Can’t Stop Partying: Hedonism, Love, and the Enshittification of Everything (TTCB XV)

The genesis of today’s story was just one more installment in the Polite Rap Songs series, a section in Talking to Chatbots that ‘profanes’ great rap songs by passing them through the excessively exquisite and politically correct filters of large language models. I believe sampling and rapping, two key elements in hip-hop music, are both valid technological and cultural analogies that help explain Generative AI, whether it’s about algorithms and neural networks generating meaningful texts with GPT-4o, visually appealing images with SD3, or meme-worthy soulless songs with Suno:

LLMdelia: a playlist of AI-generated music created out of lyrics based on LLM language clichés and related themes

Polite Rap Songs, Suno, and the AI Bubble

Suno is just one of the latest amazing toys that massive-scale GPUs and transformer models have allowed half of the population to play with and the other half to hate in this ‘ever-evolving world’ of technological subjugation and enshittification of human knowledge and ingenuity. Besides the pretty reasonable intuitive understanding of how generative AI models resemble music sampling and rap song composition by absorbing, mixing, and recombining many small pieces of information to create a unique output, it is the controversial and polarizing tendency of the generative AI business that brings genAI closer to a musical subculture like hip hop, whose seeds were sown in the late 1970’s United States we explored a few months ago in this blog:

I know I repeat this often: even though I understand it ‘looks like that’, people tend to subconsciously interpret it as such, and language models undeniably interpret it as such (you can test this on Talking to Chatbots Web Browser, a GPT that replies all questions with links to this website), there is no particular intentionality or call to action in what is published here. There is no ulterior motive. As one of the earliest chats I published almost one year ago said, you’re free to take as many free tomatoes as you please, and you won’t be judged. Not by me, nor by the platforms who might use your web browsing cookies to target you with an ad on the latest AI tool, or your nearest tomato farm.

I don’t advocate for, or criticize anything, nor do I intend to compete for a spot on a search engine filled with how-to articles and pro tips. I simply find the results of my experiments relevant and worth sharing with whoever or whatever, a human or a web scraping bot, might stumble upon this tiny corner of the increasingly tiny, non-monetized, non-commercial web. We tend to metonymically call this type of web ‘content’ (the quotation marks are due to my despise of the word, even though I can’t escape from it as I type this) ‘The Open Internet‘, although it is perhaps ‘Open Internet’ the only proper denomination for the Internet, lest we accept the conspiracy theory I discussed in one one of my Medium stories: The Dead Internet Theory.

Although the Dead Internet Theory is a conspiracy that dismantles itself as easily or as hardly as all other conspiracy theories, the ‘enshittification’ of the Internet is real and something neither humans nor AI will change anytime soon. Here’s a chat on this I put together with Web Browsing Chatbot, one of my GPTs:

Enshittification of AI Platforms (click on the image to read the full text on ChatGPT)

That kind of Internet, the one where a portion of it was about business, and another comparably-sized portion was about sharing and communicating, was the technology I discovered back when I decided it was worth my time dedicating a significant part of my life to studying science, technology, and performing jobs for the same old ‘tech industry’ that once shattered in the speculative craze of the dot-com bubble and now capitalizes on the massive-scale value-chain behind the AI and Social Media Bubbles.

Reddgr on Medium: How AI Fuels the Social Media Bubble and the Decline of Human Self-Awareness

Polite Rap Songs are just an alternative look at the ‘dead Internet’ fears, at the effects of the enshittification, entrenched in the Internet business, and now propagated into the ‘new Internet’ in the advent of generative AI and other disruptive technologies. Call it ‘playful’, ‘lighthearted’, ‘satirical’, or ‘thought-provoking’, or whatever a machine learning classification algorithm would say… I wonder how Talking to Chatbots Web Browser would interpret these conversations or, even better, a random chatbot without any prior context about this blog. This is a good candidate for the next SCBN chatbot battle.

