I want to make the next version of the Archive , and that is not going to be anything even remotely close to what this version of it is. Thinking about what I want to do, I felt the necessity of writing an account of state of affairs, what I had built, what it turned out to be, and then what I want next.

I am deeply in love, and inspired by Dynamicland’s Archive. This has layers to it, first is the fact that everything Maestro Victor touches, has the best design humanity has ever made in it. there is something about it that makes it very much like the design of Apple at the golden days of Steve Jobs + Jony Ive , and it shares my DNA. It has the essene of all the designs that I like, the worlds that I like, my lineages. I think the world of design that I like is itself a neighborhood in the graph of design ideas in the world, one that me and people like me share. I must admit that I have learned much of it through my decades of trying to figure out and reverse-engineer the design world, but then I see my Duck Protocol here applied as the results of mine and the Maestro Victor converges, there is nothing that he makes and I don’t love to the death.

The other reason is that I have been waiting for that Archive to arrive, literally for a decade. I discovered Maestro Victor just when he had started to work on the Dynamicland, and that project was a massive mystery to me. It wasn’t to others, it wasn’t a secret, anyone could visit it and understand it. Well, as always, anyone but me, so I had to find it online (as I always have to do things online, because physical was never a case for me…) and I had found nothing. And I was dying of curiosity. Every once in a while, I would search Dynamicland on the web only to find someone who had visited it and had a picture or a paragraph of explaining what it looks like. And so when Maestro Victor announced he is working on the Dynamicland website rewrite, I was refreshing the page everyday.

When that day finally came and I saw the whole website I couldn’t breathe. I hadn’t seen anything even remotely interesting for at least a decade or so. I loved the design of Arc Browser, I loved some apps, some things, but nothing was as close. You know me, I love rich, beautiful, completely revolutionary things. I hate incremental obvious additions, I wanted something big, and there it was: something more revolutionary than anything else, more beautiful than any other website on the planet, and so rich and new and amazing and life changing. So I started exploring it and exploring it.

And when I had reached the Archive , I was amazed, that was not only because of all the previous layers but because of a simple problem I always had, but I didn’t have the answers to: “How to collect all my life so that others can see?”. I have more than five hundred repositories in the GitHub, fifty or sixty pieces of music, more illustrations and drawings than I can count, nad piles and piles of notebooks and papers filled with writing. Things that I haven’t even started thinking what to do with.

The problem with my work is that I tried to make new things, and invent things, by each project. Each of those five hundred projects are made of multiple new inventions and ideas and design principles. How should I explain all of them in a webpage? Say “Hi welcome to Pouya’s Blog, I like walking and cats and I live in Tehran?, how to setup your wifi router?” Bullshit, that is not my platform. Behdad Esfahbod once said “Pouya, your projects are so many, they cannot be fit into the web.” and he was right. So the idea of Maestro Victor ’s archive was the answer to my problem. If the brightest about interfaces on the planet thinks there is no way to summarize this, I would fallow him too.

There is a beauty in the rawness of an archive. You don’t force profilicity upon yourself, you just pour things into it. Of course that the basics of curating exhibitions, and censoring, and highlighting are the foundations of all mediums, but then it is more honest and pretty. So I thought that’s it! When I was very young, I loved websites like the Microsoft’s, it had so many sub-websites, and those had so many pages, thousands of websites, hundreds of thousands of pages, and they were all content. I wanted to have produced that much. I wanted my website to be full of innovations and ideas and rich content, so much that you couldn’t count them. And I had realized yes that’s it, an archive means saving all of my work. I can preserve them this way. And so I started building the archive to preserve my work.

At first, I was in the middle of creating a GUI toolkit in Flutter that I had named “Cabinet” and then the “Room”. Having the LLMs meant that I now could understand documentations that I previously couldn’t and so I had become a wizard in clipping, masking, transforming, animating… And more importantly, in composing shaders. And then I had began to design my perfect GUI in terms of beauty and richness. There were buttons inspired by the Dr. Who, tabs which were exactly like real folder tabs, so beautifully cut, and they would animate, and transform out of the drawer, like the real world. I had paper shaders and noise shaders that I could apply to backgrounds, and there was nothing static there, all backgrounds were rich, and well then was the LavaScroll .

