Hi Friend,
I’ve been running Apple’s beta software this year and I keep forgetting it’s there. And that got me thinking.
This is the smoothest beta cycle in years. Apple shipped one genuinely big new thing, Siri AI, the one we’ve all been waiting on, and then spent the rest of the summer fixing bugs and sanding down rough edges instead of stacking six more headline features on top. They delivered on a prior promise instead of making three shiny new ones they couldn’t keep.
If you are listening, Apple, I think this is the way things should work going forward. Each year, solve one big problem and polish everything else.
The Mac, the iPhone, and the iPad are grown up. iOS is nearing twenty years old. These platforms don’t need seven or eight tentpole features a year to justify themselves anymore. Nobody is buying an iPhone because it added a dozen things you’ll never open. Pick one big thing. Do it really well. Spend the other eleven months making the thousand small things people actually touch every day feel solid. Make that the plan every year, not the emergency measure you reach for after a rough patch.
I’ve even got a list! Here are three nominations for next year.
Global, Reliable Device Search
The first is search that actually understands me. Most people have no idea how to find anything on their own device, and it’s not their fault. Spotlight still treats my files, my Mail, my Messages, and my Photos like separate keyword boxes that don’t talk to each other. I know the photo exists. I know I mentioned the thing in a text and then again in an email, but I have to go rummage through four apps to reassemble one memory. A local, secure, AI-powered search that actually understood what I meant and searched everywhere at once would be a massive jump for the average person. It would help people like me too, the ones who already know every keyboard shortcut. Microsoft has already gone down this road with Windows. I think Apple could do the same thing more safely and with a bit more panache. This is worth a full year of focus on its own.
A Quality AI Harness
The second is a real on-device AI harness. I’ve said before that I haven’t found a good harness for local AI models yet, and I mean that plainly. The frontier models have moved fast. Claude has Cowork, this whole scaffolding around the model that lets it actually do multi-step work for me. The small models running right on your Mac or your iPhone have nothing like that. They answer a question and stop. If Apple built the harness so on-device Apple Intelligence could take an instruction and carry it through several steps on its own, that would be the natural next chapter after shipping Siri AI.
Make Apple Home Awesome
The third is Apple Home. I’ve got a house full of HomeKit gear, and I still can’t fully trust it. Scenes drop. An automation that ran fine for a month stops firing, and I don’t find out until something doesn’t happen. A light that should turn on at sunset decides tonight is the night it won’t. I don’t want new Home features. I want a year where the whole team does nothing but make the thing I already own boringly, gloriously reliable.
I’m not predicting any of this will happen. I just wish it would. One big, useful update a year, and a long list of bug fixes like we actually got this time. I want the stuff I own to work. I want the promises kept. Give me one big new thing done right, and then leave the rest of the year to make my Mac feel like an old friend who has their act together.
That would be plenty.
Your pal,
David
p.s. This week I announced A New Home for the Field Guides The Field Guides. They have moved off Teachable and into the MacSparky Circle community, with better video playback, a real mobile app, and people to learn alongside instead of a one-way lecture. If you are having trouble accessing a Field Guide, let me know.
This Week in the MacSparky-Verse
It was a full week around the MacSparky-verse.
- Claude Cowork Made My Quarterly Review Better My friend Jim Eagar wrote up how he uses Claude Cowork for his quarterly reviews, and it’s worth your time. Cowork does the digging through ninety days of notes and calendar events; Jim still does the thinking.
- Aerial Photographer of the Year 2026 A roundup of this year’s best aerial photography made me rethink how much credit the format gets as an art form.
- Turn Any iPhone Into a Dumb Phone Assistive Access can lock any iPhone down to a handful of apps, which beats buying a separate dumb phone and losing the camera and Emergency SOS along with it.
- Talking AI Agents With Michael Whalen I joined composer Michael Whalen’s podcast to talk AI agents, not music generation, and walked through how I’ve built mine to handle the busywork of a creative business.
- Omni Group Building Toward MCP Support Ken Case confirmed the Omni apps are building first-class MCP support now that Apple’s WWDC plans for App Intents are settled.
- On Mac Power Users 858, Ian Byrd of Byrdseed.TV joined me and Stephen to talk running his video business off a Mac Studio and an iPad, and how a Hawaii blackout pushed his AI tooling onto a cloud server.
In the MacSparky Labs, it was a full week of planning, tinkering, and talking Apple. A few of the things members got this week:
- A walkthrough of Circle’s new notification controls, tier by tier, so you only hear from what you actually want
- The weekly Lab Report, covering Apple’s price hikes across Music, One, iCloud+, and iPhones in Japan
- The Q3 Quarter Planning meetup, working through focus, Arete, and where AI fits into it all
That was a single week, and the Labs run like this every week. If that sounds like your kind of place, come join us.
Read this post on macsparky.com