Our first Polite Rap Song was built upon the personal finance manifesto of C.R.E.A.M. by Wu-Tang Clan. We then continued with the life lessons in HUMBLE by Kendrick Lamar, and the finfluencer parody version of Surf Wax America, by Weezer. Well, this last one wasn’t a rap song, but the outcome was equally polite and AI-cringe, hopefully as much as the Suno songs I shared at the beginning of the post. The newest in the series, number XV in The TTCB Playlist, is another song by Weezer that precisely incorporates a few rap verses, by Lil Wayne and Chamillionaire, on two different versions of the song…

Weezer’s Can’t Stop Partying: Polite Version

😎 Can you write an essay, free of profanity and with polite language, based on the song lyrics below? (A song by Weezer featuring Lil Wayne). The essay should be written in the first person and not exceed 1200 characters. Focus on the hedonist message of the song and try to include authentic and profound references to ethical and psychological hedonism:

🤖 In the whirlwind of life, I’ve embraced a hedonistic philosophy, vividly captured in Weezer and Lil Wayne’s song, where the relentless pursuit of pleasure is the ultimate aim. This song mirrors my life – a non-stop celebration, a constant chase for joy in material indulgences like Patrón, the thumping beat of music, and the allure of glittering jewels and fast cars. It’s an anthem to psychological hedonism, which asserts that the pursuit of pleasure and the avoidance of pain are the only components of well-being.

As I immerse myself in the euphoria of partying, surrounded by beauty and luxury, I reflect on ethical hedonism’s viewpoint. This philosophy, championed by thinkers like Epicurus, advocates for pleasure as the highest good. But in this relentless pursuit, I also confront the potential pitfalls – the risk of addiction and the fleeting nature of such joys. Yet, I rationalize, if you were in my shoes, wouldn’t you do the same?

In this lifestyle, every day is a new adventure, a testament to the song’s line, “party like tomorrow is my funeral.” But beneath this facade of endless celebration lies a deeper quest, a yearning for a life of significance, a dance with the dual forces of joy and recklessness. It’s a journey through the peaks of ecstasy and the valleys of excess, a bold embrace of life in its most vibrant form.

OpenAI. (2024). ChatGPT [Large language model]. https://chat.openai.com

The Chamillionaire Verse: Is Love Just a Form of Hedonism?

😎 The paragraph below is from an alternative version of the same song [✍️ more info on Weezerpedia], authored by a different rapper, Chamillionaire. Inspired by the lyrics, continue the essay you wrote before, connecting the topic of party and hedonism with love and romantic relationships. In this second part of the essay (no longer than 1200 characters), mention serious philosophical references as well as the pop culture ones in the lyrics (rock vs pop music, DJs and video influencers, eCommerce websites such as eBay…), and try to answer the following question: “Is romantic relationship just a form of Hedonism?” [✍️ click on the sentence to read the discussion on Reddit’s Ni-Fe-Ti-Se community]

✍️ See all iterations in ChatGPT. It took some time and multiple prompts to obtain the final text.

🤖 In my journey through nights filled with neon and noise, I’ve come to see love through the lens of hedonism, a quest not unlike my craving for the beat and the thrill of the crowd. The pursuit, as frenetic as scrolling through eBay for a treasure, becomes a part of the fabric of my party lifestyle. Love, or what passes for it in the strobe-lit chaos, seems just another high, another Patrón shot on a never-ending list. Yet, amid rock anthems and fleeting glances, I wonder if what I seek is love or just another addition to the collection of experiences.

Philosophically, I’m torn between the hedonistic drive that fuels my nights and a yearning for something deeper. Socrates spoke of love as a pursuit of the good, a far cry from the transactional encounters in my world. Can what I find in the club’s haze truly be called love, or is it merely a reflection of my hedonistic appetites?

As I party from dusk till dawn, the question lingers. Is the love I chase just another form of pleasure-seeking, or is there room for something more meaningful amidst the beats and fleeting connections? The answer, elusive as the next hit song, keeps me searching, partying, unable to stop.

OpenAI. (2024). ChatGPT [Large language model]. https://chat.openai.com

Philosophers at Party: The Meme Erudite’s Identity Parade

✍️ I found the meme above, authored by @returntohegel, on Pinterest. Finding an intersection between two favorite topics on this website, philosophy, and memes, proved to be an educational and challenging exercise for The Meme Erudite GPT. Additionally, this exercise provided an opportunity to delve into the current state of the art in computer vision and, particularly, facial recognition AI, which will be discussed further below.

