Deep Learning Philosophy

Palinode Productions: Building the Living Cosmos of Ideas

Dear friends,

I’m really excited to share with you a project we’ve been brewing over the past year. I still can’t believe I get to call this my job, it’s more like a professional blessing. I’m able to dedicate my expertise and brainpower to something I find deeply meaningful, exciting, stimulating, and utterly weird. I’ve never done this kind of self-promo here, but I really think this project will resonate with many of my fellow nerdy-inclined readers, so please! Let me tell you a little bit about Palinode Productions.

I’m not going to try and sell you anything here, don’t worry. :)

Deep Learning Philosophy

I’m not really a pitch deck person, and I pray to gods I never have to become one, so you’ll have to bear with my usual meandering train of thought. At Palinode Productions, we develop a platform for new ways of learning. The need for it arose from several complex problems piling atop each other, primarily the general collapse of the education system precipitated by AI and… well, the good ol’ capitalism. The frustrated teachers who spent the past two years marking AI-generated papers, the inherent socio-economical elitism of academia … And the internet, with all its free lectures and amazing breadth of information that could have been the greatest tool to elevate us - seems to have degenerated into its lowest form, a sad shopping mall and a dumpster fire of dopamine rush algorithms.

If we are to mitigate this disaster, we need to start at the very tissue of the information exchange - the almighty Algorithm serving us content. Current social media systems are optimising for engagement, for likes, for ad revenue streams and for fixing people onto the shadows dancing on their walls. And as the ad for a product I thought of 10 minutes ago flashes in my scrollable side panel, I think to myself… those bastards really mastered the art form.

But just imagine what could happen if we chose a different set of values to optimise these immensely powerful technologies for. Such as … I don’t know… learning and curiosity? Something that leads us to be better human beings? To constantly learn about ourselves and expand our horizons, and thus grow our tolerance, acceptance and compassion? Maybe if we stop treating users as commodities, and start believing in the inherent spark of genuine brilliance inside each human being, we might start finding ways out of this miserable social slump called the 21st century.

Sounds pretty ambitious, right? Well yeah… Because it is. You all know I’m an old tech-doomer, but the vision of our founder, scholar of Plato and Professor of Philosophy at Gonzaga University, Dr. Danielle Layne is so inspiring, that even a mad miser like me dares to work towards a better future. (proof in the interview below)

Baby Steps

I found my way into the project in the most auspiciously mercurial way. Earl Fontainelle just released a SHWEP Oddcast episode we recorded earlier that month called ‘on Magic and Artificial Intelligence’. A few days later, I received an email from mysterious Palinode Productions, suggesting a meeting.

So I got to meet Danny and Tof, who’s been working with their team of freaky philosophers, philosophising an ambitious plan of re-building the internet. It took roughly 15 minutes to get me beyond excited about the project. Yes, yes! I’m totally down to translate all these wild ideas into data structures and algorithms, so we can actually build this thing. (the lucky SHWEP episode is attached below)

Comic Khora

We spent months designing and re-designing various prototypes, collecting references and comparing notes on different methods of learning and note-taking, sketching, coding, hacking, and trying to figure out what it could be. We finally decided on a basic set of essential features needed to achieve our goals and started developing a new algorithm we call Khora.

The content in Khora will initially consist of our partnered works, from professional philosophers, podcasters and writers. The user logs in, tells Khora about themselves and starts their learning journey. Seeing what kind of content the user consumes, Khora will learn how the users learn. In turn, she will serve content that challenges their current ways of thinking and their metaphysical worldviews, all in a digestible form. Slowly and gently, Khora will bridge the gaps between opposing ideas - not changing people’s minds, but allowing them to engage with different beliefs earnestly.

For example, someone identified as a hard-headed atheist-materialist will be gently nudged outside of their echo chamber. Obviously, we can’t immediately expect them to find beauty in gnostic gospels, as this will most likely trigger a defensive stance. Khora might start with Kastrup’s lectures on Analytic Idealism - a theory that might reveal the main problems with religious belief in scientifism.

Scaffolding the Universe

The inspiration for the backbone of Khora was my beloved note-taking app Obsidian, but on a LOT of semantic steroids. This means that we are, essentially, building a vast knowledge graph of Philosophy. This graph is built bottom-up, from content to the ideas to the network to higher ideals. Every component in Khora is nested into the wider context, relating through underlying ideas, time periods, institutions, analogies, references... Once the platform is sufficiently developed, professional philosophers, artists and students will all interact with content and create links, each in their own precious way, creating completely unique set of exploration paths and connections between the thoughts.

