Login
Google Docs will have it, Microsoft is brewing it, PhotoShop is doing it, so is zoom…it’s high time that the advanced technology of CogDogBlog hop on the artificial intelligence inevitable train.
As demonstrated this week at the Reclaim Open 2023 conference, behold the new generation of blogging powered by AI Levine:
This is all the mad creativity of Michael Branson Smith, who along with Jim Groom (he has a blog, right?) contacted me a few months ago with a request for a short video he could us to train some AI that would generate some manifestation of me as a speaker at the conference.
How could I refuse? Michael is one of he most creative and gracious people I have gotten to know in this ed tech game, initially through DS106, but hey we have spent time together hanging out at his home in Brooklyn.
Michael shared a few prototypes, all of them definitely weird with artifacts, but still sounding like me. Each time he asked if I was okay with it, I said I trust you completely, MBS, do as you see interesting.
I did not travel to Reclaim, but enjoyed the online perch for the conference (I have to say that the Reclaim Hosting set up for live streams is one of the best I have seen. Live streams are viewed adjacent to a Discord chat, with its TV Guide like directory guide at the top.
Michael shared some details of his intent to learn something about AI by doing it himself (DS106 spirit showing). I am forgetting the tools used (it’s in the video and hopefully a future MBS blog post). The original video I sent did not perform well, so MIchael relied on one from the YouTube I recorded in 2019 as an intro to students at a Coventry University class I was doing something with. Meet AI Levine’s puppet:
Michael did say generating the AI audio was the bigger challenge- i believe he used ElevenLabs for text to speech. AI Levine’s blabbering was based on a ChaGPT query to talk about SPLOTs derived from 4 blog URLs MIchael gave it to chew on. The “hallucinating” is readily apparent in the history portion but the gist of it is maybe palpable.
As was obvious to most in the room, the video generated was weirdest with the mouth, and Michael ended up blurring AI Levine a bit in post editing.
In emails the day or two before, Jim Groom suggested some kind of irate outburst from me in chat where they could then bring me in to the room (via a whereby.com link). That would be fun, and of course, anyone ought to get bent out of shape by being deep faked. That’s too easy.
My play was then to act dismissive, to assert that I am actually a supreme AI construct:
For a long time I have been “training” myself by reading the entire internet. Everything I have made, written, created was generated from a complex process applied to this data. SPLOTs could only have been generated this way.
(no idea if this is what I actually said)
All in all it was quite enjoyable, as was seeing on the live stream and the whereby video much different views. All surreal.
Thanks Michael for creating this fallible facsimile in every AI sense of hallucinatory mockery and thanks Reclaim Hosting for putting this on the program.
For me, this again, puts on the table some impressive technical feats to produce, well the equivalent of flavorless mind porridge. Generative content stays within the guardrails of statistic probability yielding an appearance of pseudo meaning. Downtown Dullsville.
What I’ve not seen AI do is to make meaning and connections between the unlikely related. For example, in my regular training on internet content (aka reading blogs), my own neural pathways ran off in different directions in Martin Weller’s fab post Yankee Shed Foxtrot where he weaves also different, seemingly unrelated topics- a book on “shedworking” and a song by Wilco.
Reading this as a human means I do not regurgitate sentences based on some vectors of word probability. For me and my [Un]Artificial [questionable] Intelligence, a series of other un stochastic connections. My comment on Martin’s blog was one of those that really should have been a post here…. so I will steal my ow words:
What you demonstrate in action here, oh ye faithful blogger, is something AI can’t- make connections between ideas and things that are statistically out of range. Yes, we have neural networks in our grey matter, but the computerized ones that borrow the name cannot (well Stephen will assert its eventually possible) to connect this way.
What foir me your post does is trigger other connections. The idea of “shedworking” (the name new but the idea resonates) for some reason reminds me of a book Scott Leslie recommended (see how connections, work, I can remember it was on a camping trip where we canoed to an island in BC) a book called “Shop Class as Soulcraft” http://www.matthewbcrawford.com/new-page-1-1-2 that honors the value of work we do with our hands, it need not be mechanical, or done in a shed, but is a testimonial to the ideal of craftspersonship and what it does for us (I can extend it to creating web pages by hand).
As far as an academic metaphor (you are cornering the market!) I love the idea of conceptually makeing space and time for crafting. And yes, institutions are not providing this per se, but there is a two way play here- us as individuals have some responsibility for also not making the time and place ourselves for doing this. People who design their shedworking places are not just given them by an organizational entity, they take it on themselves. We as individuals have a stake in this.
And then for the music metaphor, I dont know if you are into the Song Exploder podcast https://songexploder.net/ It’s a brilliant concept where musicians break down a finished song into its origin story, how the tracks came together, layer by layer. It’s a fabulous metaphor (to me) of what we do little of in our work (outside thr last ferw bloggers standing) of not just pumping their finished work, but the craft of how it was made, how it emerged, if you will from their audio workshed.
