With the watchOS 27.3 beta , Apple started testing Siri Recap. ‌Siri‌ Recap lets the Series 12 and Ultra 4 Apple Watch models listen to conversations to generate a brief summary of what was discussed. Actual audio is sectioned off from apps, Apple, and the rest of the OS, and then discarded after a summary is generated. It's a privacy-forward way of tracking what's going on around you, but after two weeks of testing, it's clear that the privacy safeguards limit its usefulness. Overly brief summaries, lack of speaker labeling, and no original audio to compare against make ‌Siri‌ Recap unreliable. Before I get into my experience, I want to make it clear that ‌Siri‌ Recap is a beta feature in beta software. It hasn't been released yet, Apple hasn't finished refining it, and it won't be available to the public until watchOS 27.3 launches in December. Still, the bones are in place. Some of the issues I encountered could improve before release, but others are the result of the way it's built. My first recaps were from YouTube videos and Instagram Reels. ‌Siri‌ Recap can tell when I'm watching TV shows and movies, but it can't distinguish a YouTube video from a real-life conversation. I was surprised it summarized dialogue I wasn't involved in instead of waiting until it recognized my voice (which ‌Siri‌ can do). Listening in on ambient conversations makes it feel a little stalker-like. When ‌Siri‌ Recap is turned on, it's listening, and there's no indication for the people around you. Most of my other recaps summarized short, inconsequential conversations with my partner that I don't need or want to revisit. I don't need to remember that my breakfast on Friday had one too many pieces of vegan bacon, or that my cat was playing with a silica packet. I don't need to remember discussions about Grand Theft Auto VI or that I saw a rabbit in the driveway the other day. There is an option to only have ‌Siri‌ Recap on at certain times of the day or in certain locations, which would eliminate that issue. Ironically, ‌Siri‌ did the best at summarizing shorter interactions I don't care about. I've taken my cats to a few vet appointments recently, so I had an opportunity to test ‌Siri‌ Recap with genuinely important conversations where I get a flood of information. After the first visit, I learned ‌Siri‌ and I don't have the same opinion on the details that matter, and I found that summaries of longer conversations are just too concise and inaccurate to be helpful. I told the vet my cat is picky and will only eat one kind of dry food, but stays hydrated because she drinks a lot of water. In a separate part of the conversation, I said she was gaining weight even though her prescription food isn't for kittens. ‌Siri‌ combined those two parts of the conversation into a summary that doesn't make sense, and I had that issue in almost every long conversation. A few minutes later, I asked for a recommendation on heartworm medication because I've read about some problematic products that can make cats sick. The vet relayed a story about needing a particular brand for her cat because she has a rabbit at home, and rabbits are sensitive to a lot of medications. ‌Siri‌ Recap doesn't label speakers, so the summary of that conversation turned into a confusing amalgamation suggesting I have a rabbit at home. Inappropriately combining unrelated concepts is likely something that can be fixed, but the lack of speaker labels is intentional. During a pre-surgery consultation, I think I ran into another limitation. Apple says potentially harmful content isn't summarized, which could cover a broad range of topics. Important parts of the conversation with the vet aren't in the summary, like some of the scarier surgical risks and resuscitation preferences. ‌Siri‌ doesn't say if anything was skipped because of the content, so I don't know if that information wasn't considered important or if it was filtered out. Some of what ‌Siri‌ recapped was useful. If conversations were fact-heavy, ‌Siri‌ did a good job getting everything down, especially when dates were involved. I scheduled a follow-up appointment and ‌Siri‌ got the date right and the exact total of the bill when it was spoken aloud. Apple does say ‌Siri‌ Recap doesn't include financial info, so I'm not sure what counts. ‌Siri‌ is also surprisingly good at interpreting spoken medication names like benazepril and clopidogrel and correctly guessing what they're meant to be. During a third vet visit, the recap did get many of the major points, but going over those notes made me realize this kind of summarization may not work, even when accurate. I had my own memory of the conversation and knew what I thought was important, but some of what ‌Siri‌ kept wasn't what mattered to me. Those ‌Siri‌ details I re-read were cemented in my memory over other information when I revisited the conversation later in the week, subtly changing how I remember the appointment. I worry having notes that aren't my own will overwrite what I would have