📆 August 6th 2024 📆 

Developers get some love from Github with new models helping develop video & 3D content.

• Github launches Github Models to help developers test and run a variety of models within their applications

• Character-1 video generation models is released by Hendra - showing that the video gen business is not slowing down

• Nvidia hint at delays of their Blackwell AI chip amidst design flaws. With may snapping at their heels I wonder if any can use this to catchup?

• 3D models continue to get attention with StabilityAI launching Stable Fast 3D. Black Forest Labs emergence this week only strengthens the need for Stability to push on innovation.

• As always, too much to put in headlines. Check out the Other News Rollup section for a quickfire review of the top stories in AI.

Dive in for all the details!

🆘 Quick Tips 🆘 

Want to test out the latest model from newly formed Black Forest Labs but don’t want to pay for it? Hugging Face already has some hosted spaces were you can try out the new model for free.

Head on over to Hugging Face and load up Black Forest’s FLUX.1 dev space (https://huggingface.co/spaces/black-forest-labs/FLUX.1-dev). As a reminder from a previous newsletter, be descriptive in your image prompts.

A photograph of a bear drinking from a river using a hyper realistic style. The image should be captured through the trees and zoomed in on the bear, 16:9 aspect ratio

In these uncertain times finding an edge is critical to staying ahead. If you want to discuss how you can weather uncertainty with better data and efficiencies schedule a free 30 minute call

💫Help Support my Work💫

I really enjoy researching and writing about AI & appreciate your support. If you enjoy this content please use any of the options below to help

• Buy Me a Coffee ☕️- AI is really fuelled by coffee - keep the tank filled

• Spread the Word 🗞️- Share on social or directly with your friends

• Follow me on X - Always happy to chat AI and software dev

📰 News 📰 

GitHub Models Enhance Developer Access to Generative AI

GitHub has introduced "GitHub Models," a new initiative designed to provide developers with enhanced access to generative AI models, including notable names like Meta's Llama 3.1, OpenAI's GPT-4o, and Mistral Large 2. This initiative aims to simplify the process of experimenting with generative AI, enabling developers to integrate these powerful models directly into their workflows. This access is intended to foster innovation and accelerate the adoption of AI in various application areas​​.

The GitHub Models program is part of a broader trend of democratizing AI tools, making them more accessible to developers of all skill levels. This move aligns with GitHub's vision of streamlining the development process and empowering developers to create more sophisticated applications without the need for extensive AI expertise​.

Read More At: Github Blog

Hedra Secures $10 Million in Seed Funding for AI Video Platform

Hedra, an innovative AI video creation platform, has successfully raised $10 million in seed funding. This funding round was led by prominent investors, including Index Ventures, Abstract, and A16Z Speedrun, highlighting the growing interest in AI-driven video production technologies​​.

Founded by former employees of tech giants such as Nvidia, Google, and Meta, Hedra aims to revolutionize video content creation with its newly released foundation model, Character-1. Marketed as the fastest video foundation model available, Character-1 allows users to generate videos quickly and efficiently without sacrificing control. Since its launch, over 350,000 users have created more than 1.6 million videos using the platform, many of which have gained popularity on social media​ )​.

Hedra plans to enhance its platform further by integrating story, sound, and video generation into a single workflow, allowing for the seamless creation and customization of digital avatars and characters.

Read More At: VentureBeat

NVIDIA's Blackwell AI Chips Face Delays

NVIDIA's upcoming Blackwell AI chips are facing significant delays due to late-discovered design flaws, potentially pushing their release to early 2025. Major clients, including Microsoft and Google, have been informed of the setback. Initially expected to ship this year, the chips promise performance boosts of up to 30 times compared to the H100 model. The company, along with its manufacturer, Taiwan Semiconductor Manufacturing Company, is currently conducting test runs to address these issues. The delay could impact large orders worth tens of billions of dollars from various tech giants.

Read More At: Engadget

Stability AI's 3D Image Generation with Stable Fast 3D

Stability AI has launched a ground-breaking AI model called Stable Fast 3D, designed to rapidly generate high-quality 3D assets from a single image in just 0.5 seconds. This model represents a significant advancement over previous technologies, such as the Stable Video 3D, which required up to 10 minutes for similar tasks​.

Stable Fast 3D builds on the foundation of the TripoSR model, introducing major architectural improvements to enhance speed and quality. The model uses advanced techniques for material and illumination estimation, producing detailed and consistent 3D assets​​. It is particularly beneficial for industries such as gaming, virtual reality, retail, and architecture, where rapid 3D prototyping is essential​​.

The model is accessible through Stability AI's API and is available on platforms like Hugging Face.

Read More At: StabilityAI

🔨 AI Powered Tools 🔨 

• FLUX.1 - Black Forest Labs offers FLUX.1, an AI-powered image generation tool with three variants: FLUX.1 [pro], which provides high-performance image creation and visual quality for commercial use; FLUX.1 [dev], designed for non-commercial applications with open weights for experimentation; and FLUX.1 [schnell], optimized for fast, local development. Brought to you by the minds behind Stable Diffusion

• Llama Coder - Code generation tool designed to assist developers in rapidly creating applications from their ideas. It utilizes advanced models like Llama 3.1 and Together AI to offer features such as intelligent code suggestions, error checking, and seamless integration into existing development workflows.

• Semantic Scholar - Research tool designed to enhance the discovery of scientific literature. It indexes millions of research papers across various fields and employs AI to help users find relevant studies quickly. The platform features tools like Semantic Reader, which offers an enriched reading experience with contextual insights, and a robust API for developers to integrate scholarly data into their applications.

🗞️ Other News Rollup 🗞️ 

AI Cyber Challenge Support - Google supports DARPA’s AI Cyber Challenge with resources and expertise to enhance cybersecurity through innovative AI solutions.

OpenAI Model Safety - OpenAI's next AI model will undergo U.S. government safety checks, enhancing commitment to AI safety protocols.

Nvidia Antitrust Probe - The DOJ investigates Nvidia's acquisition of Run:ai amid concerns of market dominance and anti-competitive practices.

AI Solution for Karoshi - A new AI app aims to reduce Japan's karoshi crisis by providing personalized stress management solutions.

AI Crime Prediction - Argentina launches an AI unit to predict crime, raising concerns about effectiveness and civil liberties.

ChatGPT Watermarking Debate - OpenAI debates watermarking ChatGPT text due to user concerns, despite its effectiveness in detecting AI-generated content.

Meta's Celebrity AI - Meta is negotiating with celebrities for voice use in its new AI assistant, potentially paying millions

Securing Open Source AI - A new technique makes open source AI models harder to misuse, enhancing security against harmful modifications.

Figure 02 Unveiled - Figure 02, the latest humanoid robot from OpenAI-backed Figure, launches on August 6, 2024

Google and Character AI Deal - Google licenses Character AI's technology, enhancing its product development while poaching key executives for DeepMind.

🎶 Prompts 🎶 

Prompts used to generate some of this issues images. Unless otherwise stated all images are using Dall-E and ChatGPT Plus.

Thank You for Subscribing