The paper computer

It's essentially a poor man's hacked up DynamicLand - projector, camera, live agent. There are so many things you could do if you had a strong working baseline for this. My kids used it to create stories, learn how to draw various things, and watching safe videos they could hold in their hand.

There's something weirdly compelling and delightfully physical about holding a piece of paper that shows a live rocket launch, with the flames streaming down the page. It could also project targeted pieces of text, such as inline homework advice, or graphs next to data. It doesn't take long to imagine any other number of fun use cases, and it feels a lot more freeing and inspiring than keeping everything bound to a screen.

Github - https://github.com/Pugio/Orly (hacky minimal prototype that did the thing)

Video Pitch - https://youtu.be/-9l1x7GnmxU (filmed an hour before the deadline on an old phone with no sleep)

This reminds me of those predictions from 1900 about the year 2000, when they thought we'd all live in enormous skyscrapers and get around by flying cars. Instead we moved out to suburbs because improved logistics systems meant we could buy things from suburban shopping centres rather than having to go into city centres. Revolution, not evolution.

Surely the real advantage of an 'actually good AI' would be getting the AI to do the work itself, rather than just allowing the work to be done in a format with which the human is more comfortable. The underlying problem is that there are too many things vying for our attention.

[0] https://www.youtube.com/watch?v=7wa3nm0qcfM [1] https://dynamicland.org/

https://daily.tinyprojects.dev/paper_website

https://news.ycombinator.com/item?id=29550812

I question the idea of pastoralism though, I would argue this is another kind of construct. Laurel Hatcher Ulrich’s ‘age of homespun’ talks about this in detail, and how handcraft revivals were an expression of fear or anxiety about the radical changes brought about by industrialisation, and became a sort of myth making device for the rejection of technological overlords.

In any case, Paper Computer charts neat reformulation of the personal computer into something more interesting. If all individual computing tasks become distributed back into real spaces, objects and physically manipulable media it becomes more of an interpersonal computer, and distributed computing power can be pushed to things that don’t ordinarily engage with computational tasks such as wind or plants or anything within the shared working environment.

Human picks up all the sheets out of the printer, writes out replies with pen

Human puts the stack of answered email sheets in a multi-page scanner

Scanner physically scans them, agent transcribes them and matches them back to the incoming emails via the unique ID on each sheet, sends replies

You could adjust this flow for anything where human input is just one part of a larger sequence: just add print -> write -> scan into your flow where you'd normally have a human type. It's kind of a rebirth of faxing

https://x.com/daviddorg/status/2037050583274954882

https://x.com/daviddorg/status/2033937383012635065

https://yearunplugged.com/newsletter

Using paper and space to organize ideas is nice, but that's a niche use-case. And in any case, you'll have to digitalize it anyway afterwards, so better start on the digital version immediately, and be good at it. Everytime I start a new project, I'm tempted to take a pencil and paper, but then I refrain and use draw.io or the like because I know it will be winning on the longer run.

For the rest, you can easily customize your phone / browser / anything to be less distracting.

As for using AI just for convenience, this looks like very expensive in terms of resource.

UNIX Principle anyone ? Do one thing, and do it well - seems like in this 'age of AI' the industry is rediscovering by detour best practices, decades old, all over again.

But otherwise having 'interfaces' printed out to you and an LLM multi-modal later working from your notes on it sounds really interesting and less stressful than modern 'computing'.

The Office's Michael Scott would be proud - Paper may just be the future of Digital after all!

Just the other day, I noticed my thinking was so hijacked by distractions while building something (with AI help) that I started writing in a notebook to stay on track. The last time I'd written in the notebook was 3 years ago; in this case writing stuff down in it really helped to get me unstuck.

I'm excited to imagine workflows that could make computing a more physical activity. Thanks for writing and sharing this.

(My blog post btw if you’re curious https://bhave.sh/make-humans-analog-again/)

The only compromise would be a limited area like a physical desktop that had affordances like an overhead camera and some form of paper output.

That said, I do much prefer reading on paper, or at least on e-ink, for many of the same reasons outlined in the post. Computers and phones are just too distracting, and too dynamic.

And I'd love some way to write down shopping lists or appointments, and have them available wherever, without having to pull out the phone. Our current method is a whiteboard + a photo whenever we need it, which doesn't quite cut it.

I see this seemingly everywhere. People are looking for these extreme solutions to solve the problem of getting distracted by an app like Instagram or TikTok on their phone. Wouldn’t uninstalling the app, and going a step further, deleting the account, be the more pragmatic solution here? We control what is installed on our devices, what accounts we have, and which notifications we receive. If someone has enough agency to move to a pen and paper, surely they can uninstall some apps?

While I like the idea of having a magic paper notebook that would somehow interact with computer systems, that idea seems like mostly science fiction without having significant levels of technology all around you (cameras, projectors, etc) which would kind of defeat the purpose imo.

I watched the first video on Dynamic Land and I think I’d feel very uncomfortable in a room like that. Look the wrong way and catch a projector’s light in the eye, and once big tech gets into the game, who knows what happens with all the data from the cameras. I’ve grown rather paranoid.

A phone with just utilities installed, no social media, or going a step further to something like an e-ink tablet (something like Remarkable), seems like it would get most of the way there and actually work today. The biggest concern then becomes the web browser, but the big tech companies do most of the work for us by making sites insufferable to use while logged out and without an app.

Something might be able to get rigged up with RocketBook as well, for an actual pen on paper experience, but having to take a picture of the pages is kind of a pain. I have one and the novelty wore off very quickly; it has sat in a drawer for years now.

I’ve struggled with this idea a bit myself, as I sometimes romanticize the idea of using analog tools, but when they exist alone on an island, that seems to come with some considerable downsides in the modern world.

Apple Notes can be good for some of this too. Instead of using ChatGPT, Apple Notes can use the phone camera to do live OCR on text and add it into a note. I’ve used it a couple times and it’s pretty handy, when I remember it.

Also, check the spirograph too, among the slide ruler and any abacus.

Emacs, and technologies built on it, such as org-mode, come somewhat close to ideas expressed here by having plain text in a buffer be the unifying data format. You can organize stuff by just moving snippets of text around.

I think it's difficult in practice to design data manipulation interfaces based on real-world objects because atoms are heavy and bits are not. Data is just much more malleable and transformable than real world objects, at least at the pre-Diamond Age tech level we're at. But maybe ML will help make this easier by allowing computers to track and scan the objects more easily.