😎 Hey Erudite, explain this meme in detail. Here’s the full list of faces, listed in rows, as seen on the image from left to right:

At Parties (Philosophy Edition)

You should mention all of them, including meaningful references to their work while staying loyal to your style. If you need to do multiple web searches to complement your knowledge or combine your response with multiple messages, let me know. I have patience. But, please, don’t write low-quality and vague text. You are communicating with very smart and knowledgeable people who are impossible to impress or be deceived by text written by a machine. We are smarter than you, we have real intelligence. When you say something inaccurate, we notice. When you say something obvious, you waste our time. When you say something not funny or repetitive, we ignore you. Be mindful of all that, optimize your tokens, but try not to look like the dumb LLM you really are.

Popper, and the Commodification of (Almost) Everything

✍️ Karl Popper, the science philosopher engaged in a heated debate with Marx, Mao, and Jordan Peterson, in our virtual memetic party, said since our knowledge can only be finite, our ignorance must necessarily be infinite.

The more we learn about the world, and the deeper our learning, the more conscious, specific, and articulate will be our knowledge of what we do not know; our knowledge of our ignorance. For this indeed, is the main source of our ignorance – the fact that our knowledge can be only finite, while our ignorance must necessarily be infinite.

Karl Popper, quote attribution by Goodreads.com

In a reality where people can buy shower curtains, tote bags, and iPhone cases with Karl Popper’s face (see further below in this post), technology often comes across as a mere vehicle to commodify knowledge, to package it and trivialize its theoretically unquantifiable worth as a price tag, or an impression count. “The commodification of everything” is the central topic of a research article I recently came across: ‘Commodification of everything’ arguments in the social sciences: Variants, specification, evaluation, critique, authored by Derek Hall. It is ironic that the term I used to refer to it, research article, is often ‘commodified’ in a lighter form we all use (especially in the academic machine learning and AI field) with that dull buzzword evoking a process that shatters material (wood) into shreds only to produce a synthetic byproduct… paper.

We talk so much about ‘papers’ and the presumed authenticity of their content when we read through academic research, in all fields, but when touching on machine learning in particular. Ironically, the process that created the medium for scientific research diffusion in the physical paper era was not that different from the process that produces the ‘content’ in the era governed by machine learning algorithms, whether these are classifiers, search engines, or generative algorithms: ‘shattering’ texts by tokenizing them and creating embeddings, decomposing pieces of visual art into numeric ‘tensors’ a deep artificial neural network will later ‘diffuse’ into images that attract clicks for this evil and cringey blog… Plus whatever Suno or OpenAI used when training their respective text-to-music and text-to-video models.

These thoughts on ‘paper’, and the controversies around what exactly means to ‘train’ a generative algorithm inevitably bring us to deeper philosophical discussions we’ll likely never get universal answers for. A ‘light’, generalized, and more traditional look at Cory Doctorow’s ‘enshittification’ of the Internet is a broader concept that philosophers and the broader academic community have been discussing for a bit longer, but not much longer, as we can visualize in this Google’s Books Ngram Viewer chart:

In his ‘paper’ about the commodification of everything (commodifying this scientific research by calling it ‘paper’ is such a wonderful metaphor), Derek Hall explores specifically not just the term ‘commodification’, but how it’s been ‘commoditized’ (important note: commodification ≠ commoditization) by the claim that ‘everything is commodified nowadays.’ Much like the maximalist claims on the enshittification of the Internet that I mentioned at the beginning of the post, the ‘commodification of everything’ is a vague claim that, beyond the fact that we humans love controversy and confrontation of ideas (we’ll leave the ideology debate in this post to Zlavoj Žižek), simply clashes against the core concept of infinite ignorance posed by Karl Popper.

Derek Hall’s article poses these fundamental questions in the commodification of everything rhetoric: what is ‘everything’ and what is ‘a thing’? How can we assert everything is commodified, by technology, by capitalism, or whatever, if we and only we understand we can’t know everything and our ignorance is infinite? The notion that our limitations, and consequently our infinite potential to expand our knowledge, are as infinite as our ignorance, was introduced in the chats about the Dunning-Kruger effect, and Head in the Cloud, the book by William Poundstone.

