Per "Gimmic" Johansson
Freya Holmérmy new video is out!
this one is a lot more personal than my usual stuff, but I've felt a need to express this for a long time
Generative AI is a Parasitic Cancer
https://www.youtube.com/watch?v=-opBifFfsMY
YouTube
ari :autism: @acegikmo i watched this today and holy shit, the dead internet theory is not a theory anymore
jana (ACCOUNT MOVED)@acegikmo watched it as it came out yesterday Freyja, and loved it! Thanks so much for putting these thoughts into words
Matthew Dockrey@acegikmo It should make me mad, but really I feel claustrophobic and panicked. We've built a dead mall, and we're all stuck in it.
Phosphenes'Dead mall' is a good term. All culture and life replaced with just buying things.
Dieu@attoparsec @acegikmo the winning move of AI is not be good at anything but to make humans bad at everything.
Spy Stuff@acegikmo Just watched your new video—raw, thought-provoking, and powerful. Tackling generative AI as a 'parasitic cancer' takes guts, and you nailed it. Can’t wait to see the discussion this sparks! 🤯💻 #GenerativeAI
l'empathie mécanique@spythriller @acegikmo guts?.. This position is kinda mainstream. The discussion is going on for a few years by now. Haters continue to hate, merchants continue to sell, Internet gets enshittified with each passing day.
Yo Soy Freeman@acegikmo Thank you Freya, is important to have recognized voices speak about this. Really good video! Happy to see you posting again around here!
antimu0n 🤘 metal lynx ♿ 🏳️🌈🏳️⚧️@acegikmo just 10 minutes in and I want to [REDACTED] an AI datacenter.
It's so incredibly exhausting.
When people talk about AI making things accessible I do the desperate cry laugh, because I'm a disabled person and AI is as hostile by design as a too steep wheelchair ramp.
I'm however cautiously optimistic, for one because in recent weeks the signs are getting stronger that AI is running into fundamental walls that can't be solved with their usual (and only) strategy of throwing more compute (ie money) at it, and secondly because of the point you also touched on: those pesky humans simply refuse to stop creating art, it's almost like they enjoy the process or something.
A greater concern I have about AI right now and in the foreseeable future is that, because of its fundamental limitations, so much of it is mechanical turks. There are huge numbers of 3rd world workers employed to prep training data and correct flawed outputs, often under extremely exploitative conditions.
@acegikmo This video is so good! Thanks for making it! :D
It's just so weird how we over time have gotten more and more high quality and easily searchable content on the web, and then all of a sudden we start mass producing garbage and promoting it to the degree where searching the web for answers to simple questions becomes impossible.
Those glTF-related search results you showed were hilarious though 😅
It's just so weird how we over time have gotten more and more high quality and easily searchable content on the web, and then all of a sudden we start mass producing garbage and promoting it to the degree where searching the web for answers to simple questions becomes impossible.
Those glTF-related search results you showed were hilarious though 😅
Charlie. Probably.This is the death of imposter syndrome. @acegikmo
Any flesh-and-blood human with half a brain-cell who has any first-hand experience in a subject and a modicum of competence with words is being handed a golden invitation to write on their topic of choice, sure in the knowledge that it will be more worthwhile than AI-generated excrement. Many shy voices who actually deserve to be heard remain silent if they fear being below par. The par is now a rounding error and tending towards zero.
All we need is a search-engine to surface human-authored matter, only, and we'll be golden.
Thank you for using "parasite" and "cancer" together for this topic — I have honestly wondered why those aren't used ubiquitously because they are ideal descriptors.
unawarewolf 🐺@acegikmo
It's strange how anti-plagiarism people are til the techbros bust out their automatic plagiarism machines. Literally just a few stick figures are better and more creative than AI slop.
Hopefully it makes you feel better to know that AI's are getting fed their own slop though, and likely won't even be able to approximate content given enough time.
Strypey@gimmechocolate
> likely won't even be able to approximate content given enough time
Hopefully long before then we'll get across to the public just what a waste of computation and electricity the whole shell game is, and it will all get shut down. Along with the hyperscale datacentre model it's an example of trying to find a use for.
I wrote this in an attempt at educating and agitating on the topic;
https://disintermedia.substack.com/p/invasion-of-the-mole-trainers
F. Jairala@acegikmo This was not a stupid video. This was a video that expressed how you feel about something, wether we agree or not (and I do agree with you) you are still human and life will still have an impact on you! That’s why communication and expression are so important. Not even to mention that opinions are a tricky thing but also necessary. Not everything cab be fact or scientific based and even if it was, there’s always room for our own experience of things!
Celes 🌙@acegikmo as an application developer who often wants to add the functionality of GLB to 3DS in my .NET *or* Java, this was an invaluable source of information
and in a more serious note, thanks for putting this out there!
Vicente ⁂@acegikmo will watch later today!
Arto KoistinenMe: *trying to get Raylib GPU skinning to work on a model loaded from a GLB with absolutely no help from Google*
Me: "Oh cool, a new video from Freya!"
Me: ....
