America will regret this. 99 Senators across both parties rejected an earlier version of it.
It will leave our children and mentally ill vulnerable.
It may end privacy as we know it.
It will escalate the chance of uncontrollable and unaligned AI.
And it will set a precedent for government overreach that will endure for decades.
SCOOP: A Powerful A Dark Money Group Is Secretly Funding High-Profile Democratic Influencers.
Big liberal content creators are being paid up to $8,000 per month to push party messaging.
@icreatelife@XestyTV Thanks for being so open with sharing prompts. I do have a question about the training, are you training on individual photos and then prompting them in together or are you using photos of products together?
The future of humanity is being written by a few thousand AI devs with relatively no oversight or public input.
I'm not pro-regulation since fastest path (less regulated) wins. Just think it's fascinating how the values of all humans are being aggregated by a few thousand people.
Making the Simpson’s living room a reality.
Another controlnet experiment with Stable Diffusion. It was really tricky balancing the different models (and their weights) with this one. I used scribble, hed and depth, plus img2img.
Researchers release ControlNet: control large diffusion models using inputs beyond just text prompts.
You input:
edge maps, poses, etc + a text prompt
to output: highly specific images from models like Stable Diffusion.
Here's how people are using ControlNet 🧵:
We remade a 2 second snippet of this classic GTA V cutscene using Stable Diffusion, Dreambooth, ControlNet, VideoLoopBack, Davinci Resolve & Adobe Premiere. 🧵 (1/8)
The diffusers 🧨 library just did a release incorporating ControlNet, it runs so fast! 🏎️💨
Blog: https://t.co/LOftkb95kh
Colab: https://t.co/ADxxFX04xc
Here's a simple (< 100 lines of code) app to run #ChatGPT question-answering on any uploaded document (using @langchain DBQA w/ ChatGPT API): https://t.co/dGGY69YHDz
Voy a compartir un pequeñísimo trozo de código para poder usar los mapas de profundidad de @clipdropapp en #ControlNet. Como se ve, son mucho mejores que con MiDaS.
Vaya por delante que no soy programador, pero esto funciona si lo pegáis en un Collab.
https://t.co/r8W65UiJXh
"You did not seriously think that a hobbit could contend with the will of Sauron?"
Keyframe generated with #ControlNet#img2img using the canny edge detection model & animated using #Ebsynth#aiart#stablediffusion
The simple fact is that we just don’t have enough leverage to convince billion dollar companies that the ethical thing is to NOT optimise profits and minimise expenditure by using these tools (during an economic downturn)
Introducing AI Training
You’ve seen the magic. Now, make it your own. Train your own custom AI model in just a few minutes and turn anything into everything.
Available now: https://t.co/Y7iO2LlkHw