I turned off Siri on the Mac years ago and never looked back. Similarly, I found Apple Intelligence fruitless and never interacted with it. But the new Siri AI coming to macOS 27 Golden Gate at least got me slowly rethinking things.
I’m still starting to explore Siri AI, as I’ve only had access to the macOS 27 developer beta for over 24 hours. It’s also in early dev beta testing, so there should be plenty of development routes before it’s released later this year. I don’t know if it has finished indexing my files and folders on our M5 MacBook Air and M5 Max MacBook Pro review unit. Unlike the iOS 27 dev beta, there is no “persistent index” box on the settings page. I asked Siri if she could tell me, but she told me to click the button in Settings that doesn’t exist.
Our partners got a headstart testing Siri AI on the iPhone and Apple Watch, and got a read on its familiar vibe, and so far they’ve had good feedback using it. My feelings are a bit mixed.
When I sit down at my laptop I don’t need a voice assistant to search for things I’m curious about or check the weather like I would on a phone; I can do that faster and more accurately with a keyboard and mouse. So I tried to think of ways to let Siri AI help me on macOS – things that might be useful for me in my daily work.
I’d be happy to automate some of the time-consuming balancing I do when I review laptops, but while Siri AI can launch apps, it can’t take actions within them (not that Apple ever said it could). Then I tried to see if writing with vibe a few shortcuts would get me there instead. This is not a feature of Siri AI, but a new part of Apple Intelligence. I asked Shortcuts to run a test on Geekbench or Cinebench, capture the results on a screenshot, wait a few minutes, and repeat the process two more times. But the found automation could not do the tests again. Apple Intelligence created a shortcut that opened Geekbench and took screenshots (but forgot to run the benchmark), and created a shortcut for Cinebench that had “Wait for test to run” as an actual step. Maybe if the developers keep expanding the App Objectives this might work one day.
So if Siri can’t help me run my benchmarks, maybe she can at least help me get a little faster at data entry. In my typical workflow, I run each benchmark three times, taking screenshots as I go, and later averaging the results before cataloging them in a spreadsheet. Apple’s WWDC keynote showed someone using Ask Siri in Spotlight to analyze data from local files. So I tried selecting collections of those screenshots in the Finder and asked Siri to calculate the average scores for me. It worked very well – most of the time.
It was smart enough to separate single-core CPU scores from multi-core CPU scores and GPU scores, average test results, and organize them into easy-to-read tables. But it could be thrown away if I include screenshots of too many types of tests, especially if I mix with synthetic score results (Geekbench, PugetBench, etc.) and time-based results (Blender offers tests and our 4K video export test). And sometimes it is thrown off by the CPU level data seen in the Cinebench screenshots. Ideally, I’ll be able to have Siri AI accurately calculate 15 or more averages from a bunch of my screenshots all at once — that could save some serious time. But for now, it can only help me a little. And unless it gets better I’m still willing to keep doing it myself, especially since Siri messed up the numbers a few times by pulling the wrong data.
1/6
So far, Siri AI seems to be more powerful within the Apple ecosystem than outside of it, even with apps and files that are already on my Mac but from non-Apple apps. When I asked Siri to find my photos of cats or babies, it pulled results from Apple’s Photos and Messages apps. This may be enough for most people, but not for me. Most of my messaging is done through Signal, and photos from my phone are uploaded to Google Photos, not iCloud. Siri also missed thousands of photos I have in my Lightroom Classic catalog, or files stored locally in the photos folder and I kept asking it to access them directly. It’s possible that those files haven’t been identified yet, but I have no way of telling.
At the moment, I’m getting the same vibes as when I tested the Copilot Vision last year. Like Copilot Vision, you can use Siri’s Visual Intelligence to ask questions about things on your screen. And like Copilot, it’s limited. I asked Siri to check the benchmark results in a spreadsheet in Google Sheets, but it can’t see all the data if it’s not visible on the screen at once. I was able to see the entire spreadsheet by downloading it as an Excel file and pointing Siri at it in the Finder, but when I asked for a laptop with a single-core Geekbench score it gave me multicore data. Not good.
I fired up Siri while using Lightroom Classic, on a black and white photo from my Ricoh GR IV Monochrome review, and asked Siri how to make it look like a shot from street photographer Alan Schaller. Siri offered precise value adjustments for exposure, contrast, and more, and adjusting those values gave me a good result. Unfortunately, when I asked Siri to judge the result, it gave me a bad impression, saying I would look at it and get a “permanent feeling,” which is the kind of behavior Apple says it shouldn’t display. (I thought we were past this.)
Then I uploaded an old photo of Garry Winograd and asked how I could change my Lightroom settings to match that photo; Siri recommended that I set the exposure to the value it had been at. So, some hits and some misses.
It’s still very early days for Siri AI, and a lot can change between now and the final release. But what is already clear is that the experience may be very different on the iPhone, where most of your data is stored within Apple’s applications, and on the Mac, where you may be riding between all kinds of applications and ecosystems in ways that limit what Siri can do. Anyway, less praise, but this is still a very useful and useful Siri. Apple’s first real AI baby steps.
