[links] Once upon A.I. / AlphaFold imagines all the proteins / AI-generates The Kubrick Times
The Terror of Everything, Everywhere All At Once
What a great, cool idea for AI-use: The Kubrick Times - "For 2001: A Space Odyssey, Kubrick’s team wrote 36 futuristic New York Times headlines to appear on iPad-like devices. HAL 9000 isn’t available, so we used OpenAI’s neural network tech to turn those headlines into full AI‑generated articles. GPT-3 wrote text from prompts based on the 36 original headlines, along with additional fact boxes from related phrases. DALL·E 2 produced images using a similar process. The fake ads use AI-generated photos and slogans. New York weather data for the year 2001 was sourced from Visual Crossing Weather."
In a similar vein: The Daily Wrong, "AI-Generated Lies Every Day" illustrated by Dall-E, and prompt.press, "AI generated artwork inspired by current events"
I want this AGI T-Shirt: "Scale is all you need"
GPT3 produces descriptions of famous paintings that go into Midjourney as prompts: AI Reimagines 10 Famous Landscape Paintings
Absurd AI-Generated Professional Food Photography with DALL-E 2 - "A Hamburger in the shape of five dimensions, professional food photography"
Tommaso Marinettis 1932 The Futurist Cookbook illustrated by AI: DALL-e meets Marinetti
Once Upon A.I. – Inspired by Yuval Noah Harari What does the mind-blowing story of humankind look like through the eyes of artificial intelligence? 'Once Upon A.I. – Inspired by Yuval Noah Harari' is an experimental short film featuring over 260 strange and spectacular images, generated by 13 different A.I. algorithms. It's a collaboration between Sapienship, Love Tomorrow, AKA De Mensen, Once Storytelling, No Computer, and... several machines. The film was originally created for Tomorrowland festival’s Love Tomorrow Conference in July 2022.
AI art is challenging the boundaries of curation - AI is emphasizing curation in the creative process on various levels.
This is not very surprising given how good language models have become, but still: an AI impersonating Daniel Dennett fooling Dennett-Experts is something: In Experiment, AI Successfully Impersonates Famous Philosopher
The Mirror of Language - Prompt design is the new alchemy: "The age of thinking machines will be a new age of mystics extracting truth from hidden worlds."
Everybody linked to this story in The Verge about writers incorporating AI into their textproduction and I hated it. They got the worst kind of writers (an ex-manager who now produces lowbrow cheap fantasy crap en masse and an SEO-spammer/life-coach) and never mention cut up techniques or experimental literature, once. Algorithms are part of literature for a long time, and you at least could mention people like Kenneth Goldsmith and his groundbreaking book Uncreative Writing, or I'd have to asume that you just don't know what you're talking about. The usual tech journo standards at work, i guess.
Artificial intelligence model finds potential drug molecules a thousand times faster and DeepMind uncovers structure of 200m proteins in scientific leap forward - AlphaFold predicts the structures of pretty much every protein and they are developing techniques to make this even faster. I want economists to find out how much cheaper the development of pharmaceuticals will become because of this technology and I want to know if those cost reductions will "trickle down" to those who need them. Also: "AlphaFold’s models are also being used by scientists at the University of Portsmouth’s Centre for Enzyme Innovation, to identify enzymes from the natural world that could be tweaked to digest and recycle plastics." I want the equivalent of water treatment plants, but with stuff that eats plastic. Can we do that please?
Great reply from Googles Blaise Agüera Y Arcas to the "Is LaMDA sentient"-debate: The Model Is The Message. More on AI Consciousness: Why We Need New Language For Artificial Intelligence.
This video muses about how the stunningly great Everything Everywhere all at once is about the Internet: The Terror of Everything Everywhere All At Once, but I would argue that it's actually more about AI. Dall-E 2 is nothing but a visual "Everything Everywhere All At Once"-machine. Helpful in this context is looking at Artificial Intelligence not as intelligent machines, but Large Language Models as a Cultural Technology, as explained by Alison Gopnik from UC Berkeley. Dall-E is a library of every image and you access pieces in that library (locations in latent space) with the new alchemy of prompts.
AI improves itself: Nvidia AI Research Team Presents A Deep Reinforcement Learning (RL) Based Approach To Create Smaller And Faster Circuits
There are not much details about these newsbits from AI research in China and my bullshit sense is on red altert, but there you go: Researchers develop a 'mind-reading' device to help censor porn in China / China Says It's Developing an AI to Detect Party Loyalty / Researchers claim China's new 'brain-scale' AI rivals the human brain / GPT-3 Scared You? Meet Wu Dao 2.0: A Monster of 1.75 Trillion Parameters / China Launches 'Mother Ship,' World's First AI Unmanned Drone Carrier
SNeRF: Stylized Neural Implicit Representations for 3D Scenes - Style Transfer for 3D-scenes, video here
Human and machine intelligence work together to find 40,000 ring galaxies
Brain-Computer Interface Startup Implants First Device in US Patient
How's the metaverse going? "someone completely recreated Kmarts with working scanners, p.a. systems, and more - and have a Discord with 3,000 people that occasionally roleplay working here unironically."
