I love this initiative from Kagi (the ad-free, paid search engine)! They now bump up relevant results from the small web, like blogs and personal websites. And there’s a fun way to stumble upon new sites. A great companion to Marginalia Search. Yay for all efforts to humanize the web!
Kagi Small Web offers a fresh approach by promoting recently published content from the “small web.” We gather new content, published within the last week, from a handpicked list of blogs and surface it in multiple ways[.]
One week of the photoblogging challenge already! Is this panorama captured by a human 🧔♀️ or dreamed up by a text-to-image model 🤖?
Here’s the website with a teaser for Hayao Miyazaki’s next film, The Boy and the Heron.
All is well in this realm, but dream 🤖 or reality 🧔♀️, can you tell?
Fellow microbloggers, I’m sorry I lied. Two out of my five challenge entries so far were generated by a diffusion model. A couple identified the dreamed-up sunset. No one (!) suspected the fake clouds.
The rest was shot by me. Thanks, everyone, for participating so far. Keep the guesses coming! 😊
Well, it’s no secret that I love me a lush forest. But is this one documented by your fellow human 🧔♀️ or just a latent diffusion model’s 🤖 dream?
It’s day four already, the prompt is orange, and you know the drill by now: was this pic snapped by me 🧔♀️ or dreamed up by a text-to-image model 🤖?
This potholder might look ordinary, but it is very precious to me. Or maybe it doesn’t even exist. What do you think – is this pic snapped by me 🧔♀️ or dreamed up by a text-to-image model 🤖?
The prompt for day two of the September Photoblogging Challenge is buildup. What do you think – is this pic snapped by me 🧔♀️ or dreamed up by a text-to-image model 🤖?
The prompt for day one of the September Photoblogging Challenge is abstract. What do you think – is this pic snapped by me 🧔♀️ or dreamed up by a text-to-image model 🤖?
Snapped by me or dreamed up by a text-to-image model?
I’ll post a picture every day for the September 2023 Photoblogging Challenge. To keep it interesting, I will sometimes use my computer to manifest a “photo.”
And you – yes, you – are welcome to guess whether it’s snapped by me 🧔♀️ or dreamed up by a text-to-image model 🤖.
It might be fun! I, for sure, will learn a lot, and maybe we will all be better at spotting fakes at the end of the challenge.
Here are some constraints I’ve set up for myself, besides the rules stipulated by Micro.blog.
- Real photos must be snapped by me.
- It’s okay to use photos from my archive.
- I will write the prompts when generating images.
- Text-to-image models must have a permissive license and run on my machine. No cloud allowed.
- Cropping, color adjustments, and light retouching are fair game for all pics – real or fake.
- I’m allowed to lie and deceive on the blog for the duration of the challenge.
- But I will reveal all the lies and generated images when the challenge is over. Full transparency.
So, for the rest of September, don’t believe everything I share here. And start guessing! Here’s the pic for day one.
Here’s more than 500 CSS loading animations from Temani Afif. 🤯
Our Flag Means Death is incredible and as close as we will get to a Monkey Island TV show, according to Ron Gilbert. That’s all I need; I’m in! 🏴☠️
Oh, man, now I want to find an old knitting machine and start hacking on it. This is just marvelous! 🧶
Retreat is an independent electronic machine knitting practice based in Berlin. All items are crafted by a real human with the help of a hacked domestic knitting machine from the 90s.
This reminded me of the don’t write it down advice in REWORK. (via @grumpygamer)
But people will say, "Do you keep a notebook?" and the answer is, I think a writer's notebook is the best way in the world to immortalize bad ideas. My idea about a good idea is one that sticks around and sticks around and sticks around.
These 40 hats by Anders are stunning! I think I know LeChuck’s hat, Ash’s cap, Samus’ helmet, and about two dozen more. How many can you place? 🤠
The history of video game hats - One "hat" per year from the past 40 years.. see if you can guess them all #pixelart
Simon Willison shares how to make large language models work for you. It gets a bit technical and programmy a couple of times, but most of the talk is pretty accessible to anyone curious about LLMs. If video is not your thing, read the excellently annotated transcript instead.
If Ben were to voice Mario in the future, I wouldn’t be too sad. (via Tobias)
My official audition to play #Mario.
I’m on my way to see Oppenheimer. 20 meters away, my niece exit from the cinema. She’s just seen Barbie with two of her friends. I’m not aware of her at first. But then I hear her. Loud and clear. Shouting.
Svampen!
That’s what she calls me. She runs towards me and jumps into my arms. I do my best to catch her and keep her there for a long hug. It feels fucking great. And I savor the moment because I know it won’t be long before she’s way too cool to greet her old uncle like that.
And maybe it’s for the best; I won’t be able to catch her without falling backwards for much longer.
Wordpress.com promises to secure your online legacy for a century (if you pay up $38,000). I love the idea! It’s just too bad one won’t be around to see if they fulfill their end of the deal. A lot can happen in a hundred years. (via ma.tt)