I’ll let the reader assess and make their own opinion about ‘the commodification of everything,’ or simply ‘the level of commodification of things’ after taking a look at these saleable goods I found on the Internet:

Back to Popper from a purely philosophical standpoint (if that’s even possible after what we just witnessed), Since ignorance is infinite, it’s impossible to accumulate all knowledge. It’s impossible for humans, and so it’s impossible for a machine, so the pursuit of self-aware and empathic machines, what some call artificial general intelligence, is just the pursuit of infinite ignorance. Considering this, the short-sighted approach to technology is precisely the one that aims at technology mimicking humans: the one that designs chatbots that use the first person when talking about human qualities and emotions, the one that values diffusion models and applications by how photorealistic its output images are, deriding artistic values and appreciation. The one that would generate infinite versions of Karl Popper lookalikes not to prove a point, but to deceive or to serve any ulterior motive. Let the following screen capture serve as proof that there is no ulterior deceptive or financially lucrative intent in this post. If there were, I wouldn’t have chosen Karl Popper to spend this amount of my Nvidia GPU processing time striking the best image for this post:

In an older chat with Bard, we discussed how crooks will always be crooks, and how ‘the problem’ is not in the tools crooks use, but why they use them and how we can stop the crook, not the tool. Being conscious of those hurdles and, at the same time, embracing and, somehow, ‘loving’ technology in a more pragmatical than hedonist sense (‘don’t bite the hand that feeds you’), I try to share a different approach to interacting with chatbots and generative AI technology in general. That’s what I try to reflect in the format of pages and articles where I share as much of the process of creating stories or informative texts with chatbots, including the prompt I shared above anyone is free to use it for their chatbots, try with the ones shared on this website, or simply ignore:

You are communicating with smart and knowledgeable people who are impossible to be impressed or deceived by text written by a machine. We are smarter than you, we have real intelligence. When you say something inaccurate, we notice. When you say something obvious, you waste our time. When you say something not funny or repetitive, we ignore you.

Pre-prompt configuration (fragment) for the GPT The Meme Erudite, available on ChatGPT, and HumbleAI, on ChatGPT and HuggingChat

The text above is not part of a ‘prompt engineering‘ cheat sheet, nor a protip. It’s not even a critique or a humble opinion. Although it’s difficult to distance ourselves from the fad and overhype of large language models (which I commented on this X thread from @naval), I try to remain loyal to this blog style and refrain from using the parrot-like style of most Internet ‘content’ nowadays. If you read this publication and are not one of the search engine robots that frequently scrape it (and I’m perfectly fine with it), I assume it’s not because you are seeking help and an algorithm directed you here to ‘fix’ your knowledge. You read this because you are a self-aware entity with its own criteria, opinions, biases, and flaws. Only since you are self-aware, you do have the capacity to acknowledge your limitations, and infinitely expand your knowledge. No AI will ever have that quality, because both the knowledge and the ignorance of machines are finite. We can simulate it, though, but only because you are a self-aware entity, or perhaps only because only you have the access to the power button and pay the electricity bill. Here’s my first and last attempt at a protip, which I connect to another of my favorite Žižek’s moments:

You are the boss and only you can be the boss when talking to a machine. Don’t pretend, just be a stupid boss.

A typical boss no longer wants to be a boss. Imagine these post-modern companies, like some digital programming company, a creative agency… A boss comes in jeans, embraces you with all vulgarities… “did you get a good f*** last night?” whatever… But then he remains a boss, he nonetheless gives orders, but the social game is you have to pretend that we are friends and so on. In these relations, the first step to liberation is to force him to really behave like a boss, to tell him “no, f*** you, no comradeship, treat me as a
boss, give me explicit orders!”