How is crypto going? NFT artist "Robness" mints an NFT of a journalist's childhood photo to harass her.
text-based nft platform ▄▄▓▓███▓██▓▀ ░ ▀▀ ▄▄ ░▓▓█ ███▓██▄▓▀▀▀ ▄▄▄▄▄▄ ▀▓▀▀▄▄█▓███▀█▓███▓▄▄▄▄ ▀▀░▓█¦Ì▓▓███▀ ▄ ░▀▀▀█▓▄▀██ ▒████ ▀▀ ▄▄███▓▀ ▓█▀ ▀█ ▀ ▀▓ ░███▓ ▓█████ ▀█ ░███░ ▓█ ▓▓█▀ ▄▄█▓ ▓▄▀▄▄ ▀ ▀ ▓████▄██████▀▀ ▄▓░███▄███▓▓▄▄▄ ███▒ ░███████▓▓▄▄▄ ███ ▀▀ ▄▄ ▓█████ ▓█¦Ì█░██▄▄ ██████▓▀▀▀ ▄▄▓███▄▄ ▀▀▀▀▀████████░ ▄▄ ▀▀▀▀▀████████¦ ░▓ ▓█░██▀ ▀█ ▀▀ ▀▀▓███¦ ▀▀▀ ▄▄▄██▄▄▀▀█▀▀▄▄▄██▓ ▄▄▄ ▀▀▓██▓░ ▓▓██▀ ▓ ▄ ▀▀▓██▓░ ▓¦Ì█▓███▄█▀ ▀▄▓ ▄▄▄▄ ▀▀▀▀▓█████████▀ ▄ ▀▀▀▓▓
Like Bill Gates before him, Mark Zuckerberg is having a ‘Pearl Harbour’ moment | John Naughton | The Guardian If TikTok is the future and FB is in serious decline, then i wonder where textual discourse will go. Yes, there is rudimentary textual discourse on photo/videosites in the form of memes and screenshots, but people write and the web is a hypertext-medium, still. Stuff like Substack and all those newsletters clearly point into this direction, the new-ish PKM-stuff too. I always think of reddit then, which is growing constantly, year by year, with barely anyone taking note. See also: A Text Renaissance
Did a Random Person on Twitter Name the Latest COVID Variant ‘Centaurus’? Yes, a rando on Twitter named the latest Covid-Variant. Power to the people, I guess.
Generation of genome-edited dogs by somatic cell nuclear transfer If one animal on earth is already genetically edited af then it's dogs. Maybe, actually, we can lift some of the pain traditional breeding has inflicted on some species (german shepard backpains for instance) by precise genetic engineering.
We've pulled videos from human brain activity for some time and now we are doing the same with puppers: Through a Dog’s Eyes: fMRI Decoding of Naturalistic Videos from Dog Cortex.
Yes, there are people researching how to downsize Matt Damon: The Incredible Shrinking Man – Researching the implications of downsizing the human species to better fit the earth. From there: Japanese Miniatures: The 1/8th Project. Great stuff if you think, like me, that "The incredible shrinking man" is one of the greatest SciFi movies ever.
I left a long-ish comment on The New Numbers on Music Consumption Are Very Ugly by Ted Goia:
I follow this phenomenon for quite a while now - i nicknamed it "The Big Flat Now" after the headline of a very good article about this.
I have two theories, one is about the web and one about technology and it's not very sophisticated either: a) The internet is an archive and as such we are constantly looking into the past while interacting with it. b) Digital technology, especially used as a tool to explore creative spaces, aesthetics and sound, has explored every space of sound there is. Digital sound creation contains all traditional sounds via sampling AND adds pure electronic sounds to it, and all of those are largely explored, leaving only arrangement and composition as a field to innovate on. Even using stuff like a turntable as an instrument is an anachronism now, DJs use turntables creatively for nearly 50 years now.
I remember the 80s (pls forgive me for even mentioning them, I'll return to them later again): I remember that we had a vast explosion of styles, new soundscapes, new arrangements, new structures nearly every year, Punk went to Hardcore to New Wave to Goth, simultaneously electronic music exploded into House, Techno, Breakbeats, while Hip Hop stole the show with new structures and even a new way of vocalization (rap). An innovative explosion like then is unthinkable today.
Just look at electronic music, listen to any modern dance track in a club, preferably something from the underground, but not experimental: You will not find any difference in sound from 2002 or 2022. This is mindblowing for a field in which production itself, digital sound creation and arrangement, gives you total freedom.
My guess is that democratization through digital sound creation created one last explosion of innovation which pretty much explored all soundscapes there are.
And the effin 80s. You would be mistaken to think that the stranger things wave of nostalgia started a few years ago. Actually, the 80s nostalgia wave started around the year 2000 with releases on DJ Hells Gigolo Records (Fischerspooner). The first retro electro track i can recall was i-Fs "space invaders are smoking gras" which was one of the first to cite 80s electrostuff. Dopplereffect from Detroit also started to cite 80s electro influences back then. The Lust for the 80s started at this point and never ended. This is not normal.
Maybe you are right and these are signs of a broader decline, but I'm afraid it's deeper than that. I think, as described above, that we created technology, then democratized them, which simply eluminated creative spaces that took ages to explore in the generations before in a very short amount of time.
Also i'd like to add a thought that, while innovation itself seems gone, the quality of new music has increased. If you spend a day on Bandcamp and give new bands a listen, you will find hundreds of bands whichs songwriting skills equal Lennon/McCartney. Maybe innovation got replaced by refinement of skill while the output got atomized by technological fragmentation. Both together would explain the phenomenon.
For old farts like us this may look bad, but, you know, as long as the kids are dancing on their video app, i'm optimistic for the future.