Great video, and I feel pretty much exactly the same!
I think the reason these personal AI gen things will never pick up is that stories are inherently social, and a personalised AI gen story is inherently anti-social. It has less meaning than a dream. But unfortunately, that won't stop the AI slop 😩
Blue Luma@acegikmo why at the end of your videos description don't you include a link to your Mastodon account as you do for Twitter, Bluesky, Instagram... ?
Is it voluntary ?
Is it voluntary ?
Jimjam@blueluma @acegikmo I will make sure to check out the video later today when I’m back home. Without watching it though, I already agree with general consensus that generative AI is largely a waste of time, resources, and talent.
But I agree with Blue Luma with the question or sentiment that you should have also (or only) linked your mastodon handle. The other social media platforms listed, especially Twitter, are near vomit-inducing at this point and mastodon could use more love!
cuan_knaggs@acegikmo this made me go look, and yes, there are multiple wordpress plugins that let you automate this shit generation
CounterVariable@acegikmo Sincerely, I understand how you feel Freya. My take is that the AI paradigm seems like a desperate final attempt to maintain control.
But in reality, information and art, are inherent to the individual rather than any external system.
I think of something that Benn Jordan said in a recent video. When the dust settles, we’ll still be around to to making the art without the need of the middlemen.
I have noticed that I’ve been buying more textbooks recently. :)
Maya Zimmerman@acegikmo After dealing with researching lzw last week, this video was a salve.
@acegikmo
1. this is devastating (in the way that journalism like this needs to be).
2. thank you.
/me@acegikmo I watched it last night and I agree with everything you said
Bret Mogilefsky@acegikmo Thanks for articulating this, and I feel the same grief for what we're losing. As that beloved firehose we drank from, filled with the care of excited strangers making things for each other, is steadily contaminated by AI-generated sewage, I'm choosing to search for joy and connection in the places and people closer to home, which are still tangible and meaningful. It really feels like a dark age is coming, and I hate that we have to accept loss and act in a survival rather than growing mindset, but at least we have each other. (Mastodon in particular has been a real refuge of actual humans interacting free of algorithms, and at least so far not exploited, and I really appreciate that.)
J M Berntson@acegikmo Oh, hey, the YouTube algorithm delivered this to be earlier today.
Hard agree.
Frederico Machuca@acegikmo you might find this useful. https://github.com/NotaInutilis/Super-SEO-Spam-Suppressor
GitHub - NotaInutilis/Super-SEO-Spam-Suppressor: An anticapitalist blocklist targeting websites abusing SEO tactics to spam web searches with data pollution and security risks: content farms, scrapers, copycats, generative AIs, scams, advertisements, malwares, and useless wasteful garbage in general. It is best used with uBlacklist.
An anticapitalist blocklist targeting websites abusing SEO tactics to spam web searches with data pollution and security risks: content farms, scrapers, copycats, generative AIs, scams, advertiseme...
Fr4nzAs you said most technology can be used for good or bad.
I would prefer if ai content would be clearly marked, but yes that will be hard. However i guess it would be even in the interest of the ai companies themselves.
But your video shows how broken Google search engine is. I think fixing search would help more here. To disscurrage ai slop.
Maybe incorporate social media in search too. So based on what friends and even friends of my friends consider value should be higher ranked for me.
I would prefer if ai content would be clearly marked, but yes that will be hard. However i guess it would be even in the interest of the ai companies themselves.
But your video shows how broken Google search engine is. I think fixing search would help more here. To disscurrage ai slop.
Maybe incorporate social media in search too. So based on what friends and even friends of my friends consider value should be higher ranked for me.
Outfrost@acegikmo I wholeheartedly appreciate you talking about everything in this video. I didn't even realise how much of a drain on my hope and wellbeing AI has been, and how much I've had to avoid discussions of it. I studied computer science in order to make a living from solving human problems and creating things that others would appreciate. Now I'm forced to watch these two things get eaten away by a fire that so many people seem to be roasting marshmallows over, instead of putting it out.
@acegikmo Between this, and being chewed up and spat out by corporate employers time and time again, it's been a difficult journey trying to find myself again. But I believe, as you do, that we will persevere; that as long as we have the desire to create, we will find a way. Those among us who care about genuine expression and meaningful connections will be guiding lights for generations to come.
@acegikmo (cross-posting this here so it doesn't get lost in the youtube comments)
Death Star Robot@acegikmo Not advice: Someone told me the universe is a fractal of Holons, there is always an inside and an outside. To me, my interpretation, which could be wrong, is that this video is like the inside of the holon trying to blame the outside of the holon.
@acegikmo No you can't choose to be happy if you are depressed, but maybe you can spend less time on things that make you feel depressed, like this video is 1 hours and 19 minutes, and move on to something more fun faster. I don't know if you can do that, but if you could, you might feel a tiny bit better, but not happy.
