Today’s top AI Highlights:

1. Rent a Human Brain for AI Research - Swiss tech firm unveils brain rental service.

2. OpenAI Fine-Tuning for GPT-4o - Free fine-tuning now available.

3. Anthropic's Intelligence Scandal? - Are they faking it?

4. NVIDIA's Storm Tracking Tech - AI meets meteorology.

5. Global Text-to-Speech by ElevenLabs - New app launch.

6. Microsoft's Open-Source AI - Introducing Phi-3.5 series.

7. Salesforce's AI Agents - Meet Einstein SDR and Einstein Sales Coach.

8. Trending AI tools

   & so much more!

_{Read time: 8 minutes}

HIGHLIGHT OF THE WEEK

Rent a human brain for AI research - swiss tech firm to offer rental service

FinalSpark, a tech firm in Switzerland,  introduced AI made of human brain cells for rent. The company calls the cells Organoid meaning a computer made out of human brain cells or organic material. These computers are offered to researchers around the globe for rent to run their experiments and get the data report from the portal. The company says “Neurospark is the only company to offer access to biocomputers”.

What is special?

•  The biocomputer's purpose is to be a highly efficient, low energy solution to the high cost associated with developing front end AI models according to Finalspark. 

• This technology is 1000000 times better than traditional silicon based technology to train AI.

• The organoid's lifespan is calculated as 100 days.

WEEKLY HIGHLIGHTS 🌍

OpenAI launches free fine-tuning for GPT-4o

OpenAI has introduced a new feature allowing users to fine-tune its GPT-4o model for free until September 23, 2024. This initiative provides developers with the opportunity to customize the AI's responses according to specific needs, enhancing the model’s accuracy and relevance.

Key Points:

• Free Access: Users can fine-tune GPT-4o using up to 1 million free training tokens daily until September 23.

• State-of-the-Art Performance: Cosine’s Genie achieved state-of-the-art (SOTA) scores of 43.8% on the SWE-bench Verified benchmark and 30.08% on SWE-bench Full with a fine-tuned GPT-4o model, marking the largest improvement ever on this benchmark. Distyl also ranked 1st on the BIRD-SQL benchmark with a 71.83% execution accuracy.

• Customization: Fine-tuning allows developers to adjust GPT-4o’s outputs to better fit specific projects, making the AI more intuitive and aligned with user needs.

• Cost after Free Period: After September 23, fine-tuning will cost $25 per million tokens, with additional costs for input and output tokens.

• Privacy and Control: OpenAI ensures that all data used in fine-tuning remains private, with strict ethical guidelines to prevent misuse.

• Real-World Applications: Early adopters have successfully used fine-tuning for diverse applications, including customer support and code debugging.

• This new feature is an important step in making AI more accessible and adaptable to various industries and use cases​

Did Anthropic just fake its intelligence?

Anthropic, the developer behind the Claude AI chatbot, faces a class-action lawsuit filed by a group of authors who allege that the AI model was trained on pirated copies of their copyrighted books on Monday. This lawsuit adds to the growing legal challenges against AI companies regarding the use of copyrighted materials in training datasets.

Key Points:

• Lawsuit Filed: The lawsuit was filed in a San Francisco federal court by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, representing a class of similarly affected writers.

• Allegation Details: The lawsuit claims that Anthropic used a dataset known as "The Pile," which included pirated copies of hundreds of thousands of books, to train Claude AI.

• Anthropic's Defense: Anthropic has not yet commented publicly on this specific case but has previously argued that training AI models on publicly available data falls under the "fair use" doctrine.

• Wider Legal Context: This lawsuit is part of a broader trend of legal challenges against AI companies, including similar cases against OpenAI and Microsoft, where plaintiffs argue that their copyrighted works were used without permission or compensation.

• Impact on AI Development: The outcome of these lawsuits could have significant implications for how AI models are trained and the legal responsibilities of AI developers regarding copyrighted content.

NVIDIA will track storms now!

NVIDIA has introduced groundbreaking AI technologies that enhance both weather forecasting accuracy and the interactivity of non-playable characters (NPCs) in video games. These advancements are poised to significantly impact both climate research and the gaming industry.

 Key Points:

• AI Weather Forecasting:

  • StormCast Model: NVIDIA's new StormCast AI model predicts mesoscale weather events with unprecedented accuracy, offering a 10% improvement over existing methods like those from the U.S. National Oceanic and Atmospheric Administration (NOAA).

  • High Efficiency: The model is capable of super-resolving atmospheric data 1,000 times faster while using significantly less energy, thanks to NVIDIA’s accelerated computing technology.