I won’t be telling you much more at this point, but the structure is intentionally kept very fluid - we don’t focus on formal reasoning and argument components, but rather on capturing the multi-faceted nature of learning. We let the weirdness and random synchronicities seep through the Khora’s marrow. Because human thinking itself isn’t rigorous, despite many vehemently insisting so.

Oh, and when it comes to the user experience itself, forget the feed. Our designers are exploring totally novel ways of interacting with content, trying to push outside the pre-charted routes and rulebooks and let their imaginations run wild.

Philosophy of Coding

And you know what’s great? The most exciting part about this project might not even be the algorithm itself. It’s the way how Palinode Productions completely re-thinks and restructures the usual modes of working.

Playfulness and curiosity is embedded in the very DNA of the project, and it really brings out everyone’s best side. And once you have an environment like this, where you don’t discard any idea just because ‘that’s not how things are done’ - you actually find out you’re building something truly unique. It’s not just a job anymore.

The algorithms I’m working on tweak the latest AI technologies (knowledge graphs, augmented retrieval, AI agents etc.) - a lot of them custom-designed by myself, just fooling around with the ideas and trying them out. I decided to re-take linear algebra classes (it’s been a while, and I LOVE waking up and watching lectures goddamnit bless the MIT open courseware) so that I can develop my own embedding algorithms and transformations, playing and twisting the meanings to the bone.

We discuss the philosophy and values of what we’re aiming for in all our meetings. What’s crazy is that many algorithmic problems I’ve encountered were actually solved … philosophically?

For anyone working in tech, this might sound insane - if you have a coding problem, you go through the motions: a 15-minute Google search that spews out a pile of stack overflow posts filled with snarky comments from infinitely programmatically superior users. Two or three medium blogs behind a paywall, most likely not worth it. Oh and don’t forget that one lonely GitHub ticket, filed in 2019 with three likes and no response. And if even after an hour, there’s really nothing to save the day, and you realise you might have to think for yourself for once, it’s time to make a coffee. There is a path to get towards an optimal solution, with some trials and errors and a lot of mathematical thinking. (Or, the 2024 way - you let GPT solve it for you in the most dumb way possible and buy a stronger server)

But when I was explaining some of the struggles I’ve been facing with certain algorithms, the philosophers would rather intuitively point me to a specific platonic text or an idea. I’m supposed to read WHAT to optimise the graph traversal? Can you even imagine how surprised I was to find out that the right data structure revealed itself to me while reading Timaeus?

After years in corporations, with all their clunky plug-and-play frameworks and bloated Python libraries, I’m finally rediscovering coding as an art, where beauty and elegance threads through the core. 

Partners and Friends

These are the early days of the Khora algorithm, and we are still just preparing the scaffolding for our Cosmic Brain. This is the reason why the partnered content is such a crucial component of Khora’s development. We already have many awesome thinkers and channels on board - from Dr Jeff Kripal, Dr Tom Cheetham, Dr Gregory B. Sadler, SHWEP, History of Philosophy Without Any Gaps, and many more prolific writers, lecturers and artists who decided to allow us to use their content in the initial stages. You can find the growing list in our news section!

Here is where I ask you, fellow writers and thinkers, to reach out to me if you would like to be part of this journey with us as well. Our partners share their content with us, and we ingest it into the Khora, embedding it into a wider context and linking it with all the other information already in place. Not only this makes our brain smarter and richer for future users, but you also get to see what content relates to your work, in what way and maybe you get to explore some of the rabbit holes for yourself.

We are slowly building our dream team of like-minded philosophers, programmers and business people, and we're very happy to welcome Adina Bezerita (FRSA), scholar, philosopher and founder of the Numinous Threads (2016) as a member of our Honorary Executive Committee!

Regarding the technicalities, we are able to process YouTube channels, pdfs, documents, websites, downloadable mp3s of podcasts and possibly more. Our primary focus is on content related to philosophy, but in the wider sense of the word - including esotericism, popular science, and opinion pieces. etc. When we use the content, the copyright isn’t affected. We don’t display any parts of the content on our site and we don’t claim any ownership. We only store external links, so all the interaction with the content remains at the partner’s end (likes, listens, etc.). We only do AI analysis on the primary source, extract the summary and process it through our ingestion pipelines to embed the content into our Khora structure. In case the author is concerned about the AI processing of the primary source, there are ways around this too.

So, my dear friends, thank you for staying with my little ‘advertising’ all the way here. Hope I got you a little bit excited about the project (hit that lil ❤️ if that’s the case) - I’ll be uploading updates on the Palinode blog, and from time to time link it out from here as well. Let me know your thoughts, questions or feedback in the comments!

And please, hit me up if you would like us to include your work.

Palinode Productions Website

Stay curious,

k