Welcome to the AI Daily column! This is your daily guide to exploring the world of artificial intelligence. Each day, we bring you the hottest topics in the AI field, focusing on developers to help you understand technological trends and innovative AI product applications.

Explore Fresh AI Products Click to Learn More: https://top.aibase.com/

1. Google Releases New Gemma22B Model, Surpassing GPT-3.5-Turbo and Mixtral-8x7B

Google has introduced the high-performance Gemma22B model, featuring 2 billion parameters, which performs excellently across various hardware environments. Equipped with the ShieldGemma safety classifier and Gemma Scope tools, it aids researchers in content filtering and decision analysis. This significant breakthrough in intense competition showcases continuous efforts in AI technology development.

image.png

AiBase Highlights:

🌟 The Gemma22B model, with 2 billion parameters, performs excellently across various hardware environments, surpassing multiple models including GPT-3.5-Turbo.

🔍 The new model is equipped with the ShieldGemma safety classifier and Gemma Scope tools, aiding researchers in content filtering and decision analysis.

🚀 Google needs to accelerate the launch of Gemma3 to maintain market competitiveness.

2. ByteDance Introduces SpongeMusic, an AI Music Product on Par with Suno: Generate Music with a Single Click of Inspiration

SpongeMusic is an AI music creation tool comparable to Suno, allowing everyone to easily realize their musical dreams. By inputting inspiration or lyrics, SpongeMusic quickly generates musical works, bridging the gap between ordinary people and music creation, making creation more accessible. It offers rich customization features, supports various music styles, and allows users to enjoy a unique music creation experience.

image.png

AiBase Highlights:

🎵 Create music effortlessly using AI technology, simply by inputting inspiration or lyrics.

✨ Offers rich customization features, supports various music styles, meeting individual creative needs.

💡 Discover new inspiration during the creative process, enjoying a unique music creation experience.

Details Link: https://top.aibase.com/tool/haimianyinle

3. Alibaba Anticipates Launching an AI Conversational Procurement Engine in September

Alibaba's International Digital Commerce Group plans to launch an innovative AI conversational procurement engine aimed at revolutionizing the B2B e-commerce procurement experience, particularly serving small and medium-sized enterprises. The engine integrates multiple e-commerce platforms, using AI technology to parse natural language and provide precise procurement demand conversion. It possesses robust data analysis capabilities, predicts procurement trends, and offers personalized recommendations to assist businesses in making intelligent decisions.

image.png

AiBase Highlights:

🚀 The engine will revolutionize the B2B e-commerce procurement experience, focusing on serving small and medium-sized enterprises.

💬 Integrates multiple e-commerce platforms, using AI technology to parse natural language and provide precise procurement demand conversion.

📊 Possesses strong data analysis capabilities, predicts procurement trends, and offers personalized recommendations to assist businesses in making intelligent decisions.

4. Runway Announces the Launch of the Gen-3 Alpha Turbo Version

Runway recently announced on social platform X the exciting news of the upcoming release of the new AI video model Gen-3 Alpha Turbo, available at a significantly reduced price. This move will inject new vitality into the AI video creation field and has also sparked discussions about the balance between technological innovation and copyright protection.

image.png

AiBase Highlights:

🚀 The Gen-3 Alpha Turbo model generates videos 7 times faster than the previous Gen-3 Alpha, while maintaining excellent performance.

💰 Runway has reduced the price for the new model, possibly to expand broader applications and create more overall revenue.

🔍 Runway faces copyright disputes, plans to scrape content from YouTube channels to train AI models, sparking controversy but has not responded directly.

Details Link: https://top.aibase.com/tool/runwayml

5. 360 Floating Ball to be Replaced by AI Assistant: Integrating 15 of the Most Powerful Models

At the ISC.AI2024 12th Internet Security Conference, Zhou Hongyi announced that the 360 floating ball will be replaced by an AI assistant, with an additional AI assistant floating button. The AI assistant collaborates with 15 top large-model companies, fully integrating into 360's national-level entry product, allowing users to use it without installing plugins. Zhou emphasized that the AI assistant offers a one-stop experience, enabling users to access the most powerful AI models in China, select the most suitable model for effect comparison and cross-validation, to choose the most satisfactory solution.

image.png

AiBase Highlights:

🤖 The AI assistant collaborates with 15 top large-model companies, fully integrating into 360's national-level entry product, allowing users to use it without installing plugins.

🔍 Users can select the most suitable AI model based on their needs, for effect comparison and cross-validation, to choose the most satisfactory solution.

🖥️ The 360 floating ball will be replaced by the AI assistant, with an additional AI assistant floating button, providing a more convenient AI function experience.

6. Huawei Pura 70 Series Phones Add AI Image Expansion Feature

The Huawei Pura 70 series phones, after upgrading to the HarmonyOS 4.2.0.172 version, have added a remarkable AI image expansion feature, bringing users a richer and more interesting photography experience. Users can easily enjoy the creative fun brought by AI technology, save discarded photos, and unleash creative potential.

image.png

AiBase Highlights:

✨ The AI image expansion feature allows users to intelligently understand the content of the picture, fill in the blank areas, and provide a broader field of vision.

