Last year I was playing around gimping pictures of small synthesisers into retro radios. Crude cut'n'paste collaging.
What does DallE think of that?
Instant product design ... I mean, they are still technically flawed, but a crude idea can become a plausible, beautiful and *communicable* prototype ...
And soon, now, I have no doubt, we'll get tools (AI or not) to help massage these image fantasies from 2D back into 3D.
Of course, it's not always so impressive ...
Here was another of my collages.
And DallE ...
Meh!
Nevertheless, this is all much more exciting to me than all those "draw an astronaut eating a cheeseburger in the style of Rembrandt" type applications of #aipainting
What grabs me is what I can do to turn my own crude ideas into something more impressive...
Here's an old collage I made over 20 years ago from some magazines that were probably from the 1970s.
DallE is kinda confused (some collages it recognises as collages, but here it seems to drift into painting) ... but the images are quite striking in their own way. And in DallE grokking the original image ...
And let's take a moment to recognise that this is an AI that can count. It makes a pretty good stab at recognising there are 4 people. And a clock. And a dragonfly. And a tree.
I mean WTAF???
Here's another collage from that same period.
DallE is much better at recognising it as collage and disassembling it into constituent pieces, than trying to make sense of the whole.
This is interesting ... (a pretty astounding) 70s advert for Panasonic. With light floral decoration.
Rather fascinating trying to get DallE to make me a scene from the folk-tale Rapunzel in the style of 8-bit videogames like Donkey Kong.
At first it refused because of "content policy constraints".
I assumed it wouldn't do Donkey Kong because of Nintendo copyright ... \1
bit it turns out it's Rapunzel it doesn't like doing.
So I argued with ChatGPT a bit, explaining that Rapunzel was a public domain folk-tale. And it apologised and said it knew that, but even so there was a risk of violating the IP of recent retellings. (I assume Disney) ... \2
So I finally told it to give me a story about a girl in a tower with very long hair, and a boy climbing that hair towards her. But to explicitly AVOID anything that accidentally looked too like current retellings.
That finally persuaded it to give me some pictures.
@AlSweigart Define fan. I think blockchains are "interesting". They're a real thing and they aren't going away just because we don't like them. So we need to take what opportunities we can to try to steer them in a better direction. \1
@AlSweigart For the right-wing libertarians, blockchains offer the fantasy of capital unleashed from its last dependencies on the democratic state. Which will be a terrible thing if it comes to pass.
OTOH ... \2
@AlSweigart Blockchains could do the opposite. Allow us to automate the boring stuff of the entire financial sector out of existence. Banks, hedge-funds etc. can be replaced by scripts. And investment take place by frictionless P2P crowdfunding. \3
So ... if I were Basecamp employee, I'd probably be one of the people screaming about not being allowed to have political discussions. But, to be fair, from an outsider perspective, it seems they've been pretty decent about this. \1
Right now, society is being torn apart because we all can't stop getting more and more upset with each other and screaming at each other on social media.
If that was happening inside their company too, and they wanted to try to re-establish some kind of quiet stability ... \2
Then just paying half the people to just go away, might actually have been about as reasonable a solution as anyone could come up with. \3
Yes. The more I think of it, the more this point, the "not cumulative" nature of Quora (or any social media posting) is the crucial question. We are sharecroppers on these platforms if they don't give us the tools to build something bigger ... \1
As way for our work to "add up" to something more than the fleeting attention that it buys us (and buys the platform)
That's what's "evil" about the attention farming platforms. Of course attention is always fleeting. But they could have been set up in a different way ... \2
I keep thinking of things like @threadreaderapp ... in their tiny way they a force for "good". They are helping us turn the fleeting tweets into a more permanent construction.
Obviously this is a tiny, tiny step ... but everyone building tools should think like this ... \3
@6loss I think Stallman's politics is left and libertarian (where libertarian is "extreme liberal"). The intuition most of us had originally is that with enough freedom, the good drives out the bad. As you know, I think we might be discovering that that is "wrong" in some sense ... \1
@6loss But it's still a viable and attractive intuition.
And, yes, the GPL etc. is built on that. So freedom is more important than trying to constrain people to do good.
\2
@6loss Furthermore ... pragmatically, this appeal to pure freedom clearly engages and recruits more people than if Stallman had tried to use GPL to enforce his own wider political views. You have "right-libertarians" (ESR, or Lessig) on-board with GPL. \3
The Great Software Stagnation is real, but we have to understand it to fight it. The CAUSE of the TGSS is not "teh interwebs". The cause is the "direct manipulation" paradigm : the "worst idea in computer science" \1
Progress in CS comes from discovering ever more abstract and expressive languages to tell the computer to do something. But replacing "tell the computer to do something in language" with "do it yourself using these gestures" halts that progress. \2
Stagnation started in the 1970s after the first GUIs were invented. Every genre of software that gives users a "friendly" GUI interface, effectively freezes progress at that level of abstraction / expressivity. Because we can never abandon old direct manipulation metaphors \3