclaw - the future is bright!














