9to5Mac is powered by Memoket: The Memoket Gem is a 0.4-ounce AI wearable that fits next to your Apple Watch. One press captures a conversation and KITE’s memory connects it to what you said weeks ago. 20 hour battery. Available now starting at $168.99 on memoket.ai and Amazon.
While it’s great that Siri’s AI can finally deliver on Apple’s promise of a truly useful personal assistant for the AI age, expectations about agent assistants are rapidly changing. But Apple is showing that it may no longer be watching from the sidelines.
Siri’s AI is great, but it’s also very 2024
Now that Siri AI (which is still officially designated beta) has been out for a little over a month, it seems fair to say that its two defining characteristics are that it’s finally useful, but it can be painfully slow at responding to even simple questions.
And while I’d spent much of the last few months mocking Silicon Valley’s obsession with AI assistants that can book flights and hotels, I hate to admit that the first time Siri AI really impressed me was when I asked how much I’d paid for a hotel room for an upcoming trip to San Francisco and how far it was from my first date in the city. Both answers were exactly correct.
Could I have looked it up myself? Safe. But putting aside the fact that I was in the middle of a 10K run at the time, finding out that information on my own would undoubtedly have taken longer than it took Siri to give me the answers.
As a result, over the past few months, I’ve found myself relying more and more on Siri’s AI for questions like “What was the contact information for that plumber who fixed the kitchen sink leak a couple of years ago?” or “What was the name of that PR person who introduced me to so-and-so?”
Information that I know I have somewhere, whether it’s an iMessage conversation or an email receipt, and that while I could just look it up, I’d now rather have Siri AI do it for me.
And for the first time in 15 years, I can ask these questions without expecting Siri to half-fail at asking them, only to get frustrated when she inevitably gets it wrong and proves me right.
All is well and good. The problem is that Apple has only now caught up with what AI assistants could do in 2024.
Apple’s next AI challenge: building an agent control system for Siri
Especially in recent months, the frontier AI market has increasingly shifted toward agent assistants, with products from OpenAI, Anthropic, Meta, and xAI (among many others) going beyond question answering to actually perform tasks on behalf of users.
I’m a ChatGPT subscriber and use Codex and ChatGPT Work extensively for day-to-day tasks involving files and information spread across my Mac and various cloud services.
And it still seems like magic every time I ask one of these tools to do something that involves opening, creating, editing, or organizing files on my computer, as well as interacting with websites and services I’m logged in to, and it actually does it.
The downside is that every time ChatGPT completes one of these tasks, the new Siri AI starts to feel like the old Siri again.
It’s still early days, sure, but it already seems like this is another area where Apple seems to be falling further and further behind, as AI labs continue to improve their capabilities, sometimes rolling out major updates just weeks apart, while Apple remains largely tied to its annual software release cycle.
But there’s an important limitation of today’s agent assistants that I think Apple is particularly well positioned to address.
As impressive as tools like Codex, Dots, and ChatGPT Work may be, they still rely heavily on the user to provide the permissions and context behind any task they are supposed to perform. They can access files, search emails, and interact with websites, but they generally don’t know what you’re doing, what you’ve discussed with other people, or what you might need next unless you explicitly tell them.
And as these tools become increasingly capable of managing our computers and phones on our behalf, I suspect that the next big leap in their usefulness will come not just from improving what they can do, but from improving their proactive understanding of user context in real time.
This is where Apple may have an advantage.
Apple recently showed that it is finally willing to leverage the trust it has spent years building around privacy, even if it means testing the limits of what users feel comfortable sharing with an AI assistant.
Last month introduced two Apple Watch features: Siri Recap, which can automatically generate summaries of conversations throughout the day, and Live Rewind, which lets users catch up on the last 15 seconds of a conversation they may have missed.
Both features rely on the Apple Watch’s microphone to capture information about what’s happening around the user, potentially giving Siri access to a whole new source of personal context.
Of course, the privacy implications are huge, especially since the conversations involve other people who haven’t necessarily consented to having their words processed by an AI assistant.
And while these concerns are certainly worth discussing at length (and are likely why these features have yet to be rolled out), Apple has repeatedly demonstrated that it is willing to go to great lengths to address the privacy challenges that come with giving technology access to sensitive personal information. Just look at the efforts put into developing its reference image system.
As assistants become more capable, I suspect that their greatest limitation will have less and less to do with what the underlying models can do, and more to do with how much we are willing to let them know about us.
Imagine an assistant that doesn’t just wait for you to ask it to find a document or complete a task, but already understands what you’re working on. One that can gather information in the background so that it is ready when you need it or even take action on your behalf based on what is happening around you.
This is the kind of assistant that could make today’s agentic tools seem as limited as the old Siri (and, increasingly, the new Siri AI) now.
And while Apple still has a lot of catching up to do when it comes to actually developing these capabilities, it is in a much better position to compete in this next phase of AI than when the race for generative AI began.
9to5Mac is powered by Memoket:
Apple just added voice recording and note-taking to iPhone and Apple Watch. Helpful, but check battery at 2pm. Dedicated recording is one of the fastest ways to wear out a Watch, and to do so your iPhone must be turned off and unlocked. More importantly, neither device links what was said in today’s meeting to a decision made three weeks ago. They capture moments. They don’t build memory.
Memoket Gem was created to do more. A 0.4-ounce wearable AI agent that fits your Apple Watch on the same wrist. One button, one press, 20 hours of battery. Gem captures conversations and links them over time using KITE, a long-term memory engine that links what was said today to what mattered weeks ago. When a topic comes back, Gem automatically surfaces the original conversation. It all comes together in ChatGPT, Claude, Slack and Notion.
153 languages supported. Available now on memoket.ai and Amazon starting at $168.99.
FTC: We use automatic affiliate links that generate income. Moreover.
