Daily AI Newsletter: Breaking: Midjourney has released the highly anticipated V6 version

In today’s email:

🤯 Ray Kurzweil is sticking to his long-held predictions: 2029 for AGI and 2045 for the singularity
👀 Nvidia CEO: We bet the farm on AI and no one knew it
👩🏻‍🔬 These scientists aren’t using ChatGPT — here’s why
🧰 18 new AI-powered tools and resources. Make sure to check the online version for the full list of tools.

Breaking: Midjourney has released the highly anticipated V6 version

David Holz, CEO of Midjourney, has excitedly announced the alpha release of their V6 model, which is available now! This groundbreaking update can be activated by selecting V6 in the settings dropdown menu ( use /settings to access that) or by simply appending “ --v 6” to any prompt.

The V6 model introduces several enhancements:

Enhanced Prompt Adherence and Length: The model now follows prompts more accurately and supports longer inputs.
Improved Coherence and Knowledge: There’s a notable upgrade in how the model understands and processes information.
Advanced Image Prompting and Remixing: The V6 model offers better capabilities in image prompting, including minor text drawing with specific instructions.
Upgraded Upscalers: The resolution of images can be increased twofold, featuring both ‘subtle’ and ‘creative’ modes.
Minor Text Drawing Ability: The model’s ability to incorporate text into images is a significant addition. Users can now include specific text by writing it in “quotations” and using the ‘ — style raw’ command or lower ‘ — stylize’ values for better integration. This feature allows for combining textual and visual elements in a single artwork.

Hire a world-class AI team for 80% less

Trusted by leading startups and Fortune 500 companies.

Building an AI product is hard. Engineers who understand AI are expensive and hard to find. And there's no way of telling who's legit and who's not.

That's why companies around the world trust AE Studio. We help you craft and implement the optimal AI solution for your business with our team of world class AI experts from Harvard, Princeton, and Stanford.

Our development, design, and data science teams work closely with founders and executives to create custom software and AI solutions that get the job done for a fraction of the cost.

Book a free consultation session now

p.s. Wondering how OpenAI DevDay impacts your business? Let’s talk!

Apple wants AI to run directly on its hardware instead of in the cloud

Apple's latest research, as detailed in their paper "LLM in a Flash," signals their intent to compete in the generative artificial intelligence space. The research focuses on running large language models (LLMs) like ChatGPT on smartphones, a task traditionally reserved for powerful data centers due to computational demands. This approach aims to overcome current limitations in device memory and computational power, enabling more efficient LLM inference on personal devices.

The paper gained attention after being featured on Hugging Face and marks Apple's second publication on generative AI this month. It aligns with Apple's broader strategy to run AI directly on iPhones, contrasting with rivals like Microsoft and Google who rely on cloud-based services. This shift is part of a wider industry trend, with companies like Samsung planning to launch AI-focused smartphones and Qualcomm predicting a new era of mobile experiences enhanced by AI.

Apple's research could lead to more sophisticated virtual assistants and advanced photo editing on smartphones. The move also emphasizes privacy, as processing data on the device reduces cloud dependency. While this research is indicative of Apple's direction, it is not a definitive guide to their future product features.

Google announced VideoPoet: A large language model for zero-shot video generation

"VideoPoet" is a new large language model (LLM) developed by Google Research, designed for advanced video generation tasks. It can perform various functions like text-to-video, image-to-video, video stylization, inpainting, outpainting, and video-to-audio conversion. Unlike traditional models, VideoPoet integrates multiple video generation capabilities within a single LLM framework, leveraging the learning strengths of LLMs across language, code, and audio.

Key aspects of VideoPoet include:

Multitasking on different video-centric inputs and outputs.
Use of specialized tokenizers for converting video, image, and audio into discrete tokens.
Training across multiple modalities for versatile video generation.

VideoPoet stands out for its ability to generate coherent and interesting motions in videos, interactive video editing, and control over camera movements in generated content. It has demonstrated superior performance in terms of text fidelity and motion interestingness compared to other models. VideoPoet signifies a step forward in the application of LLMs to video generation, with potential for future expansions in "any-to-any" generation tasks.