For years, Siri has been the punchline in Apple's ecosystem. Users turned it off, ignored it, and moved on with their lives. But something shifted when Apple Intelligence arrived, and now the company is making another bet that this time—really, this time—voice assistants might actually be useful on your Mac. After 24 hours with the new Siri AI in macOS 27 Golden Gate, one tech reviewer discovered something both promising and deeply revealing: Siri is genuinely smarter, but it exposed the fundamental problem nobody wants to admit about talking to your computer.

\n\n

The gap between what Siri can do and what users actually need it to do is wider on a desktop than anywhere else. On your iPhone, a voice assistant makes sense—your hands are busy, your screen is small, voice is faster. But at a Mac, you're already sitting with a keyboard and trackpad in front of you. You can Google something faster than you can describe it to Siri. You can check the weather in three seconds. The question Apple is wrestling with is genuinely difficult: what problems does a smarter voice assistant actually solve on a computer? The company showed off demos at WWDC of Siri analyzing data in files and running complex tasks, but when tested in real workflows—benchmarking laptops, automating repetitive work, analyzing screenshots—the limitations became immediately obvious. Siri can launch apps but can't control what happens inside them. Apple Intelligence's new Shortcuts feature promised to fill that gap through "vibe coding," but when asked to automate benchmark tests, it created shortcuts that either forgot the actual task or included a step that said "Wait for you to run the test." The irony is sharp: an AI assistant asking you to do the work it's supposed to automate.

\n\n

This reveals something crucial about where AI assistants actually stand in 2024. They're impressive at language, remarkable at summarization, genuinely helpful at explaining things. But they're still fundamentally brittle when it comes to real-world automation—the kind of work that would actually save professionals time. A reviewer testing Siri for data analysis in spreadsheets found it could understand the task but struggled with execution in ways that felt almost comical: the assistant understood what was being asked but couldn't quite bridge the gap between understanding and doing. And here's the thing that matters: Apple has massive advantages. It controls the entire OS. It controls the hardware. It knows exactly what apps are installed. Yet even with all that power, the limitations are still obvious after one day of real-world use. That's not a bug in Siri—that's a fundamental constraint of how AI works right now.

\n\n

For most Mac users, this probably doesn't matter yet. You're not benchmarking laptops for a living. But if you're someone who actually relies on automation—whether that's running tests, processing data, or handling repetitive tasks—the new Siri AI feels like a promise that hasn't quite landed. The early feedback from iPhone and Apple Watch users has been more positive, which actually makes sense: on those devices, voice assistants are solving a real problem (your hands are full). On a Mac, they're solving a problem that doesn't really exist. And that's the uncomfortable truth Apple's new Siri is exposing: no matter how intelligent the AI gets, a voice interface is just not the right tool for the job when you've already got a keyboard in front of you.

\n\n

Apple clearly has more work to do before this ships later this year. The indexing of files isn't even complete on the test units yet, and Siri itself admitted it was confused about what settings options actually exist. But the real question isn't whether Siri will get smarter—it almost certainly will. The question is whether Apple can figure out what problems voice assistants are actually supposed to solve on a Mac. Until then, a lot of users will probably do what they've done for years: ignore it, and reach for their keyboard instead.