Let's dive into the world of Siri AI on the Mac and explore the potential and pitfalls of this new feature. I've been playing with Siri AI for a little over 24 hours now, and it's an intriguing, if somewhat limited, experience.
The Early Days of Siri AI
I've been a long-time skeptic of Siri and Apple Intelligence, but the new Siri AI on macOS 27 Golden Gate has piqued my interest. It's still in its infancy, and I'm testing it on the developer beta, so there's a lot of room for improvement. One thing that immediately stands out is the lack of an 'indexing in progress' indicator, which leaves me wondering about the current state of my files and folders.
Finding Siri AI's Place
When it comes to laptops, my initial thought was to use Siri AI for time-consuming tasks like benchmarking. However, I quickly realized that Siri AI can't take actions inside apps, which limits its usefulness for such tasks. I then explored the potential of vibe coding Shortcuts, but even that fell short of my expectations. It seems Siri AI is more of a helper than a full-fledged assistant at this stage.
Data Analysis: A Glimpse of Potential
One area where Siri AI shows promise is data analysis. I tried using it to calculate average scores from benchmark screenshots, and it worked surprisingly well most of the time. It could distinguish between different types of scores and present the results in a clear format. However, it struggled with more complex scenarios, like mixing synthetic and time-based results. Despite these limitations, it offers a glimpse of what Siri AI could become.
The Ecosystem Factor
Siri AI's capabilities seem heavily influenced by the Apple ecosystem. It excels within Apple's apps and services, but struggles with non-Apple apps and files. When I asked Siri to find pictures of cats or babies, it only searched Apple Photos and Messages, ignoring my preferred apps like Signal and Google Photos. This ecosystem bias is a significant limitation, especially for power users who rely on diverse apps and services.
Visual Intelligence: A Work in Progress
Siri's Visual Intelligence, reminiscent of Copilot Vision, allows you to ask questions about things on your screen. I tested this by asking Siri to evaluate benchmark results in a Google Sheets spreadsheet. However, it couldn't see all the data unless it was visible onscreen, which is a major drawback. Additionally, Siri's responses were sometimes inaccurate, providing multicore data when I asked for single-core scores.
The Human Factor
One aspect that stands out is Siri's occasional display of human-like behavior. When I asked Siri to judge a photo I edited to resemble a street photographer's style, it complimented my work, which is not the intended behavior, according to Apple. On the other hand, when I asked for similar settings to a classic Garry Winogrand photo, Siri recommended keeping the exposure as is, which felt more like a missed opportunity.
Final Thoughts
Siri AI is still in its infancy, and it's clear that the experience will vary greatly depending on the device and the user's ecosystem. On an iPhone, where data is largely contained within Apple's apps, Siri AI might shine. But on a Mac, where users often bounce between various apps and ecosystems, its capabilities are more restricted. Despite these limitations, this is the most useful Siri has ever been, and it marks Apple's first real steps into the world of AI. It's an exciting development, and I can't wait to see how Siri AI evolves and adapts to the diverse needs of Mac users.
