Gorilla Newsletter 65

Welcome back everyone 👋 and a heartfelt thank you to all new subscribers that joined in the past week!

This is the 65th issue of the Gorilla Newsletter - a weekly online publication that sums up everything noteworthy from the past week in generative art, creative coding, web3, tech and AI - with a spritz of Gorilla updates.

If it's your first time here, we've also got a discord server now, where we nerd out about all sorts of genart and tech things - if you want to connect with other readers of the newsletter, come and say hi: here's an invite link!

That said, hope that you're all having an awesome start into the new week! Here's your weekly roundup 👇

All the Generative Things

Generative Architecture: Unbuilt by Jacek

Unbuilt is an iconic piece from the early days of fxhash by Jacek Markusiewicz, who you might recognize from his more recent project Barbarians, that released back in February/March 2024 as a Verse Solos piece.

This past week we got a recount of the original inspirations behind the beautiful generative architectural project directly from Jacek himself, in form of a video space on TwiX hosted by Anaglyphic, who guides the conversation with a number of excellent questions:

In the recount Jacek reveals that Unbuilt is a tribute to Palladian architecture, an architectural style that derives its name from Andrea Palladio, a classicist architect from the 16th century (1508–1580). He describes the architectural style as "very mathematically definable", with the Wikipedia section on the topic echoing this, stating that the main focus of this architectural style lies on symmetry and perspective "to emulate the grandeur of classical buildings" of ancient Greece and ancient Rome.

Moreover, the project hails to the research of Gines and Stiny who studied the architectural rules that Palladio describes in his books as a form of architectural grammar - quite literally describing it as a generative approach to creating the structures:

In Unbuilt Jacek uses this concept of a generative grammar and applies it in the setting of French sacral architecture, to abstract and capture the architectural idea of a cathedral and then producing variations off of it. I recommend checking out the research that inspired Unbuilt, it's quite fascinating and might give you some ideas for your own generative artworks.

While there were a few hiccups with the VOD, I did enjoyed the format; being able to see the artworks while listening helps quite a bit. And apparently this is one of a first in a series of such "Welcome to my practice" style videos that Anaglyphic's planned - the second one's already out as well if you're curious:

Neural Mesh Cellular Automata + SIGGRAPH 2024

SIGGRAPH 2024 will be taking place from the 28th of July till the 1st of August, and considering that it's internationally the most prestigious computer graphics conference, there's bound to be more than a few interesting papers (which I'll be happy to share in future issues of the Newsletter) - you can find the full 2024 program, with links to each individual presentation here:

I learned about one of the interesting papers that's going to be presented by chance, through a Twitter thread from Ehsan Pajouheshgar, who's collaborated with Google Research on MeshNCA, short for Mesh Neural Cellular Automata: it's an approach for applying animated textures to 3D objects without requiring UV maps - here's a link to the thread where Ehsan elaborates on it:

While quite a bit of the math involved eludes me (tweets 3, 4, and 5), the showcased results are simply stellar - hence very worth checking out the entire thread.

I've written about cellular automata on a few occasions for the Newsletter before; most recently we had a look at Particle Lenia, a modern particle based approach to cellular automata. One type of cellular automata that I haven't tackled before however are NCAs, the neural variant. They're hybrid models that combine the grid-based structure and local interaction rules of cellular automata with the adaptive learning capabilities of neural networks - here you might remember that paper from a few years back with a self regenerating pixel lizard:

As far as I understand it, with an NCA each cell in the grid holds its own mini neural network that computes the next state based on the state of the neighbouring cells - in the context of NCAs we refer to this aggregated state as a cell's perception. A really good explainer video on the topic is by Data Garden:

I think at this point I might have to do a proper full-length article on cellular automata and all of the different variants that are out there.

Meta Morph Awards

We've got two generative artists claim trophies in the Metamorph awards that were announced last week - and I have to say very well deserved! For one we've got Greweb winning the generative category with his plotted animation piece "L'Homme et la machine", and for another Frederik Vanhoutte winning the NFT category with his submission "Isoverse - Invisible Cities" - both stellar pieces:

Link to Greweb's Tweet | Link to Winterbload's Tweet

You can also check out the other winners over on the Meta Morph page:

Other Cool Genart Things

Peckham Digital took place last week, a two day festival of creative computing, while I don't think that the session were recorded, we do get some insights from Amy Goodchild's most recent Newsletter where she also shares updates about her ongoing generative handwriting project.
An article on Hackaday shares algorithmic artwork by Prof. Halim Boutayeb who leverages Maxwell's equation in combination with a number of electromagnetic sources for the purpose of making colorful and fractal-esque plots.