⚙️ Easy to use, users just need to open the photo app, select the edit function, click AI retouching, and choose image expansion.

📸 The AI image expansion feature will be supported by Pura 70 and Pura 70 Beidou Satellite Message Edition by the end of August 2024, based on advanced AI composition algorithms and image generation technology.

7. Elon Musk Responds to xAI Considering Acquisition of Character.AI: Acquisition Not Considered

Elon Musk's xAI company is not considering acquiring Character.AI, this move is to further test and develop its own Grok chatbot. Character.AI, founded by former Google employees, provides a platform for designing personalized chatbots and AI assistants, attracting a large number of young users.

image.png

AiBase Highlights:

✨ xAI company is not considering acquiring Character.AI, focusing on testing and developing the Grok chatbot.

💡 Character.AI, founded by former Google employees, provides a platform for designing personalized chatbots and AI assistants.

🌟 Character.AI's products have both free and paid versions, with the paid version costing $9.99 per month for a subscription.

8. Unveiling the "Black Box" of Language Models! Google DeepMind Releases a Set of Visualization Tools, Gemma Scope

In the world of artificial intelligence, language models are like an enigmatic black box, where we input words and it outputs meaning. Google DeepMind's latest research, Gemma Scope, peels back a corner of this black box. Through the training of sparse autoencoders, Gemma Scope reveals the true meaning behind the activations of language models, taking an important step towards the interpretability and safety of AI.

image.png

AiBase Highlights:

🔍 Gemma Scope reveals the true meaning behind the activations of language models through sparse autoencoders.

🧠 JumpReLU SAEs control activations, optimize reconstruction loss, and regularize the number of latent features.

📊 Gemma Scope performs excellently across different subsets of datasets, especially on DeepMind mathematics.

Details Link: https://storage.googleapis.com/gemma-scope/gemma-scope-report.pdf

9. Yum! Brands Tests AI Ordering Service in Over 100 Taco Bell Stores in the US

Taco Bell's parent company, Yum! Brands, is testing the introduction of voice AI ordering services in stores to improve order accuracy, shorten waiting times, and reduce employee burdens to earn more profits. The results of AI experiments in the fast food industry are mixed, with McDonald's withdrawing its plan and Wendy's being satisfied. The scale of the promotion is relatively small, with the test having lasted for about two years, and the technology has also been extended to KFC in Australia.

image.png

AiBase Highlights:

😜 Taco Bell's parent company plans to introduce voice AI technology into hundreds of stores by the end of the year.

😕 The results of AI experiments in the fast food industry are mixed, with McDonald's withdrawing its plan and Wendy's being satisfied.

🤔 There are many Taco Bell stores in the US, and the scale of this promotion is relatively small, with the test having lasted for about two years, and the technology has also been extended to KFC in Australia.

10. TikTok Pays Nearly $20 Million Per Month to Become One of Microsoft's Largest AI Cloud Computing Customers

TikTok pays nearly $20 million per month to Microsoft for access to OpenAI's artificial intelligence models, highlighting the importance of AI services in the tech industry. Microsoft earns substantial revenue from the large-scale use of OpenAI models, but may face challenges from TikTok's potential development of its own large language models.

image.png

AiBase Highlights:

🔍 TikTok pays nearly $20 million per month to Microsoft, accounting for a quarter of Microsoft's cloud department revenue.

💰 Microsoft's cloud AI business is expected to achieve $1 billion in annual revenue, benefiting from TikTok's use of OpenAI models.

⚖️ TikTok's parent company ByteDance is accused of secretly using OpenAI technology to develop AI models, sparking breach of contract disputes and investigations.

11. Google Adjusts Search to Help Hide Explicit Deepfake Content

Google has recently launched new online safety measures aimed at effectively removing deepfake content and preventing it from appearing prominently in search results. This initiative provides users with a more reassuring search experience.

image.png

AiBase Highlights:

😃 Google launches new online safety features, massively deleting deepfake explicit content.

😜 Adjusts search rankings, downgrading websites containing fake explicit images.

😕 Continues to take measures to address online dangers and explicit content.

12. US Copyright Office Calls for Legislation to Ban AI Deepfake Technology

The US Copyright Office has released its first report on the impact of AI in the field, calling for legislation to ban AI deepfake technology, emphasizing the threat of unauthorized digital reproductions to entertainment, politics, and personal rights. The report proposes several recommendations, including liability arising from the distribution of unauthorized reproductions, protection lasting a lifetime, and the establishment of a safe harbor mechanism.

image.png

AiBase Highlights:

⚖️ New laws are needed to define and combat AI-driven deepfake technology, curbing the distribution of unauthorized digital reproductions.

🔒 Liability should arise from the distribution or provision of unauthorized digital reproductions, with protection lasting at least a lifetime.

🛡️ Regulations should include a safe harbor mechanism, encouraging online service providers to remove unauthorized digital reproductions.