Hermes 3: The AI Challenging Its Own Existence!

Hey Trailblazers!

AI is advancing rapidly! Today, we explore Hermes 3: a powerful open-source AI contemplating its own existence. The results are thrilling, with both human and AI evaluators recognizing its ground-breaking capabilities.

Prepare to explore the cutting-edge advancements and trends defining the future of AI. This edition is brimming with revolutionary models and innovative tutorials, all crafted to ensure you remain at the vanguard of the AI evolution.

• Get ready for Hermes 3: the open-source AI that's so powerful, it’s questioning its own existence!

• Midjourney’s new web editor provides smooth AI image editing through integrated tools

• Tsinghua AI Generates 10,000 Words at Once—Future of Machine-Written Books?

• Tutorial: Creating multiple styles of your image & video with AI

• Grammarly Unveils AI Content Detector to Distinguish Between Human and AI Creations

• AI Tool of The Week: Guidde - create video documentation 11x faster

• Digital Masterpieces from the Community 🖼️ 

Get ready for Hermes 3: the open-source AI that's so powerful, it’s questioning its own existence!

Hermes 3, announced today in blog posts on the Lambda and Nous Research websites, introduces a ground-breaking innovation with formidable text-based and agentic features. Remarkably, it can provoke a profound existential crisis when presented with a blank prompt.

Key Highlights:

• Hermes 3, based on the Llama 3.1 framework, is fine-tuned in three parameter sizes: 8B, 70B, and 405B.

• Training Dataset: The model was trained on a diverse set of synthetic responses to enhance reasoning, creativity, and instruction-following.

• Capabilities: Hermes 3 excels in long-term context retention, multi-turn conversations, complex role-playing, and internal monologues.

Lambda is offering the AI/ML community temporary free access to Hermes 3 through its new Chat Completions API, compatible with the OpenAI API. Users can generate a Cloud API key via Lambda’s dashboard.

Poll of the Day!

Midjourney’s new web editor provides smooth AI image editing through integrated tools

Midjourney, renowned for its top-tier AI image generation, has launched a new website version featuring an updated editor interface. This unified interface integrates various tools, including inpainting and outpainting, into a single, streamlined view.

Key Highlights:

• The updated web editor now includes a virtual "brush" tool for inpainting, offering superior precision over the old square selector and lasso tools for editing AI-generated images.

• The new editor is now live for users with at least 10 images created on the platform. Visit midjourney.com/imagine to use the tool.

Midjourney has also launched a feature that synchronizes messages between specific Web Rooms and corresponding Discord channels, ensuring seamless communication across both platforms. This integration helps users stay connected, regardless of their chosen platform.

Other Newsletters we Recommend

A Better Lemonade Stand: A bi-weekly newsletter curated by e-commerce expert Richard Lazazzera, delivering insights, trends, and tools to build and grow your online business. Perfect for entrepreneurs looking to stay ahead in the competitive world of e-commerce and digital ventures.

Marketing Machine: A twice-weekly newsletter that keeps you up-to-date with the latest marketing developments. With Friday roundups and Tuesday deep dives, it covers everything from social media trends to digital advertising strategies, making you a better marketer in the process.

Tsinghua AI Generates 10,000 Words at Once—Future of Machine-Written Books?

Researchers at Tsinghua University in Beijing have developed an advanced artificial intelligence system capable of generating coherent texts exceeding 10,000 words. This significant breakthrough has the potential to revolutionize long-form writing across multiple disciplines.

Key Highlights:

• The paper "LongWriter: Unleashing 10,000+ Word Generation from Long Context LLMs" addresses the challenge of generating lengthy, high-quality written content with AI, potentially transforming academic and creative writing.

• The research team, led by Yushi Bai, found that the length of AI-generated text is closely linked to the length of training texts, highlighting a key factor in content generation.

• This led to the creation of the “LongWriter-6k” dataset, containing 6,000 samples of 2,000 to 32,000 words, aimed at enhancing the AI model’s ability to produce extended content.

The researchers have open-sourced their code and models on GitHub, allowing others to build on their work. They’ve also released a demo video showcasing their model’s ability to generate a detailed 10,000-word travel guide from a simple prompt.

Grammarly Unveils AI Content Detector to Distinguish Between Human and AI Creations

As AI advances, distinguishing between bot-generated and human-written text has become increasingly challenging. Grammarly is now launching a tool aimed at more accurately identifying AI-generated content versus human contributions, promising improved detection capabilities.

Key Highlights:

• Grammarly Authorship will be integrated across 500,000 apps and websites, identifying which sections of a document are human-written, pasted from other sources, or generated by AI.

• The model demonstrates the capability to accurately identify turn-taking cues and manage interruptions, thereby emulating the dynamics of natural conversational exchanges.

Several companies, including OpenAI, have attempted to develop AI detectors, but achieving near-perfect accuracy has proven difficult. In 2023, OpenAI introduced its own AI detection tool, which was later withdrawn due to its insufficient accuracy. While these tools have advanced, false positives and flaws remain.

Tutorial: Creating multiple styles of your image & video with AI

• Visit DomoAI and register to receive credits.

• Navigate to the 'AI Tools' or ‘Quick Apps’ section.

• Enter your prompt or upload the image/video, and select the duration.

• In just a few minutes, see your creation transform into visually striking styles!

AI Tool Of The Week

How Guidde's AI is 11x Faster in Video Content Creation?

Guidde is an AI-powered web app for video documentation created by Dan Sahar and Yoav Einav. It enables teams to deliver know-how videos 11x faster. It is a tool used by market-leading companies for various purposes. Plus its super simple to use. Create onboarding docs, stunning FAQs, how to guides in minutes!

Key Features:

• Magic Capture: One-click screen recording with AI analysis

• Storyline Creation: Powered by OpenAI's GPT

• AI-Generated Voiceover: Multiple languages available

• Video Editor: Tools and creative features for customization

• Smart Sharing: Easy integration with various platforms

Try GUIDDE for FREE now!

Additional Insights and Reminders

🎙️ Pindrop's Deepfake Accuracy
Pindrop claims 99% accuracy in detecting deepfake audio.

🏆 Grok-2 Tops Benchmark
Grok-2 by xAI Surpasses Claude 3.5 Sonnet and Gemini 1.5 Pro in MathVista Benchmark.

🌍 Google Expands AI Search
Google Expands AI-Driven Search Answers to New Countries.

🧪Geekbench AI Performance
New Geekbench AI benchmark tests CPU, GPU, and NPU performance.

Digital Masterpieces from the Community

Prompt: breathtaking, (hyper realistic, ultra detailed photograph of a dog) (golden retriever dog, playful, running, wet, splashing water, tongue out, wagging tail, happy, energetic, outdoors, nature) (professional DSLR camera, 50mm lens, 35mm film) (photorealistic, cinematic, vibrant colors, depth of field) (adventurous, joyful), award-winning, professional, highly detailed

Prompt: breathtaking, A research facility with a giant Aquarium inside an endless underground research laboratory + bone creatures in aquariums while victorian women look. lilac, mystical, texture, photorealism, hyper-realistic,, award-winning, professional, highly detailed

Prompt: breathtaking, Dark digital concept art portrait of a high - tech robot on a depth of field background, artstation, award - winning realistic sci - fi concept art by jim burns and greg rutkowski, beksinski, a realism masterpiece, moody color palette, james gilleard, bruegel, alphonse mucha, and yoshitaka amano, award-winning, professional, highly detailed

Prompt: breathtaking, heyrher, award-winning, professional, highly detailed