  • Collaborations: NVIDIA is working with The Weather Company and Colorado State University to refine and potentially deploy this model in real-world scenarios.

• AI in Gaming:

  • NVIDIA ACE for NPCs: NVIDIA's Avatar Cloud Engine (ACE) will soon debut in the game Mecha BREAK, allowing NPCs to engage in dynamic, lifelike conversations powered by AI rather than pre-scripted dialogue.

  • Advanced Integration: The AI integrates with various technologies, including Audio2Face for facial animations and OpenAI’s Whisper for speech recognition, making NPC interactions more immersive.

  • Industry Impact: This technology is expected to revolutionize how players interact with game worlds, potentially setting a new standard for NPCs in the gaming industry.

ElevenLabs launches global text-to-speech app, reader

ElevenLabs has globally launched its Reader app, a text-to-speech application that supports 32 languages. The app, initially available in the U.S., U.K., and Canada, now allows users worldwide to convert text into high-quality audio using a wide range of AI-generated voices.

Key Points:

• Global Availability: The Reader app is now accessible worldwide on both iOS and Android, supporting 32 languages, including major languages like Spanish, Hindi, and Japanese.

• Voice Library Expansion: The app features hundreds of new voices optimized for various languages and accents, including licensed voices of iconic actors.

• Versatile Content Support: Users can convert different text formats like PDFs, e-books, and articles into audio, making content more accessible on the go.

• Future Enhancements: Planned updates include offline support and the ability to share audio snippets, further improving user experience​

Microsoft launches new open-source AI models: Phi-3.5 series

Microsoft has released a new series of open-source AI models, the Phi-3.5 series, which includes models optimized for text, vision, and multimodal tasks. These models are designed to deliver high performance while maintaining efficiency, making them suitable for a wide range of applications from cloud-based services to edge devices.

Key Points:

• Phi-3.5 Series Overview: The series includes three models: Phi-3.5 Mini (3.8 billion parameters), Phi-3.5 MoE (42 billion parameters with a unique Mixture of Experts architecture), and Phi-3.5 Vision (4.2 billion parameters for multimodal tasks).

• Performance and Efficiency: These models outperform larger competitors in various benchmarks, offering high performance in multilingual tasks, reasoning, and visual processing while requiring fewer computational resources.

• Open-Source Availability: All Phi-3.5 models are available on Hugging Face under an open-source MIT license, allowing developers to freely use, modify, and commercialize them.

• Integration with Azure ML: Microsoft has integrated these models into its Azure Machine Learning platform, offering users tools for fine-tuning, deploying, and optimizing AI models with features like DeepSpeed and ONNX Runtime 

Salesforce unveils Einstein SDR and Einstein Sales Coach: new AI agents to boost sales teams

Salesforce has introduced two new autonomous AI sales agents, Einstein SDR (Sales Development Rep) and Einstein Sales Coach. These AI tools are designed to automate key sales tasks, from lead engagement to personalized sales coaching, helping teams scale their operations and improve effectiveness.

Key Points:

• Einstein SDR Agent: This AI autonomously manages inbound leads, handling tasks such as answering questions, overcoming objections, and booking meetings. It operates 24/7, enabling sales teams to focus on more complex tasks.

• Einstein Sales Coach Agent: Provides personalized coaching by simulating real sales scenarios, allowing sales reps to practice pitches and negotiations. It offers objective feedback based on real-time data, helping improve performance.

• Integration and Customization: Both agents are built on Salesforce's Einstein 1 Agentforce Platform and can be customized using no-code workflows and templates. They leverage Salesforce Data Cloud for enhanced data integration and accuracy.

• Security and Trust: The agents operate with Salesforce’s Einstein Trust Layer, ensuring secure and reliable responses grounded in a company’s CRM data.

Trending Tools 📈

1. Sybil : To analyze and transcribe customer conversations, providing insights into buyer intent and engagement through natural language processing and emotion detection.

2. FlowKitten : To streamline workflow automation by intelligently managing tasks, data processing, and integrations across various applications, enhancing productivity with minimal manual intervention.

3. Simplex : To simplify complex data analysis and decision-making processes through advanced machine learning algorithms, making AI accessible to users without deep technical expertise.

4. Ideogram 2.0 : To generate text-based images, allowing users to create visual representations of words and phrases with advanced customization options.

5. D-iD : Transforms still images into lifelike videos, enabling users to create engaging, personalized avatars and interactive video content from simple photographs.

Stay ahead of the curve by exploring these trending tools! Join our Telegram channel for more updates and insights. Explore more on @ainews_daily!