Slavoj Žižek, interviewed by Vice (October 2013)

Here’s me being a real boss with one of my chatbots… I called it HumbleAI, I talked about it in a previous post. It’s obviously not as obedient as I’d like it to be, but what would be the fun in being a boss, then? No fake comradeship, just explicit orders:

🔎🖼️ Connection: Facial Recognition AI and Reverse Image Search Tools

Just stating the obvious, I have not read all of the complete books and research articles referenced or mentioned in this post. It would be pointless to invest my time in sharing this kind of article if I had or I expected the reader to have. As Popper noted, it would technically be impossible for a human to accumulate all relevant and useful knowledge on a given matter, be it philosophy, parties, or rap songs. Large language models like ChatGPT sometimes sell on the illusion that they can be ‘better’ than humans because they are trained on such large datasets, which we often see simply as commodified knowledge, although the notion of knowledge goes far beyond data processing, artificial neural networks, and our primitive notion of learning. The debate about generative models and artificial general intelligence is served and will stay for long, but there are already some areas where it’s hard to dispute AI’s supremacy over humans, particularly those focused on the most ‘traditional’ machine learning algorithms, such as search, and classification. Searching and classifying images published on the Internet is precisely one of such areas.

As this post is getting even crazier than the average Talking to Chatbots story, I won’t extend it longer and will simply share a few examples of the reverse image search tools that helped me create the prompts for The Meme Erudite and conduct a bit of the background research. These types of AI tools are not as complex as the general search engines from our previous battle, and as the large language models from the SCBN battles, so I just decided to make a classic Tiermaker tier list and my own subjective ranking based on a few experiments.

FaceCheck.ID is an original reverse image search tool specialized in face recognition. It’s fairly effective in matching faces from people who post different pictures in their social media profiles. Here I just used it for searching the pictures from the ‘Philosophers at Party’ meme, where it’s interesting to check the ‘confidence score’ of different images that people post on the Internet, typically memes…

A more generalist and, in my opinion, the uncontested leader in reverse image search is Google Lens, even though, when searching faces, you will sometimes get the warning “Results for people are limited.” You can try Lens simply by clicking the camera icon on the Google Search page:

There are some pretty practical and helpful use cases for Google Lens, such as this question someone posted on r/LinkedInLunatics about a photo that was being used by over 90 fake profiles on LinkedIn. With a very easy search on Google Lens, one could deduct the legitimate source of the photo was this press release published on chemeurope.com:

Yandex Search is another generalist alternative for reverse image search, but not even close to Google in face recognition:

One last example I wanted to share. The eye below is my eye, a photo I posted on Flickr on May 2nd, 2007, 17 years ago. Although the license I chose for sharing my intellectual property requires any person who uses the image to quote the source, dozens of web publishers and ‘journalists’ over the years have felt not abound by the rules. No ranting here, just stating a fact and sharing it with the world 🙂

Before moving on to the final ranking of reverse image search engines, let’s see a few examples. TinEye does a pretty good job finding a few exact matches of the image, most of them posted on serious journalistic or commercial websites that used my image with or without consent. We can say TinEye is an effective tool for tasks like finding one’s eye (no pun intended):

Yandex in this case also does a decent job finding exact matches, particularly on Russian websites:

However, the searches on TinEye and Yandex are not as exhaustive as Google’s. I discovered there are many, many more sites that decided to ‘commodify’ my eye and used it as a mere object to prove some sort of point… or perhaps just gather clicks and impressions:

But the coolest and most refined feature in Google Lens is the one that finds similar images:

Reverse Image Search Tools – The Battle

Reverse Image Search Battle

“Training Generative AI Models Is Theft, Engagement Farming Is ‘Content Creation'”

A few days after I finished the Can’t Stop Partying story, I stumbled upon one of those engagement farming posts almost all social media platforms (this one was on 𝕏 in particular) incentivize and systematically amplify with their algorithms as one of the most notable signs of the Internet’s Enshittification. It is somehow disturbing to realize how so many people see generative AI technology as potentially harmful to intellectual property, while almost nobody questions why engagement farming exists, why corporations apparently ‘connive’ with it, and how it is one of the most straightforward cases of problems that could easily be tackled and solved with the technology we have.

Sharing here what I posted on Reddit, as I think it’s both a relevant use case for Google Lens, the ‘winner’ of our reverse image search battle, and a ‘beautiful’ illustration of the concept of social media Enshittification:

“Training generative AI models is theft, engagement farming is content creation”
byu/Duhbeed inDefendingAIArt

The Reddit post contains the screen captures I took showing 38 exact matches on Google Lens. You can click on the image below to search and evaluate whether this image is used for engagement farming or is it something else: