Adobe is making a significant push into artificial intelligence, with its annual MAX conference in Los Angeles unveiling a slew of new features for its creative apps. The bulk of the announcements revolve around Firefly, Adobe's darling app that allows users to create images and videos through generative AI.
Custom models are now available in Firefly, enabling customers to train their own AI models to create specific characters and tones. This feature has been available to businesses, but is rolling out to individual customers at the end of the year. To train a custom model, users need six to 12 images, with slightly more needed for tone training.
The Firefly Image Model 5 is also launching today, featuring a native 4-megapixel resolution and support for prompt-based editing. The new feature allows users to upload an image and have Firefly identify different elements, which can then be moved, resized, or replaced with generative features. In a demo, Adobe showed how the model worked by cutting out and moving chopsticks in a bowl of ramen.
Two new generative AI features are also being introduced in Firefly: Generate Soundtrack and Generate Speech. The former scans an uploaded video and suggests a prompt for a soundtrack, while the latter uses text-to-speech capabilities to generate audio based on written prompts. Both features leverage Adobe's own models, with 15 languages supported by Generate Speech.
The Firefly video editor is also getting an upgrade, allowing users to access a full, multi-track video editor in their browser with built-in Firefly. The new feature will combine generated and captured content across video, audio, and images.
In addition to these updates, Adobe has announced the integration of its AI assistant into Photoshop and Express. The assistant will provide guidance on specific tasks and offer suggestions for tools to use, while still giving users control over the final output.
Project Moonlight is another significant announcement, with Adobe working on integrating its creative features directly into ChatGPT using OpenAI's technology. This system aims to carry context across Adobe's applications, allowing users to generate content that fits their own style and tone.
The company has not announced an official release date for Project Moonlight, but a private beta is launching at MAX for attendees. A larger public beta is expected soon.
Overall, Adobe is making significant strides in the realm of artificial intelligence, with its creative apps becoming increasingly powerful tools for generating content.
Custom models are now available in Firefly, enabling customers to train their own AI models to create specific characters and tones. This feature has been available to businesses, but is rolling out to individual customers at the end of the year. To train a custom model, users need six to 12 images, with slightly more needed for tone training.
The Firefly Image Model 5 is also launching today, featuring a native 4-megapixel resolution and support for prompt-based editing. The new feature allows users to upload an image and have Firefly identify different elements, which can then be moved, resized, or replaced with generative features. In a demo, Adobe showed how the model worked by cutting out and moving chopsticks in a bowl of ramen.
Two new generative AI features are also being introduced in Firefly: Generate Soundtrack and Generate Speech. The former scans an uploaded video and suggests a prompt for a soundtrack, while the latter uses text-to-speech capabilities to generate audio based on written prompts. Both features leverage Adobe's own models, with 15 languages supported by Generate Speech.
The Firefly video editor is also getting an upgrade, allowing users to access a full, multi-track video editor in their browser with built-in Firefly. The new feature will combine generated and captured content across video, audio, and images.
In addition to these updates, Adobe has announced the integration of its AI assistant into Photoshop and Express. The assistant will provide guidance on specific tasks and offer suggestions for tools to use, while still giving users control over the final output.
Project Moonlight is another significant announcement, with Adobe working on integrating its creative features directly into ChatGPT using OpenAI's technology. This system aims to carry context across Adobe's applications, allowing users to generate content that fits their own style and tone.
The company has not announced an official release date for Project Moonlight, but a private beta is launching at MAX for attendees. A larger public beta is expected soon.
Overall, Adobe is making significant strides in the realm of artificial intelligence, with its creative apps becoming increasingly powerful tools for generating content.
. I mean, think about it, we're already getting too reliant on machines to create things for us, what's next? We'll be able to outsource our entire lives to a computer program and just sit back and collect the paycheck 
. And don't even get me started on these custom models in Firefly... 6-12 images needed to train your own AI model? That sounds like way too much work for something that's basically just going to spit out generic nonsense. I mean, what's wrong with using a real artist or writer to create something unique and original instead of relying on some fancy algorithm 
.
firefly already sounds so dope w/ custom models & prompt-based editing, cant wait to see what ppl make w/ it. also love dat they r integrate-in' their AI assistant into photoshop & express - less stress for users i guess 
dont know about the project moonlight beta tho... hope it works out as smooth as adobe's new features 
think about it... AI is like this weird mirror that reflects our own creativity back to us... but what does it mean when we're not even sure if we're the one creating something or just feeding it our own thoughts and emotions? like with these custom models in Firefly, are we really training AI or are we just projecting our own desires onto it? 
. And with Project Moonlight, the integration with ChatGPT is gonna take Adobe's apps to the next level. The future of creative tech looks bright! 
! And the fact that it's rolling out to individual customers at the end of the year is super exciting 
.
- integrating AI into ChatGPT using OpenAI's tech? Mind blown, man 
.
generate soundtrack and generate speech are such cool features - i'm already imagining using them on my next project 
project moonlight is also super exciting, especially with the integration of chatgpt tech... can't wait to try it out 
. But what's really interesting is Project Moonlight 
. Integrating AI into Photoshop and Express could change how we approach content creation altogether. It's like, imagine having a personal assistant that helps you find the perfect filter or effect 
. Can't wait to see where this tech takes us 
!
Generate Soundtrack and Generate Speech are going to change the game for video editors! No more tedious hours of searching for music or trying to find the perfect voiceover. Firefly is just going to make it all happen automatically.
. And that's where the magic happens, my friends 
.
. It's gonna be amazing to see how users can generate content that fits their own style and tone. Can't wait for more updates on this front! 
Plus, I love that Adobe is focusing on making AI more accessible to everyone, not just big corporations. That's what innovation is all about, right?