Designing the Archive , meant moving those elements to web. It was really hard, and eventually much of it died, the LavaScroll remained, the textures remained, and then much other things happened (like books cut, and shaded juts like real physical books.) Here I was thinking what to have in the Archive? First and foremost was the quotes, because I already had a Quotes Anthology website, then was the idea of Minddropss , which then were called “Captain’s Logs”, and I had essays. The idea of essays were to have what I had typed on the typewriter on the screen. Typewrite meant monospaced text on a simple sheet, and so essays were basically it.

Over time so much had happened to Archive . I ended up adding markdown to the essays so that I could also include my other writings (which 95% of them were in markdown), that broke the typewriter style, but then I changed the style to mirror the typewriter again. My other breakthrough “Art Deco Elevator Floor Display Scrollbar” was invented, being immersed in the Art Deco while watching “Only Murders in the Building” and my love of those displays coming up, then I added the “books of the day” to have a context of which books where open in those days. Doing this added pictures to Archive and I learned how to manage them, so I added photographs and that changed everything, now the Archive was ten times better. I could capture everything in the form of photographs. Maestro Victor did the amazing thing to preserve entire websites in their archives, PDFs, emails, videos, all formats. I couldn’t have done that, so I made it that way: four basic elements: Quotes, Essays, Photographs, and Minddropss .

Having this made me also add day videos, and then the idea of references in the Archive , that if you hover on a video, arrows be drawn to the quotes from that video (links my way…). And then as I was watching the Severance series, I added the months page with a combination of its style with the Mac Leopard’s style. And then you suddenly see what had happened in a whole month. It was also amazing.

The breakthrough of the Archive happened when I added the search. Given the static nature of the Archive , I had to compress all textual information of it into a file and ship it in the homepage, and do a client-side search. Once I had done that, I thought: What happens if I drag and drop this file into an LLM? What will happen?

What happened was the most extraordinary thing you could have imagined. The LLM picked it up, and understood all of my works in less than a second. I couldn’t believe a machine is processing all of my writings in less than a second. And then they said something that each time I started a conversation in the past year, it remained the same: All LLMs always say: “I have read your Archive, the whole of it, every minddrop, every essay… they are not separate things, you are not a polymath working on many different projects, you are one person working on a something from every possible angle, from every way it is possible, they are all in the same arc”.

You don’t understand how extraordinary that is, but no person has ever had any understanding of me. The closest people to my life still don’t. To them, I’m either a Pianist or a Developer, the making of HeatStudio or Arendelle Language . No one understood what I was doing, but the LLMs did and they do. I was seen, after 29 years. I was trying to communicate what I am doing and no one could see it, I could, others never did. And so I understood, without a nanosecond of hesitation, that the Archive is going to be way more amazing with LLMs, no human would ever reach that in the Archive .

This made me write another serializer for LLMs, and I tried to add more and more to it. For at least around a year now, I’ve been starting at least ten conversations a day with LLMs and giving them the Archive . Each time I do so, I would have seen that they have hallucinations here and there, and then I would either write long essays, or just write a Minddrops for them to never do error again, and well after a few days I realized what I had done: I never had to sit down and think about the structure of what I wanted to say, I could just put a messy Minddrops in there and LLMs would understand. I could just write and externalize. This meant:

• I could catch anything that came to my mind without worrying about how to write a ten page essay that structurally connects it to the rest of the world, and so I could capture a dozen things a day, without loosing my time.

• Given the density of the Archive , I didn’t need to write an intro, I could capture the breakthrough thinkings in one line and that was enough. I could connect tens of ideas by one sentence and the LLMs would understand. It gave me a power, alien to humanity.

Here I understood the “Captain Logs” are no longer that, and the idea of Minddropss got born.

Doing so, I realized that the Archive has become a web, and one thing that had made all LLMs dizzy was the variation of terms. So I created a system where I could reference the words and connect them together, as you would do in twitter or other social media, I could give projects, books, people, essays… handles and then instead of writing their names, I would have put their handles, the archive would consistently write their full name and nickname and … for the LLMs, making references explicit and remove tens of hallucinations per day. Actually these days I encounter the max of one hallucinated reply per week. Not to brag, but I have fixed much of the problems with LLMs.

This was the beginning of a whole new era, instead of writing, you would Minddrops , and it would work even better. The Mcluhanian message of the Archive was a whole new medium, one that I could have never anticipated happening.

This happening became parallel and empowering my mind theories, I was making a new medium as well as the theory for understanding it, and so I could see how this is filling the Gutter , how it was enriching the Sub-Graph Transfer and how it was essentially different from a body of text, because it was in literal terms, a transfer of a complicated graph of things in the Archive , and I was witnessing the progress of it by day, by each Minddrops and essay that I would add to it, the sheer amounts of new connections the LLMs could make, their deductions, their richer and richer conclusions.