Web3 News

Rodeo by Foundation

The Web3 news that made a buzz this week, was Foundation's launch of a brand new platform titled Rodeo - a Web3 social media platform, that at first glance seems very similar to Zora with it's main mint feed, but with a different micro-minting dynamic that does give it an overall different vibe. Before getting my hands on an invite I found this explainer by UV quite insightful:

When I first read UV's post I had concerns about the 24 hour minting window, that it might make for a stressful experience, but so far it hasn't felt that way. Logging in with my ETH wallet it also immediately picked up my Farcaster profile, automatically connecting me with others that I'm already mutuals with over on FC:

Got a nice dopamine hit off of seeing those first few notifications roll in.

The apparent simplicity of it in its current state is quite refreshing - although there are a few subtle aspects to it that create an interesting dynamic, and that also make it a little addictive. The best way to describe it, is that it's basically a Web3 Instagram clone where media posts (PNG, JPG, GIF, MP4, and MOV are supported at this point) are NFTs that can be collected within a 24 hour window after they've been posted. When you mint something is it also shows that in the feed, if someone else then subsequently mints it after seeing that notification in their feed, you get a reward.

I'm calling it micro-minting, because minting/collecting has a minimal fee attached to it 0.0001 Base ETH (0.35 cents at the time of writing). The only drawback that I see at this point, is the retroactive FOMO (is there a term for that?) for when I see all of the future mints that I'm no longer able to collect. Overall super fun experience so far, excited to see how it progresses!

Known Origin Shuts Down

At the same time that Foundation is launching a new platform, we bid farewell to another; Known Origin, one of the pioneering NFT platforms on ETH that took off back in 2018, and was later acquired by Ebay in June 2022, now made an official and conclusive announcement that it would be transitioning out of service in the coming weeks:

Link to Thread | Link to Article on NFT Now

This announcement is unsurprising, considering that Ebay had laid off 30% of their Web3 employees back in February 2024 - a short article about this can be read here. This now raises concerns about the permanence of the digital items that have been minted on the platform in the past 5 years. While the platform contracts are stored onchain, the actual files are in fact stored on IPFS, the popular decentralised storage system that is notoriously ephemeral - a tweet by Ella voicing her concerns made the rounds:

I assumed that it was common knowledge by now that IPFS is not a reliable storage solution, and that provenance is mostly decoupled from the actual media files for most NFTs. I remember that this was a big discussion back in 2021 - with IPFS, creators and collectors need to take care of pinning all of the files that matter to them manually. This then also led to services like Club NFT to pop-up, and later fxhash to introduce their ONCHFS (On Chain File System). If you need/want to back up your NFT collections you should look into Club NFT - I'm not sponsored, but it's the only tool I know of for this purpose:

This text banner on their landing page now also has KnownOrigin in big bold letters, not sure if it was like that before | Link to Club NFT

You can however also pin your files on IPFS manually, by spinning up your own node if you want to do it yourself. Overall, it just feels a bit odd to read this in 2024 - but maybe it's another wake-up call to address this particular issue with the tech again:

Things are already better than they used to be however, with many platforms, especially the ones geared towards generative art, providing the option to mint the code directly onchain, making the individual editions that are ultimately collected eternally reproducible. Naturally, only as long as the chain exists however.

ARTXCODE Rebrand & Misunderstood Machines

ARTXCODE made the announcement that on the 29th of July they'll be launching a revamped website with a brand new identity, enhanced with new minting capabilities powered by highlight, in addition to a genesis project by the founder of Sofia Garcia. The announcement thread over on TwiX provides all of the necessary details:

Link to Tweet | Link to the accompanying press Release

The genesis project of this relaunch, titledMisunderstood Machines is an interactive, generative art piece that aims to highlight both the emotional experience that code-based artists undergo, as well as the beauty of web development's foundational technologies. The artwork renders two decks of playing cards that flip over to reveal two words side by side - evocative adjectives on one side, technical terms on the other - and the juxtaposition of the two is sufficient to tell a story:

You can also check out the TwiX space that was held for the occasion.

Sofia's article really resonated with me; my personal journey has been somewhat similar when I look back on it. Albeit the other way around, I've somehow slowly found my way more and more into web development through generative art - discovering the tech as another means for creative expression.

In fact, in the past couple of weeks I've spent more time on programming and learning various web things compared to making generative art - and it's been really refreshing, it's felt a bit more purposeful overall and I wake up in the morning excited to learn something new.

Tech & Web Dev

The CrowdStrike Disaster

This past Friday (the 19th of July) things went south real quick when CrowdStrike rolled out an update to their Falcon software, an “endpoint detection and response” (EDR) software that monitors machines for nefarious activity, affecting 8.5 million windows machines around the world, and causing the biggest IT outage in history.

