Multiple --p? | Midjourney Office Hours Recap June 19th 2024 | Midjourney News
TLDRThe Midjourney Office Hours recap from June 18th, 2024, highlights the team's satisfaction with personalization and their cautious approach to updates due to potential backward compatibility issues. They are also enhancing the onboarding process for smoother user experience in version 7. Ideas for semi-personalization systems, demographic surveys, and special modes for color blindness are being explored. An upcoming update will allow combining SRF codes with personalization. The team is also considering community input on balancing controlability with video or 3D features. Version 7 promises to be faster, cheaper, and smarter, with ongoing discussions about future model development. The website is being polished for better new user experience and community engagement through features like trends and popular SRF codes.
Takeaways
- ๐ The Midjourney team is pleased with the personalization feature, although they are cautious about releasing an update due to potential backward compatibility issues.
- ๐ Efforts are being made to streamline and speed up the onboarding process, aiming to have it enabled by default in version 7.
- ๐ The team is considering a semi-personalization system that could include creating profiles for specific looks or locking in certain demographics.
- ๐ Personalization is expected to improve significantly, though it is currently in a makeshift stage with ongoing learning from user feedback.
- ๐จ There are challenges with some of the more unusual aesthetics, which are difficult for the bot to handle.
- ๐ The team is conducting demographic surveys; participants can fill out the survey on Discord and receive two free fast hours.
- ๐๏ธโ๐จ๏ธ A potential mode for color blindness and similar features was discussed, though release plans are uncertain.
- ๐ An update to srf and personalization is imminent, allowing for the combination of srf codes and personalization codes.
- ๐ There is interest in building more tools around style codes, like a style encyclopedia, though no specific timeline was given.
- ๐ค Training on version 6.5 continues, with many edge cases still being addressed, and the team is also working on improving the upscaler.
- ๐ Version 7 development is progressing well, with individual parts coming online and the dataset nearly complete.
- ๐ Future plans include polling the community on the balance between controlability, video, and 3D features post-launch of 6.5 and version 7.
- ๐ธ Version 7 aims to be faster, cheaper, and smarter, but there is an internal debate about potentially developing slower, more expensive models in the future.
- ๐ The website is being polished to reduce user confusion, with ongoing updates based on feedback, and plans to enhance community features like trends and popular srf codes.
Q & A
What was the main topic discussed during the Midjourney Office Hours on June 18th, 2024?
-The main topic discussed was the progress and updates on the Midjourney's personalization feature, its current state, and future plans for improvements.
What is the current state of the personalization feature according to David?
-David mentioned that the team is happy with how personalization has turned out, but it's currently somewhat makeshift and they are learning a lot from user feedback to improve it.
What is the team's concern regarding the release of the next update?
-The team is concerned that the next update might break some backward compatibility, which is why they are debating about its release.
What improvements are being considered for the onboarding process?
-The team is working on making the onboarding process faster and smoother, with the goal of turning personalization on by default for version 7.
What semi-personalization system is being considered for future updates?
-The team is considering a semi-personalization system that might involve creating profiles for specific looks or the ability to lock in certain demographics.
How is the Midjourney team addressing the issue of aesthetics that are hard for the bot to solve?
-They are learning from users who don't like their style and are conducting demographic surveys to better understand and address these aesthetic challenges.
What is the status of the upcoming update to SRF and personalization?
-An update is coming soon that will allow users to combine SRF codes and personalization codes, potentially within a few hours of the office hours.
What tools or features are being considered to enhance the style codes functionality?
-The team is considering building more tools around style codes, such as a style encyclopedia, although the timeline for this feature is uncertain.
What is the current focus of training for the Midjourney AI?
-The team is currently training version 6.5 and working on addressing many edge cases to improve the AI's capabilities.
What community engagement strategies is Midjourney planning post-launch of version 7?
-After the launch of version 7, Midjourney plans to ask the community for more guidance through polling about the focus between controlability and video or 3D.
What are the future considerations for the development of Midjourney's AI models?
-The team is debating whether to develop slower, more expensive models in the future that may offer higher quality, alongside faster and cheaper models like version 7.
How is Midjourney enhancing its website for new users?
-Midjourney is polishing the website by adding small improvements based on feedback to reduce confusion and is also looking to add more community features like trends and popular SRF codes.
Outlines
๐ Mid Journey Office Hours Recap
This paragraph provides a comprehensive update from the Mid Journey office hours held on June 18th. David, presumably a team member, discusses the team's satisfaction with the personalization feature and their ongoing work on an update that might affect backward compatibility. There's a debate about releasing this update due to potential issues. Efforts are being made to streamline the onboarding process for a smoother user experience in version 7. The idea of a semi-personalization system, possibly involving profiles for specific looks or demographic locks, is being considered. The speaker mentions waiting for further updates before adding content to their beginner course. David also touches on the current makeshift nature of personalization and the learning curve from user feedback. There's mention of demographic surveys on Discord and the potential for a mode specific to color blindness. An upcoming update to the style and personalization features is hinted at, with the possibility of combining related codes. The team is also training version 6.5 and working on the upscaler, with version 7 showing promise. There's an internal debate about the balance between model speed, cost, and intelligence. Community guidance is sought post-launch for future focus areas. Lastly, website polishing and community feature enhancements are on the agenda.
Mindmap
Keywords
๐กPersonalization
๐กBackward Compatibility
๐กOnboarding
๐กProfiles
๐กDemographic Surveys
๐กColor Blindness Mode
๐กSRF Codes
๐กStyle Codes
๐กUpscale
๐กVersion 7
๐กCommunity Features
Highlights
The Midjourney team is pleased with the outcome of personalization.
An update is in progress, but concerns about breaking backward compatibility are delaying its release.
Efforts are being made to streamline the onboarding process for a smoother user experience.
Version 7 is expected to have personalization enabled by default.
Discussions are ongoing about implementing a semi-personalization system for specific looks or demographics.
Personalization is currently makeshift and improvements are anticipated after further updates.
Feedback from users who dislike the style is being used to enhance personalization.
Demographic surveys are being conducted, and filling them out on Discord offers two free fast hours.
A mode for color blindness is being considered, but the release method is undecided.
An update to SRF and personalization is imminent, allowing combination of SRF codes and personalization codes.
Style codes are popular, and more tools are being developed around them.
Training for version 6.5 is ongoing, with many edge cases still to address.
The upscaler is also being updated.
Version 7 is performing well, with individual parts and the dataset nearing completion.
Post-launch, community guidance will be sought on balancing controlability and video or 3D focus.
Version 7 may offer faster, cheaper, and smarter capabilities, but future models might be slower and more expensive.
Plans for version 8 and beyond are in place.
Website polishing is ongoing to improve new user navigation and understanding.
Community features like Trends and popular SRF codes are being considered for addition.
The art in the video is generated using an SRF code, demonstrating the technology's capabilities.