- In The Loop AI
- Posts
- Midjourney Or Stable Diffusion: Which one is the better text-2-image generator?
Midjourney Or Stable Diffusion: Which one is the better text-2-image generator?
Algot Persson
July 06, 2024
In Today's Article:
How does Midjourney V6 compare to Stable Diffusion?
First driverless car to be pulled over by the cops
Whatsapp to add personalized AI avatar feature
Read time: 1.5 minutes
Midjourney and StableDiffusion Compared
There have been 2 heavyweights in the image generation space for a year, both holding their own weight. Midjourney - the discord native platform has released its V6, while StableDiffusion has its Stable Assistant and different models you can download on your computer.
Midjourney prompting can be done through text-2-image quite easily, although it can require some knowledge of different styles if you are interested in a certain look.
Stable Diffusion prompting is objectively harder but can achieve the same results as Midjourney with enough iterations.
If you want a good result without learning either system, Midjourney is better for beginners.
Midjourney typical interface:
Stable diffusion typical interface:
Pricing:
Midjourney: around 30 dollars for a minimum membership, includes 15 hours of fast GPU generation.
Stable Diffusion: Free as long as you use a custom version on your computer. API costs are higher.
Today’s News in AI
A Waymo self-driving car in Phoenix was pulled over by police after running a red light and entering an oncoming traffic lane on June 19. Bodycam footage released by Arizona Republic shows the vehicle weaving through traffic before stopping in a parking lot. Dispatch records describe the vehicle as having "FREAKED OUT," and the officer couldn't issue a citation. Waymo stated that the car encountered confusing construction signage and briefly entered the wrong lane, remaining there for 30 seconds. There were no passengers in the vehicle. Waymo has recalled its software twice this year and is under federal investigation for safety.
WhatsApp is developing a personalized AI avatar generator called 'Imagine.' Found in the latest Android beta (2.24.14.7), this feature lets users create avatars in various settings using Meta’s AI Llama model. Users provide photos and text prompts to generate these images. The avatars can be created by typing “Imagine me” with a setting description in chats. This optional feature requires manual activation in settings, and reference images can be deleted anytime. While there's no release date yet, WhatsApp is still expanding its Meta AI chatbot and real-time AI image generation features in the US.
Reply