🚀 Midjourney Office Hours Highlights: January 29th

The V7 release is eagerly awaited, although the training timeline has slipped by about a week due to unforeseen issues. The team is exploring coding optimizations to speed up the final training phase.

For a quick overview, head straight to the TL;DR; section otherwise keep on reading.

Extended Recap

🖼️ V7 Release Updates

Two parallel V7 model trainings are underway: one with the new V7 dataset and another without the V6 dataset. The goal is to release the best-performing model, with a follow-up release likely to include a larger dataset. Users can expect improved image quality and enhanced "smartness" in the model, with better object and character references, including logos, mugs, faces, dogs, and monsters. Personalization features from previous versions will also be carried over, ensuring a seamless transition for users.

🖌️ 3D Model Developments

Midjourney is venturing into the realm of 3D modeling, aiming to handle a wide variety of scene types beyond single objects. A significant goal is the capability for moving camera angles and reframing images, which could salvage otherwise suboptimal images by adjusting composition. While timelines are uncertain due to additional training requirements and resource constraints, the team is confident that a robust 3D feature will be released this year.

🎥 Video Feature Progress

The skepticism surrounding video quality is diminishing as new models show promise. High-quality output is seen in about one out of four runs, indicating strong potential. However, rendering videos is expensive, especially for users on lower-tier plans. The team is investigating faster and cheaper video models to make this feature more accessible. Partnerships may expedite the video launch, though dates are partly dependent on external collaborators. If partner-led solutions fall short, in-house training will resume. Overall, the conviction has grown that video will become a key Midjourney offering this year.

📦 Big Batch Release ("Batch 8")

An upcoming "batch 8" feature is in development, potentially rolling out before or after V7. UI and design details are still in flux, and priority may shift toward finishing V7 and exploring video, so the exact timing for batch 8 is not firm.

⏱️ Speed and Quality Trade-offs

The community often favors higher-quality (slower) modes over faster ones, presenting an ongoing dilemma of balancing speed with fidelity. A potential solution is a quick generation mode for iteration and a slower, final mode for high-quality outputs. The team continues experimenting with ways to provide both without compromising user experience.

🖥️ Editor and Web Updates

Enhanced editor capabilities are in progress, including better selection tools and layering functions. The "Describe" feature is now integrated into the website, allowing users to drag an image into the create view to auto-generate prompt descriptions. Additional web and UI improvements are ongoing, though release dates remain variable.

🔬 Additional R&D and Secret Projects

Midjourney is expanding its physical lab space to support hardware-related research and prototyping. The company invests heavily in AI research, maintaining proprietary advancements beyond open-source norms. Several new initiatives may equal or exceed the impact of existing features, with roughly 90% tying back into the core community and product ecosystem. About 10% are more experimental, potentially offering unexpected breakthroughs. Each new project aims to be highly impactful, leveraging both software and hardware expertise.

TL;DR

🚀 V7 Release Updates: Training delays, dual model training, improved image quality, and personalization.

🖌️ 3D Model Developments: Wide scene handling, moving camera angles, and reframing capabilities.

🎥 Video Feature Progress: Promising quality, cost concerns, partnerships, and potential key offering.

📦 Big Batch Release ("Batch 8"): Development in progress, timing uncertain.

⏱️ Speed and Quality Trade-offs: Balancing speed and fidelity, experimenting with modes.

🖥️ Editor and Web Updates: Enhanced tools, "Describe" feature, ongoing improvements.

🔬 Additional R&D and Secret Projects: Lab expansion, AI research, impactful new initiatives.

Stay tuned for more updates as Midjourney continues to innovate and enhance its platform!

If you want to support me, feel free to buy me a coffee ☕️ 

If you’re not subscribed yet to my newsletter Imagine Weekly, I’d be thrilled to welcome you on board!