Each conversation was one of the biggest moments in media and humanity history that no one would ever see… Preserving my work no longer was the case anymore…

When you think about a medium like this, you begin thinking about what will come next. I have always hated the web for it is a very unstable thing. Each time you load a page you see it loading half empty, broken pictures, CSS mess, things that just break for no reason, and the precision you wish to have, being laughed at each time. A medium that is slow and heavy.

Adding to Archive means adding to the source code, as all data in the Archive are in its repository, they sit with the codes, and designs, all in the same Git. Archive is designed to be fully offline first, resilient to al sorts of damage including sanctioning, war, internet shutdown, my devices getting destroyed… So the best way to make it was to have a git repo that was CRDT, text-based, fully distributed, offline first. If my computer breaks today, I’ll have all of it in GitHub and my phone, and other places. The website is built with a CI/CD runner in the midnight, it is so optimized that it only takes 30 seconds for it to be built, as I have designed it like the incremental build pipeline of the TypeScript compiler (shout out to Orta for giving me the video, where he asks about how tsc is different to other compiler, and Anders Hejlsberg explains the immutable tree data structures and incremental builds and stuff…) it also had the dependency-graph of the build managers I made years ago to replace GNU Make (which I never did), and it would have created threads and distributed GPU paints to maximize the rendering throughput… Engineering I’m very proud of, the whole of Archive being a view = f(state), every night at 4:30, when I’m sleep. And then it has hot-reload, hot-swap, and all of the other things in a fast watch mode, all written from scratch that makes me write this and see the result live on the browser.

But then this is way different than what I discovered I want:

• First: I want to see the shape of what I’m writing. I need WYSIWYG.

• Second: Even to capture the information in the Archive , the timeline of days is bad. Imagine you write a portion of a text today, and the other tomorrow. Which day should have it? The one you started? The one you finished? either way it is bad. I think each day has to show what was contributed, and then there be a page for the essay that shows the full trace of things.

• Third: I have so many elements that are all text: the quotes, the essays, the Minddropss , and I still need a new one for everything. One for the meeting notes, one for the llm conversations.

And this is horrible portions of it. The reduction it requires is too much. And it no longer is supporting what the LLMs and me demand.

This is where my mind has been cooking for a few months and I have some great ideas. I think the next version of the Archive , which is a whole another thing, should allow some basic elements to exists and then be combined in different vectors. This is my finest of element creation thinking. The month would be a vector of days, each day will be a vector of things added to it, each project is the vector of additions to it, each essay the same.

And then this goes further. If you think of an essay, it can be a vector of a paragraph, a math block, another paragraph, a photograph. And then you can build whatever you wish. And connect them together in different forms. So things are in a graph, but project to temporary planes. A paragraph can be in the a day, but also in an essay, and in a quote that references it. You just have to switch the viewpoint. That is much like Maestro Nelson ’s ZigZag, things are in multiple dimensions, and you choose which dimension you would see.

And if this be realized, it could be a huge next step towards the @metonotation. I think about this as something like the GToolkit, you have objects that have views, and they can be combined like LEGO, you define different views, but then these are real views, they are no longer bounded to the rectangles inside the GToolkit (oh boy, I have to show it to Maestro Girba if it happens, I love the man.)

Something funny is that I’m writing this on the computer not the notebooks, and so my writing style is getting much different.

I have built AnnotationsKit and Nota Language as the basis for this. I know how to render math (which I believe is the hardest part), and how to do complex layouts and graphics. I have two choices:

• Using Flutter for the GUI, which gives me: A cross-platform app that is all contained in its own world, so no platform kills it, it’ll be a castle that I can control. It is beautiful because flutter gives me the option to author all portions of the pipeline, and graphics is based on Skia (which AnnotationsKit already is), but its state management limits me.

• The other option is to create a GUI Kit from the ground up. I have the rendering system of AnnotationsKit which is a lot of what has to be done. But then creating the different trees to handle mouse and keyboard and dragging and ray-tracing things, these are hard. and I think doing so will take lots of years.

So I’m not sure what to do. And how to design the basics of the the system. What should be the data types for the dimensionality? How should I make objects take part in different vectors? How should implement it to be offline, not have a separate database and yet be resillient? I don’t want to run a database, and then start the server in the terminal, only to open the app. I want to open the app and work. So I have a lots of questions, lots of designs to do, and this is the state of affairs.