The faulty software caused machines that it is installed on to crash on boot with a blue screen of death, and subsequently rendering them inaccessible by leaving them stuck in a boot loop - naturally, this made many operations around the world to come to a grinding halt, including airports, hospitals, and banks:

If you've never heard about CrowdStrike, it's basically a cybersecurity vendor that provides large corps with anti-hacking software (to a large extent fortune 500 companies) for windows machines. And because their software isn't just your regular run-of-the-mill McAfee anti-virus, it needs to sit in the critical path of the machine to safeguard it - this is what makes all of this such a big disaster: affected computers need to be booted up in safe mode so that the faulty driver can be removed manually, and most company devices don't allow this.

And that's pretty much the gist of it, Fireships code report covers the entire debacle in more detail if you want a quick and fun watch:

Luckily (sadly?), I'm a mac chad (and also work freelance on my own machine) so I am unaffected by this, meaning that the weekly newsletter can be delivered in a timely manner 😌.

Also, I do not wish to trade places with the CrowdStrike employee that was responsible for this.

Drag and Drop Websites

I just learned about mmm.page - a website builder of a different kind, that lets you assemble your own custom and quirky webpages simply dragging and dropping elements, and that without requiring any programming knowledge at all:

The tool's been in development for quite a while, and launched as a beta back in January 2024, and just yesterday launched it's 1.0 version - and it already has quite a buzzing community that've created many artful and quirky webpages that hail back to the early days of the internet - you can find the anonymous founder of the platform XH over on Twitter where they share a bit more about it:

In an interview with TechCrunch earlier this year XH explains some of the motivations for building mmm.page:

I grew up on an internet filled with personal websites and niche forums that existed, sometimes, for the benefit of three people who shared an interest,” XH told TechCrunch. “It was fun to explore, and I remember feeling like I wasn’t supposed to be privy to these various subcultures from the remoteness of my suburban childhood. To a kid that kind of freedom is exciting.

Next Gen CSS Variables

CSS variables get a power-up with the @property rule that is now supported by all modern browsers, web.dev celebrates this with an article that provides a run-down of the possibiliteis:

So now you can basically create variables in the following manner, to also specify a variable type and a fallback value:

@property --myColor {
  syntax: '<color>';
  inherits: false;
  initial-value: hotpink;
}

What's cool about this? One cool use-case is that it's now possible to animate gradients with vanilla css which wasn't possible before:

One of the neat applications of @property is the smooth animation of properties that were previously impossible to transition, like gradients, which are perceived as images by the browser. However, if you give variables syntactic meaning via @property, these can be used in a gradient statement.

Unfortunately the embeded codepen in the article seems to be broken, but I did however find this YouTube video by Frontend FYI that explains how to do it:

And this is now supported in all browsers!

AI Corner

AI Powered Semantic Grouping

I believe that the current advancements in generative AI can only come to full fruition if we build innovative and novel interfaces through which we can lay bare their capabilities in a useful and functional manner - we have these super powerful neural machines, but our main way of interacting with them is a simple chat box. Surely, there must be a better way? And this is exactly what Amelia Wattenberger is currently exploring with creative works.

Link to First Tweet | Link to Second Tweet

Her most recent tweets showcase an AI powered application that helps with labeling user aggregated bundles of notes/texts/ideas on an infinite by automatically generating a semantic label to the clusters, summarizes the overall content. I'm assuming that is done with an LLM, by feeding in the text, alongside a prompt to indicate what needs to be done. In another version, the tool can also automatically merge individual ideas to breed new cards that contain a text that tries to combine the ideas of the two original cards.

This is genius, because the interface here plays a crucial part in creating a new affordance. It's not only that the summarisation/merging process is streamlined, but it also creates a new way to view and analyze the data, basically asking what's the main idea if I group these couple of notes together? Seeing what the AI thinks and what label it generates is interesting, and can spark new ideas and/or perspectives - and I think that's what AI is currently all about: to act like a side-kick, a smart assistant, and a second brain. I'm hoping that Amelia will release this in one form or another at some point.

Negotiating for a better deal with AI

On a related note, what if we used AI for negotiating purposes? I came across a tweet by George McGowan, where he shares the website of a mattress vendor on which you can haggle for the price of their products with an "AI" chatbot:

I immediately had multiple questions, the first of which... why? Why let potential buyers negotiate for the price? And what's the point of using AI for this purpose?

While I have no idea what a good mattress costs these days, using this tool implies that the product is overpriced from the get-go - it seems that they're happy to sell the mattress at a lower price depending on how much time and effort the buyer is willing to invest in negotiating. That doesn't actually mean that you're getting a good deal though - in fact, the product might still be overpriced in the end. But psychologically, the fact that it is offering a deal might be enough reason for some customers to seal the deal and make the purchase. The tech that's behind this negotiating bot is called Nibble:

I put AI between quotation marks earlier, because it seems that it's not really the kind of prompt-based LLM interface that we're familiar with today, but rather just relies on some hard coded logic, that will ultimately offer you the lowest possible price if you try enough times, as some of the replies under George's tweet suggest.

