close
close

Pixel screenshots may be Google’s clearest evidence that AI is useful

Pixel screenshots may be Google’s clearest evidence that AI is useful

It’s easy to be overwhelmed by the sheer volume of announcements in the AI ​​space, and it’s a huge challenge to separate the practical from the noise.




While most of the attention has been devoted to larger projects – naturalized conversations with bubbly, courteous voices or image generation that borders on the uncanny – it’s the small projects that interest me most. At the Made By Google event this week, the company announced an app called Pixel screenshots, an app that runs on the Pixel 9 series. The technology is based on artificial intelligence, but in the rare way that combines precision with ambition.

The problem it tries to solve is as old as time: to bring order to chaos; to extract the useful from the digital garbage we accumulate every day. I am the monster who has tried and failed for 15 years to build an organizational system that works across projects, platforms, and interests. I have cobbled together a crude and fragmented semblance of meaning from Dropbox, Google Notes, Todoist, Airtable, Google Photos, and a handful of other apps and services that keep me, if not organized, then at least running on a parallel path.


A GIF showing the Pixel Screenshots app in action.

Source: Google

But I’ve realized that these tools are only as useful as the effort you put into maintaining them, and the more chaotic my life becomes, the less time I have to take care of this little piece of land. I want more push and less pull; I want someone or something to do the value creation.

Pixel Screenshots has a simple premise: you probably take a lot of screenshots on your phone, many of which are taken to remember something trivial or important, but fundamentally valuable. Until now, it was up to you to remember to do something with those images. Google Photos doesn’t even back up your screenshots by default, as they are often considered unnecessary data.


The app captures all these images, runs it through a local AI filter (in this case, the on-device large language model that’s so important to Google’s continued success with the Pixel line) and extracts what it thinks are most important: text, data, and most importantly, what it thinks you should remember about it. It then bundles these insights into collections that you can manually add or filter, and surfaces all the relevant data in Google Calendar for easy analysis.

Pixel Screenshots app from Google 1


I realized pretty quickly that this app would quickly become a central part of my online browsing habits, after realizing that for at least half a decade I’ve been using Google’s own “Favorites” feature in Chrome to manually manage my own database of web snippets, particularly recipes that I visit again and again. What started as a simple way to quickly snap up recipes that my picky kids would eat has essentially become my most-used database of insights into the subconscious ways my brain tries to structure my life. Sometimes I feel so close to the bliss of order that I can taste it, yet its erratic nature is regularly (and necessarily) reset.

I realized pretty quickly that this app would quickly become central to organizing the chaos of my online browsing habits


I had a long conversation with Simon Walker, one of the masterminds behind the Pixel Screenshots app, at Google’s launch event in California this week. After the event, I’m even more certain that the future of AI lies not in massive moon-landing projects that push us uncomfortably toward AGI, but in small, focused projects inspired by a single person’s intractable problem.

I don’t necessarily see the potential of Pixel Screenshots in its current form, limited as that may be, but in its potential as a kind of passive connective tissue that permeates digital life, an evolved version of what companies like Dropbox and Evernote tried to achieve more than a decade ago. While those companies emerged in the desktop era and built their businesses on passive data redundancy, on the mobile side it’s really only platform owners Apple and Google that are capable of offering the same experience. Perhaps the end result will be an API or a feature built into Chrome. Either way, it’s AI’s ability to draw harmony from chaos that interests me most.


I would also be remiss if I did not acknowledge that this platform dominance increased regulatory and antitrust control for Google and Apple, and therefore the Justice Department could recommend Android and Chrome are being sold from its parent company following a landmark court victory earlier this month.

A pair of hands holding a Google Pixel 9 with illuminated screen

But Google is actually in the best position to provide this kind of connective tissue across the internet, and that’s why Gemini, overhyped as it may be, is finally bearing small but significant fruit. Mobile operating systems—even Android, as open as it is—don’t encourage the same manipulation of file systems and directory hierarchies as their desktop counterparts, which is why services like Google Photos have become inextricably linked to our online identities. Google Photos in particular has evolved far beyond its original mandate as a cloud-based photo repository, and today its most important feature is memory summarization, which infuses data with nostalgia. Of course, perhaps Google is in the best position to provide this kind of connective tissue only because it can force the internet to bend to its will.


With Pixel Screenshots, Google is building on the success of the Photo Evolution by tackling the perennial problem of context collapse. The photos, files, and other stuff I store on my various screens, and especially on my phone, are quickly becoming irrelevant, and while Google’s solution isn’t all-encompassing—you still need to add notes to the most important items, for example—the integration with Gemini makes it easy to see right from the start why AI is so good for this particular use case.

Seeing Pixel screenshots in action and, more importantly, talking to the team behind the app, created the aha moment that so many people had. ChatGPT or Midjourney a few years ago. Pixel Screenshots itself may not be the answer to my problems – let’s be honest, with Google’s track record, it may not be around in a few years – but the way it focuses the special discovery and summarization power of AI is exciting.


That it’s taken me this long to get here may just say something about how weird my brain works, but I’ve spoken with enough AI skeptics to know that I’m not alone in this miasma of confusion, and that it’s often these small moments that lead to big behavioral changes.

Leave a Reply

Your email address will not be published. Required fields are marked *