I wanted to launch an idea like “Course Exploder” where educators could break down a lesson or a OER or course design in the same way.
What I enjoy about Sound Exploder is that I get these stories for songs and genres I never listen to (sort of like how university requirements had me take courses I would have never chosen on my own interests). I was just listening to the episode where Seal talked about the evolution of the mega pop hit “Kissed by the Wind” (heard the song plenty but cant say its in my listening queue). It’s rivetting to hear him talk about its origin, but there is a bit in the opening where he talks bout his start before being a mega star:
“I was living in a squat. I didn’t have any money and I was just basically getting my act together, trying to figure out who I was musically. And “Kiss From A Rose” came out of that period, when you sort of do things not for any other purpose than because that’s what’s coming out of you. I had no experience in the studio, in a proper recording studio, at that point.”
It’s that line– “when you sort of do things not for any other purpose than because that’s what’s coming out of you” that connects in my head back to Shopcraft and sheds and why we even bother trying to do stuff.
My comment on Martin Weller’s Yankee Shed Foxtrot blog post
I may be on the far end of the bell curve of what people want from the world these days, but I crave more the intelligence that makes meaning from dissimilar ideas, not just spitting back the statistically relevant blandchastic stuff.
You heard it all from me, the always A.I. Levine.
Featured Image: Combination of one frame from the Reclaim Open 2023 presentation The SPLOT Revolution will be Artificial with a conference logo and a SPLOT logo, call the whole thing CC BY.
While there is plenty of academics undergarment wadding over AI generative text (and please stop referring to it all as ChatGPT), I was first interested and still in the generation of images (a year ago Craiyon was the leading edge, now it looks like a sad nub of burnt sienna).
Get ready for everything to get upturned with Adobe Photoshop’s Generative Fill, now in beta. I spotted it and some jaw dropping examples in PetaPixel’s Photoshop’s New ‘Generative Fill’ Uses AI to Expand or Change Photos but was drawn in more by a followup post on So, Who Owns a Photo Expanded by Adobe Generative Fill? This gets into even more muddy, messy, and also interesting (time curse like?) waters.
That latter article has some really fabulous pieces of ?? Extended Album Covers found originally in the twitter stream of Alexander Dobrokotov. I’d post the tweets here for you to see, but twitter broke the capabilty to embed tweets.
The concept is rather DS106-ish a central image of an actual album cover is embedded into a much larger imagined scene (see the Petapixel post for the examples) where all the imagery around was created with this new Adobe Photoshop beta feature.
I have seen this many times with AI, you see these jaw dropping examples that imply someone just typed a phrase in a box, clicked the magic bean button, and it popped out. Most of the time, if you can find where the “making” of is shared, you will find it took hours of prompt bashing and more likely, extra post processing in regular Photoshop.
Hence why my attempts usually look awful (?)
Now I could just share say image (like the Katy Perry cover of her sleeping in soft material that turns out to be a giant cat) and say, this is cool! But I always want to try things myself. So I downloaded (overnight) the beta version of Photoshop.
The way it works is you use the crop tool to create space around a source image. This fills with just white. But then you select all that blank space along with an edge portion of the seed photo, and watch something emerge. In many ways it’s impressive.
I started with my iconic Felix photo, the one I took on his first days with me in 2016, the one I use often as an icon.
In Photoshop Beta, I enlarged the Canvas to the left a lot, and a little above, and let the magic go to work. Perhaps this is not the best example, since my truck in the background is blurred from depth of field effect.
Not quite magic.
That’s a rather awkward vehicle there. And since AI has no concept of a porch rail, it would likely extend those posts Felix is peeking through into the stratosphere.
I decided to try again, and added a prompt to the generative gizmo saying “Red truck towing a camper”
Well, that looks awkwarder too. But it generates something.
I took another stab, thinking how it might take on extending a wide landscape that is well known. This is tricky because if one knows something of Geology, they canyon to either side extends to a broad plateau.
I did one first where I went about 50% wider on each side
It certainly continues the pattern, and is not all that weird. You do get 3 variations, this one is about the same:
It’s odd, but not really too far from pseudo reality. I riffed off of this version, adding again another chunk of empty space on either side. Now its getting the geology pretty messed up and messy.
These are just quick plays, and there are also the other features in the mix to add and remove elements.
This definitely is going to change up a lot things for photographers and digital artist, and what is real and what is generated is getting so inter-tangled that thinking you can separate them is as wise as teetering off that canyon edge.
But getting back to the Petapixel leading headline, “So, Who Owns a Photo Expanded by Adobe Generative Fill?” oh my is ownership, copyright, and licensing going to get mashed up too. So all of those creative album cover expansions? It’s starting with copyrighted material. But the algorithmic extension, is that so far changed to raise a fair use flag? Heck, I have no idea.
At least if you start with an open license image, you stand on slightly less squishy ground.
I’m going back to my shed to tinker (that’s for Martin).
Featured Image: 100% free of AI!
FreshRSS 