Maybe there's a little AI involved in the text generation part of it, but that also to a limited extent. Clearly it'd be too risky to allow for an AI to have control over choosing a price - since there's more than a few ways to break them. I guess it does makes sense to slap on AI as a trendy buzzword for branding purposes however, but I'd be curious to see a more intelligent AI based version of this. Idk, that's my 2 cents here.

Suno & Udio Sued for Copyright Infringement

Hold the front page! AI startup [insert AI startup name here] is being sued by [insert other company name here] for copyright infringement. Who would've guessed? (Spoiler: Everyone.)

This piece of news is almost a month old by now, but somehow I only caught wind of it now - maybe because it's not really a new shtick - but the two generative AI music startups Suno and Udio are getting sued by more than a few record labels for copyright infringement:

A group of record labels including the big three — Universal Music Group (UMG), Sony Music Entertainment, and Warner Records — are suing two of the top names in generative AI music making, alleging the companies violated their copyright “en masse.”

This debacle ensued after the AI generated track BBL Drizzy by King Willonius took off and became viral on multiple platforms - and it's actually a pretty good track:

It will be interesting to see how this unfolds, because, while it is evident that the models are trained on copyrighted material (where else is the data coming from otherwise?), it's very difficult to factually and conclusively prove it. Another point of view suggests that training AI on the copyrighted material is fair use, because the ultimate product is of a transformative nature.

It might also relevant to bring up the lawsuit between the New York Times and OpenAI here, that has not come to a conclusion yet as of now - with the current state of affairs indicating that it will likely not end in favor of NYT:

Gorilla Updates

Simultaneously while I was in Berlin at the start of July another important personal thing happened - a couple of iterations of my generative project "Some Parts of Me" were exhibited at DiGen Art back in Lebanon, for Beirut art week, as a part of their inaugural Digital Me group show. The pieces were specifically made to perfectly fill out the two main big screens that adorn the wall right next to the entrance of the gallery:

I have a technical/inspiration write-up for the project that's almost done - but the project itself will need some more time before I fully release it, I have some ambitious ideas for it, which I still need to figure out. On a related note, I'm also in the process of giving a mini course about creative coding with P5JS for a small group of students, which was set up in collaboration with DiGen Art and Code Brave, a Lebanon based non-profit that aims to teach kids programming for better career opportunities.

Because I don't have the time to be involved with all of the course sessions, I wanted to at least provide something for the students to explore on their own time, which is how I ended up putting together a mini course site that's a little more complicated than just a simple syllabus page - and basically ended up being a mini CMS in disguise.

I will share a link to the page once it's ready. I also spent a copious amount of time on the design, I think it ended up being quite quirky although designing a user interface is never really easy, even if it's something as simple as this:

Yes, I could have equally just spun up a Jekyll blog, but it kinda felt overkill for this purpose. While it's maybe not very impressive, I do think that I found a creative solution for storing and serving markdown files via a Github repo. I'll hopefully also make a proper write-up for it when I find the time 😌

Ultimately, the cool thing about it is that it can be easily repurposed for other courses/content - right now I'm still trying to figure out how to best do that with a minimal number of actions on the user's behalf. Consider this mini web project my answer to all the crappy course syllabi that I had to endure throughout my years at uni.

And as the banner image indicates, I also picked up another older genart project this week - not sure what I'm searching for at this point, but I'm having fun exploring:

What would homework look like if you went to genart school?

Music for Coding

This past week I was captivated by Steven Lynn's "Soundtrack for an Imaginary Western". The sounds that Steven's crafted is highly reminiscent of the legendary compositions by Ennio Morricone for "The Good, the Bad, and the Ugly", with the eclectic blend of unconventional sounds, especially the use of an electric guitar, which crafts that very distinct western atmosphere. If you're feeling like having a showdown with your bugs at high noon, this is probably the soundtrack for it:

And that's it from me—hope you've enjoyed this week's curated assortment of genart and tech shenanigans!

Now that you find yourself at the end of the Newsletter, you might as well share it with some of your friends - word of mouth is till one of the best ways to support me! Otherwise come and say hi over on my socials - and since we've got also got a discord now, let me shamelessly plug it again here. Come join and say hi!

If you've read this far, thanks a million! If you're still hungry for more generative art things, you can check out last week's issue of the newsletter here:

And backlog of all previous issues can be found here:

Cheers, happy coding, and again, hope that you have a fantastic week! See you in the next one!

- Gorilla Sun 🌸