My public library has two copies; I’ve placed one on hold for myself.
If I am remembering correctly, Matthew Reinhart (in that video) and Robert Sabuda (another well-known pop up book artist) were at an event I attended in NY some years ago. I remember talking to Robert Sabuda for a bit. That sort of paper engineering is interesting. I know there are some process videos around. “Paper Paul” posts some stuff about the mechanisms he develops on Instagram ( https://www.instagram.com/paperpaul01/ ), for example. It’s a pretty different way of thinking from the origamic architecture / kirigami stuff.
Midjourney released a new version of their image generation algorithm. It’s more specific now, it allows you to ask for more particular images.
“a wooden carved figurine of a dog on a desk, ultra realistic, cinematic, --ar 16:9”
Ok not specific enough. Lets tell it we want it to be a black doberman:
OK more specific, how about at night with a fireplace:
All of which is fine, but it’s worth mentioning that these were all my first prompts. None of these are cherry-picked.
So lets see what we can get for the glowforge:
a vector illustration of a man with a cowboy hat, smoking a pipe, plaid shirt, black and white, high contrast --ar 16:9
Top was cropped, typical AI stuff there, so I ditched the 16:9 aspect ratio:
Interestingly, he had a cig both times, not a pipe. This is the first one that missed any details.
Then I tried a star chart, just to see:
“an antique star chart, black and white, high contrast, detailed, circular”
As with all of these AI things, the possibilities are endless.
a pencil sketch of a childrens book illustration of a little girl holding a balloon, walking an alligator on a leash, forest path, circular, cute
Children’s book illustration: check
Little girl holding a balloon: check
Walking an alligator on a leash: check
Forest path, circular: check
Cute: debatable
Glowforge will have to update their system to keep up, magic canvas looks pretty primitive already.
it’s interesting that he’s smoking a cigarette, but his right HAND looks like it’s holding a pipe bowl.
That “vector illustration” phrase is amazing at making stuff I can trace in inkscape and vectorize.
I wonder what astrolabe would bring up… I don’t care for all the hubbub around this AI nonsense (and I’ve been in the tech industry since the 80’s) but the star chart made me wonder.
I’ve been steadily working on my own design - based on traditional Hellenistic influences - for a number of years. Guess I should put it back on my to-do list before my GF finally bites the dirt.
Astrolabe with no details at all. I just told it to “/imagine an astrolabe”:
For kicks I asked for a realistic detailed orrery in a library, bokeh:
(Midjourney runs 4 images at a time, you can then choose which ones to iterate on or blow up)
As a vector illustration:
As a notebook sketch in the style of davinci:
One blown up:
These are all fantastic. Thanks for sharing prompts
The alligator is cute at least!
V5 looks amazing. There is SO much improvement in so short a time I’m truly astounded.
For contrast here’s the same prompt about a dog figurine in version 3, pulled 9 months ago:
And now:
It’s night and day.
Huh. I knew the Amazon was the biggest river but not that much bigger. It has more than 10x the flow of the Mississippi. Crazy!
I didn’t know about this until last week, but there is AI to help match up your resume to jobs. (There’s a free plan or you can subscribe for more scans.)
How good is it working for you so far? Poorly worded. Do you like what it has suggested for you so far?
I’m sitting here feeling disappointed, to be completely honest!
This is likely why I haven’t gotten more responses to my job applications.
I said the exact same things but just used different words and it doesn’t show as a match.
I also emailed the career coach asking her: I thought the point was to just have one resume and then update the cover letter for each job applied for. So I’ll see what she says.
But this explains a lot.
And I’m feeling conflicted! I totally get the need for this and making things easier to fill roles faster and all of those things. From a process perspective, this makes total sense. The time to fill is SO much less I’m sure. But then the human element is being removed. Like I said, very conflicted.
I had seen this one on Youtube before. It was impressive then also.
Robotically recognising the human element is very difficult. The Manager sends HR a skill set using words that HR does not fully understand based on their expectations of what HR can find. And HR looks for folks that they can defend that they looked at, and therefore anything odd is rejected. Really creative folk will usually appear odd.