remembered without them. It's almost like having a picture of something that happened in my head, and then having someone tell me it happened differently. ‌Siri‌ Recap doesn't transcribe entire conversations for privacy reasons, but that leaves ‌Siri‌ to decide what's important enough to be written down in a discussion. I don't trust ‌Siri‌ to make that kind of decision, personally, and there's no option to check what ‌Siri‌ is leaving out since there are no raw files. I wouldn't want ‌Siri‌ Recap to record conversations in full because of the privacy concerns, but Apple's summarization alternative has shortcomings that affect its usefulness. ‌Siri‌ picks and chooses what matters, and in my two-week test, ‌Siri‌ merged concepts and conversations in ways that were more confusing than helpful over and over again. ‌Siri‌ Recap isn't a feature I'm going to use, and it's not a feature I'd recommend to others. Related Roundups: watchOS 26 , watchOS 27 Tag: Siri Related Forum: Apple Watch This article, " Siri Recap Hands-On: Private, But Not Very Useful " first appeared on MacRumors.com Discuss this article in our forums
Microsoft is abandoning all versions of its To Do app across the Mac , iOS, and more. It hopes users will move to its paid Microsoft Planner service inside the awful Teams, but it's even putting obstacles in the way of doing that. It was never the greatest task management app, but if you liked Microsoft To Do you now need to wave it goodbye. Long-time users of task management apps may remember Wunderlist, which was a hugely popular To Do app for the iPhone , Mac, and iPad . Microsoft bought it in 2015, rebranded it as Microsoft To Do back in 2017 , and has now abandoned it. That's not how Microsoft is describing it, as a support document instead refers to how the app "is retiring." But as of June 30, 2027, Microsoft To Do will cease to work on Mac, Windows, iOS , Android, and the web. Continue Reading on AppleInsider | Discuss on our Forums
TSMC, a key Apple supplier and the company that makes its A- and M-series iPhone and Mac chips, has suspended some advanced chip orders for Chinese companies following the tightening of United States semiconductor export restrictions. TSMC has had to suspend orders from Chinese companies. Image credit: TSMC Under the revised rules, TSMC can no longer provide Chinese companies with chips built using the 16nm or smaller manufacturing process. That is, unless a U.S. export license has been granted or they qualify for an exemption. This new restriction means that Chinese companies are limited to chips that use much older technology built on the 28nm or 40nm processes. As a result, they will not be able to make use of the high-bandwidth memory (HBM) employed by AI data centers. Continue Reading on AppleInsider | Discuss on our Forums
Microsoft will retire its "To Do" apps on June 30, 2027, ending support for its task manager on iPhone, iPad , Mac, Windows, Android, and the web. In its retirement announcement , Microsoft directs work and school account users to Planner. Personal Microsoft accounts cannot access Planner, including those with Microsoft 365 Personal or Family subscriptions, leaving those users to find another task management app entirely. Switching to Planner will not transfer users' existing To Do tasks and lists. Those tasks will remain available within Planner until June 30, 2027, but users will lose access to them there after that date. Microsoft plans to provide an export option before To Do retires, although it has not explained what the export will include or whether other apps will be able to import it. Mac users moving to Planner will also lose the option of a dedicated desktop app. Planner is available through a browser or Microsoft Teams, and Microsoft has no plans to release a native Mac or Windows app to replace To Do. Neither option offers the same ability to manage tasks offline as the existing desktop apps. To Do grew out of "Wunderlist," the popular task manager whose developer, 6Wunderkinder, Microsoft acquired in 2015. Wunderlist's appeal at the time was its combination of simplicity, shared lists, and synced support across platforms. Microsoft introduced To Do in April 2017 as Wunderlist's intended successor, built by the same team. Initially available as a preview for iPhone, Android, Windows, and the web, it added a My Day view that suggested tasks to focus on. Microsoft said it would keep Wunderlist running until its key features were available in the new app. A native Mac app followed in June 2019 , bringing offline access, shared lists, flagged Outlook emails, and a compact window for viewing tasks alongside other apps. Wunderlist was discontinued in May 2020, completing the earlier transition to To Do. Microsoft now plans to begin rolling out an updated My tasks experience in Planner in December 2026. Tags: Microsoft To-Do , Wunderlist This article, " Microsoft to Retire 'To Do' Apps Next Year " first appeared on MacRumors.com Discuss this article in our forums