Man. This is tragic. The last vestige of something that so many people put some amazing work into at Wunderlist is fading off.
Most if the magic was gone a long time ago, but this is like the last nail in the coffin.
BREAKING: Microsoft is killing Microsoft To Do on June 30, 2027, and if you've been using it for years, you'll need to find another way to manage your tasks!
To Do is being retired across Windows, Mac, iOS, Android, and the web, affecting both personal and work accounts.
For work and school users, Microsoft wants everyone to move to Planner, which is getting a redesigned My Tasks experience starting December 2026.
For consumers with a personal account, you'll need to find another app, as Planner isn't available for personal accounts, so Microsoft is effectively asking you to find another task-management app.
To make matters worse, Microsoft isn't planning native Planner replacements for To Do's Windows and Mac desktop apps, and its web and Teams experiences don't offer equivalent offline functionality.
The thing I like about building MPC servers is that the customers for them (agents) are really good about providing feedback and even ideas for improvements.
1
3
316
“It’s OK. We’ll be fine if Nvidia stalls out without Jensen.”
Working with @AMD, our models identified up to 14% PPA improvement on flagship GPU IP in less than five months—building on designs AMD’s engineers had already spent years optimizing.
The same approach helped improve Jalapeño, our first custom inference chip, with gains verified on real silicon.
New evaluations also show our models getting better at certain chip design tasks.
We’re sharing these results and how we’re working with @Synopsys to bring our model capabilities and their trusted EDA design and validation tools to more chip designers.
openai.com/index/advancing-c…
240
Google elemental model naming logic: H < C < Ne in atomic number, but C < Ne < H in speed and H < Ne < C is price. In Artificial Analysis Index, H < C < Ne.
Maybe it’s just that Neon is a 10 in terms of both atomic count and the one that you should use? 🤷 😂
242
Does the current bubble start to pop now? Maybe. Who knows for sure. Does that mean the AI ride is over? Absolutely not. This roller coaster is just getting started. Hang on to your hats.
4
301
Genius comfortably coexists with folly and hubris. You can admire what makes somebody great and despise how they’ve fucked up.
1
277
My create/review pattern has leveraged GPT against Claude (or vice versa), until now. However, I’m finding that Fable 5.2 generated code doesn’t get as much good review commentary from Sol 6.1. So, I’ve been trending to having Opus review Fable, and vice versa.
2
4
411
Very anecdotal: In my code base, it’s feeling like Sol 6.1 is a much better reviewer of the code that Fable 5.2 writes than Astra is. And Opus seems to be doing a good job reviewing Fable-written code as well.
1
338
Based on platform.claude.com/docs/en/… I just prompted an audit of my codebase saying Claude should use my current Linear tasks as intent, and finished: “Imagine that this is now your codebase and you're responsible for it. What would you do?”
🤯 🤯 🤯
1
3
532
I’m a huge fan of the midterm results. I’m also very pro AI and pro data center. I read yesterday’s pushback on those points as a concern for how we build the future — and very much a result of how we’re doing it right now. All part of the discourse that needs to happen next.
1
3
394
New Xcode package format option (.aea). Apparently, these have already been used by software updates. Much faster to decompress.
274
What I hope happens next is that Dems don’t try to claw back the past, but actually work fully towards the future. And that the Republicans also reinvent themselves for the future. A future where the argument isn’t whether to govern for the people, but how.
7
391











