Get the 6th volume of the mahazine 🔥🔥🔥🔥🔥🔥

More details below 👇

👉 ByteDance launches Seedream 4.0

ByteDance has introduced Seedream 4.0, that combines text-to-image generation and image editing features, reportedly operating over ten times faster than its predecessor, Seedream 3.0. This model supports the use of reference images, native control signals, and multi-image workflows, enabling users to produce images in up to 4K resolution. Additionally, it allows for clear rendering of text, tables, and diagrams, and enables users to make edits, such as object removal or lighting adjustments, using simple text prompts.

Seedream 4.0 is accessible through ByteDance's Volcano Engine - Ark cloud platform and is integrated into the Doubao chatbot. It is also available on no-code partner platforms like FAL.ai, Freepik, and Wavespeed.ai. Pricing varies based on platform and image resolution, with costs around $0.03 per image for standard resolutions.

👉 Midjourney now lets you explore sref styles

Midjourney has introduced Style Explorer V1, an early version designed to help users navigate and experiment with various artistic styles. This new feature allows users to view thumbnails representing different styles, each generated through unique SREF codes that sample from a continuous latent space.

Users can save preferred styles by liking them and can easily render prompts using any selected style by clicking "Try Style."Additionally, the Style Explorer includes a search function, enabling users to find pre-rendered styles by entering keywords such as "photographic" or "anime."

👉 World Labs lets you navigate into 3D Worlds

World Labs has launched Marble, a limited access beta model that allows users to create and explore persistent 3D environments from images or text prompts. This new model generates larger, more stylistically diverse 3D worlds with improved geometry, enabling users to navigate freely without time limits or inconsistencies.

Marble also supports exporting generated worlds as Gaussian splats, which can be integrated into projects using the open-source rendering library Spark and Three.js. The model emphasizes consistency and style adherence, making it easier for users to build expansive environments by composing individual generations. Marble caters to a variety of creative styles, from hyper-realistic to cartoonish, allowing for iterative exploration in 3D.

👉 Lumalabs launches Ray 3, an Intelligent Video Model

Lumalabs has unveiled Ray 3, an intelligent video model designed to enhance storytelling through advanced visual generation. This new model features state-of-the-art physics and delivers videos in 16-bit High Dynamic Range (HDR) color, marking a significant advancement in generative video technology for professional studios. With its new Draft Mode, Ray 3 allows users to explore ideas more quickly and efficiently, enabling faster iterations and improved storytelling.

Ray 3's capabilities include visual reasoning, which helps it understand user intent and refine outputs for better quality. Users can annotate visuals to specify desired layouts and motions, making the creative process more intuitive. Additionally, Ray 3 can transform standard videos into HDR footage and export them as 16-bit EXR frames for seamless integration into professional workflows.

👉 Adobe launches AI Agents for customer experience management

Adobe has announced the general availability of AI agents designed to transform how businesses manage customer experiences and marketing campaigns. Powered by the Adobe Experience Platform (AEP) Agent Orchestrator, these agents can understand context, plan multi-step actions, and refine responses, enabling organizations to deliver personalized experiences at scale. Over 70% of eligible AEP customers are already using Adobe's AI Assistant to interact with these agents, enhancing capabilities for brands like The Hershey Company and Lenovo.

The newly introduced AEP Agent Orchestrator features a reasoning engine that interprets user intent from natural language prompts, dynamically activating agents based on orchestrated plans. Businesses can leverage various out-of-the-box AI agents to optimize audience targeting, streamline customer journey creation, analyze experimentation performance, and enhance site optimization. Adobe also plans to release the Experience Platform Agent Composer, allowing for customization of AI agents to align with brand guidelines and organizational policies, further expanding the potential applications of agentic AI across industries.

👉 AI Artist signs $3 Million record deal, sparks controversy

Xania Monet, an AI-generated artist supported by Timbaland, has signed a $3 million recording deal following her success on Billboard charts. The creation of Xania was facilitated by Telisha Jones using the AI music tool Suno, which is currently embroiled in copyright disputes. Monet’s song “Let Go, Let God” has gained traction, appearing on two Billboard charts, including a debut at No. 25 on the Emerging Talents list.

The deal has ignited significant debate within the music industry, with artists like Kehlani and SZA expressing concerns over the implications of AI-generated music on human creativity. Kehlani highlighted the potential displacement of human artists, while SZA voiced her opposition to AI involvement in music creation.

Innocean Berlin has partnered with the Royal Society for Blind Children (RSBC) to create the Bedtime Donations app, which enables parents to convert bedtime stories into free audiobooks for blind and visually impaired children. The app features a library of copyright-free stories sourced from Project Gutenberg and uses secure AI technology to transform recordings into accessible audiobooks. Recognizing the financial challenges faced by many visually impaired individuals, the app aims to provide inclusive access to educational and fictional stories.

The app operates in two modes: Reader and Listener. In Reader mode, parents select stories and record them, which are then processed by AI to meet quality standards before being made accessible. Gabriel Mattar, Chief Creative Officer at Innocean Berlin, emphasized the importance of accessibility and the app’s potential to enhance storytelling for visually impaired children. The project also benefits from input from accessibility experts, ensuring compliance with web accessibility standards and a user-friendly experience. With contributions from well-known voices, the app is designed to evolve, incorporating a diverse range of stories in multiple languages and genres.

_{Credits: Buzzwoo / Royal Society for Blind Children (RSBC) / INNOCEAN Berlin}

AI magazine for creatives

The mahazine is your curated guide to mastering AI as a creative professional, with research, tools, and insider insights that help you shape the next wave of creativity.

Inside this issue:

✍️ Prompt mastery: Practical frameworks to turn your ideas into powerful AI outputs.

🖼️ JSON prompting: Structure prompts with precision for better results.

🎨 Midjourney deep dive: Parameters and tricks to bring more soul to your renders.

📸 Marketing tools: Standout content and AI-driven product shots that convert.

🔥 AI trend radar: The most innovative campaigns to keep you ahead of the curve.

🦜 Visual explorations: Experimental imagery on summer, outdoor culture, and more.

Limited printed copies available.

— The Mahazine Team