If you had a Fantastic Summer.
Is Lilliput being off the coast of wonderland relating to the size discrepancies? “Eat me/Drink me?”
One of my favorite hobbies is thinking about the fucked up implications of this fantasy world map my parents got me for christmas
[Image ID: photo of a map. On the left side of the map is Middle Earth, with the Shire and Mordor labeled. To the direct right of Mordor is Whoville.]
RIP to two of my favorite shows of recent years. You will join the spot in my brain I’ve reserved for Infinity Train
The Owl House 🤝 Final Space
Ending/getting cancelled at a third season (unfortunately), despite being popular with all
This is potentially life saving information everyone should know.
To everyone who was there for final space, please watch Godspeed. A spiritual successor that was crowdfunded and taken shape to be a unique and amazing pilot now available on YouTube. And if you like this pilot, there’s an opportunity to make more.
Olan Rogers has another Kickstarter up called Godspeed: The Series and it has already reached its goal of $100,000.00. But the stretch goals are for more episodes and more content from the creative team and the creator-first studio: Lightfold.
It has been a rough year for animation, but projects like these can’t help but fill me with hope.
More about long COVID by a young person with long COVID. Inspired by the stories of hundreds of COVID long-haulers.
A sequel to this comic on my long COVID symptoms.
Ko-Fi
(ID under the cut)
Keep reading
So, you may have heard that Tumblr is using people’s art to train AI. It does this by taking any images people upload, and feeding them to the AI for training.
So I made a little thing you can use to defend against this. It might not be SUPER effective, but it’ll at least slow things down and hurt the overall effectiveness of AIs trained on it.
This is my anti-AI image. It’s designed to clog up the training of the AI by being mass reblogged / uploaded.
To understand how this works, you first have to understand how AI image generation works. The AI is trained on a massive amount of images, and learns to reproduce images that are similar.
Now, we want to make a significant portion of the AI’s training data consist of this image. If we do this, then the AI will struggle to generate art effectively, as it won’t have as much viable training data. Instead it’ll be more likely to just generate an image similar to this one.
Reblog it to spread the word, but be sure to take a screenshot or download the image instead, rather than just reblogging. That way it counts it as a new image, and feeds it to the AI again.
I would say you could be somewhere better than here, but I don’t have much room to talk
444 posts