What’s Your Favorite AI?

It sounds like a silly question: “do you have a favorite AI?”

After all, they’re not like socks or a book.

How do you even know when you’re dealing with one?

I know what a sock looks like so I definitely know when I see one.

In pairs.

Hiding behind my dirty clothes.

Maybe exposed on the street, wet and dirty after furtively attempting to leave the nest?

But all very clearly socks.

Something like ChatGPT – all the rage right now – is definitely AI.

But so is Google Search and the Facebook News Feed and people don’t think of them in the same way.

What about the focus feature on your phone or the way it tries to adjust the lighting optimally for you? “Computational photography” aka AI on your phone.

Despite this, I have one AI that I have fallen in love with.

And it’s conveniently also on my phone: iOS 16 has this fun feature where it can show you photos of “people”, “nature” or “cities” on the Lock Screen of your phone. But it’s not just any photos of people/nature/cities – it’s going to find them within your photos.

I’ve had the feature on for a few weeks now (just cities and nature; people would be too distracting for me) and it is making me fall back in love with my phone.

I’ve been bothered for years with the “recall” issue. That is, the mismatch between what you know and when you can remember it.

I’ve two failed startups in the space and I’m a digital hoarder, having taken over 30,000 photos since 2000. It’s becoming a hobby/art project.

So imagine my joy when I learned that my phone could pop random images that I’d taken on an hourly basis.

What’s more, there’s novelty in the aspect ratio of these photos. The vertical iPhone screen is not a standard dimension but Apple has built an AI that, amongst other capabilities, recognizes “tall skinny things” and makes them the center of the composition.

A few times I’ve looked at the screen and thought “yeah – that’s the photo I was trying to take” – but it was lost in the traditional 5×7 or similar aspect ratio.

The effect is definitely enhanced by the vibrancy and brightness of a backlit, high pixel density screen. There are just over 3 million pixels and about 460 pixels per inch on this screen. For comparison, a typical digital photo frame is running about 260 pixels per inch right now so this is vivid color.

But this is about the AI, not the phone!

Although the phone does get credit for the AI “discovering” my photos without having to do it in the cloud. It is refreshing to know that all my photos don’t have to go to an Apple server to be analyzed. Despite them knowing almost everything about my life this feels like a little hope to cling to.


My favorite AI.

It’s a thin sliver of intelligence that can find interesting vertical patterns in photos. And it’s making my life tangibly better by reminding me of all the great adventures I’ve had in the world and what a beautiful place it is.

Cheers to more of this! People used to talk about “centaurs” that were computer/human tandems, although the term seems to have fallen by the wayside recently.

This is the first tangible time I feel like I have experienced it.

I take decent photos. The AI makes them better – not by adjusting color/contrast/etc – but by finding the focus of the photos and only selecting the ones that are best suited for this new vertically-oriented medium.

My only regret is that we’re not coupled tightly enough. I would love the ability to offer feedback. I’ve cherry-picked some of my favorite images and taken photos of them but there are a few stinkers and I haven’t included them here.

I wish there was a way to indicate this to the AI so that it could learn and get better.

But I’m sure we’ll get there. And it’s amazing that technology that was cutting edge a few years ago is now quietly toiling away on my phone, making my life better in a way that doesn’t show up in any measurable statistics other than my blogging about it (and kudos to Apple for making this happen in a way that doesn’t kill my battery…).

So, yeah: I have a favorite AI. And I hope you can find one, too.

Edit 3/28/23 – Here are a few new favorites that recently popped up