BeldarakThe "dead internet theory" becomes more and more a reality and less a conspiray theory each day 😢
Aras Pranckevičius@acegikmo great video, thanks ❤️ This answer to "do you think AI will ever be able to write a good song?" question by Nick Cave -- way back from 2019! -- I often think about. https://www.theredhandfiles.com/considering-human-imagination-the-last-piece-of-wilderness-do-you-think-ai-will-ever-be-able-to-write-a-good-song/
There's also his later (2023) answer about ChatGPT writing lyrics: https://www.theredhandfiles.com/chatgpt-making-things-faster-and-easier/
Considering human imagination the last piece of wilderness, do you think AI will ever be able to write a good song? - The Red Hand Files
Dear Peter, In Yuval Noah Harari’s brilliant new book 21 Lessons for the 21st Century, he writes that Artificial Intelligence, with its limitless potential and connectedness, will ultimately render many humans redundant in the work place. This sounds entirely feasible. However, he goes on to say that AI will be able to write better songs […]
Robin Hübner
Bornach@dpwiz @aras @acegikmo
Still holds up though.
https://youtu.be/l7tWoPk25yU
If it was another human who unseated Lee Sedol from the top spot, how likely is it that person could be consistently beaten by an amateur exploiting a simplistic "double-sandwiching" strategy?
ChatGPT's HUGE Problem
@aras and thanks a ton for your continued support!! ❤️
@aras I also hadn't read the second article you linked til now - 100% agree, nice to see someone hit on the same point
Sander van Rossen🇺🇦@aras @acegikmo somehow this makes me think of the line "all that remained where fields of dreamless solitude. there's no prize for perfection, only an end to pursuit"
If you truly had AI that could make music, movies and games better than humans .. and could churn new stuff out by the second .. it would end up feeling hollow and meaningless 😬
If you truly had AI that could make music, movies and games better than humans .. and could churn new stuff out by the second .. it would end up feeling hollow and meaningless 😬
Michael Vance@aras @acegikmo I'm sympathetic to Cave's position that without "pneuma" you are missing a specific piece of what makes the creative effort "awe inspiring”. For instance I saw this piece in person the other day and while it is visually arresting the depth of my appreciation for it changed significantly when I saw its mechanism of production, small glazed tiles of ceramic artfully arranged. https://whitney.org/collection/works/63687
@aras @acegikmo I think this viewpoint is susceptible to attack though, as it invites "gatekeeping" on who gets to be an artist. I feel like any essay around AI must grapple with simple analogies such as the introduction of photography and response from painting. Art and our dialog around it are not static, our expectations shift dynamically as tools and process and thinking expand, and we have to consider changes to our landscape as part of that.
@aras @acegikmo I say this as someone deeply skeptical of AI/ML techniques, but I also feel that part of my skepticism is an artifact of "taste", of a curiosity about how things work, and specifically things that I myself grew up with, specific objects and technology and technical approaches. I don't think we can be "wholly new" in our perspectives as we age, and this is simply a fact of the human experience and the arc of our lives, and this makes us suspicious and skeptical.
April (@39c3) The Pink ⋆☾╶⃝⃤☽⋆ [AS208709]@acegikmo oh that vid is by you? Congrats giwl!
Dushman@acegikmo
One big reason why AI slop is so prevalent that not many people talk about is capitalism. People pump this shit out not to help anyone but for ad dollars. Generative AI isn't inherently bad I'd say but it's definitely being misused.
One big reason why AI slop is so prevalent that not many people talk about is capitalism. People pump this shit out not to help anyone but for ad dollars. Generative AI isn't inherently bad I'd say but it's definitely being misused.
Tacitus 🇮🇪 
The "discussion" at the end reminds me of this for some reason
Rockwell Retro Encabulator
Latest technology by Rockwell Automation
@tacitus @acegikmo
I wonder what "podcast" NotebookLM would generate when given the turbo encabulator transcript
https://youtu.be/Ac7G7xOG2Ag
I wonder what "podcast" NotebookLM would generate when given the turbo encabulator transcript
https://youtu.be/Ac7G7xOG2Ag
"Turbo Encabulator" the Original
@tacitus @acegikmo
Oh good god! Someone has already had NotebookLM generate a turbo encabulator podcast.
https://youtu.be/bCfrFdgSj74
AI becomes a Dadaist
Oh good god! Someone has already had NotebookLM generate a turbo encabulator podcast.
https://youtu.be/bCfrFdgSj74
AI becomes a Dadaist
Turbo Encabulator deep dive (NotebookLM)
crypticcelery 🔙 EH23@acegikmo being a bit “philosophical” here: I think what we are witnessing here is a fundamental attack on the information commons.
First, the genAI bros abused the fact that making useful texts, nice art, etc. available essentially for free was just the thing to do, just to train the slop machines.
And now, the slop machines are being used to fill every corner with slop, just to eke out just that little bit of “engagement”, be it to sell something or to get ad revenue.
First, the genAI bros abused the fact that making useful texts, nice art, etc. available essentially for free was just the thing to do, just to train the slop machines.
And now, the slop machines are being used to fill every corner with slop, just to eke out just that little bit of “engagement”, be it to sell something or to get ad revenue.