🔎 Story Behind AIs

Core scientists from xAI, Llama 3, and Mistral discuss AI's future.

Hello, AI Enthusiasts!

In this issue, we explore the highlights from the GenAI Summit in San Francisco, an event that drew together some of the brightest minds in AI research. Their expertise captivated a host of Silicon Valley developers, each eager to harness the potential of AI.

Don’t remember subscribing or want to stop receiving our newsletters completely? Unsubscribe here

SPONSORED BY

Perplexity AI is the best AI search engine

TODAY’S AI SECRET:

  • Core scientists from xAI, Llama 3, and Mistral discuss AI's future.

  • OpenAI is restarting its robotics team.

  • Daily trending & featured AIs boost your career and business.

  • Google states that AI Overviews has been enhanced.

Get AI Secret's latest newsletter every morning before 6 a.m. ET. Can't find it? Check your spam or promotions, then drag it to your main inbox for future ease. (click this for Gmail)

Doing this a few times should ensure that our emails consistently reach your inbox.

TODAY’S AI WORLD

  • Maven Smart System - The U.S. Army has granted a $480 million contract to Palantir for the development of the AI-driven Maven Smart System, a component of the broader AI-powered Project Maven aimed at enhancing data integration and target detection.

  • AI's Impact on Election - Nearly 80% of Americans expect AI abuses to impact the outcome of this year's presidential election, according to a survey by Elon University Poll and the Imagining the Digital Future Center.

  • OpenAI Says - OpenAI reveals the misuse of its generative AI technology by Russian, Chinese, Iranian, and Israeli groups in global influence campaigns.

  • Surpass MacBook Air - Microsoft asserts its new Copilot Plus PCs outshine the MacBook Air M3 in performance, battery life, and price.

GENAI SUMMIT

Core scientists from xAI, Llama 3, and Mistral discuss AI's future.

The GenAI Summit in San Francisco brought together prominent AI researchers to discuss the latest advancements in large language models (LLMs). The panel, featuring xAI co-founder Tony Wu, Llama series co-author Sharan Narang, and Mistral AI research scientist Devendra Chaplot, attracted significant attention from Silicon Valley developers.

Tony Wu on xAI's Future Plans:

  • Data Center Expansion: Tony discussed xAI's plan to build a data center with 100,000 H100 chips using their recent $60 billion funding. This infrastructure aims to support the development of powerful AI models.

  • Use of Proprietary Data: He emphasized the advantages of using proprietary data, like data from X, to enhance AI models' performance and relevance.

Sharan Narang on Llama's Innovations:

  • Llama 3 Development: Sharan highlighted the upcoming release of Llama 3, including a 400B parameter version and accompanying research papers to benefit the community.

  • Data Quality and Diversity: He stressed the importance of data quality, diversity, and robust evaluation mechanisms in pre-training large models, ensuring they are well-rounded and unbiased.

Devendra Chaplot on Mistral AI's Approach:

  • New Model Launch: Devendra introduced Codestral, Mistral AI's new model trained in 80 programming languages, designed to optimize performance per parameter.

  • Synthetic Data and Efficiency: He discussed the potential of synthetic data to enhance overall data quality and efficiency in training models, addressing the data scarcity issue.

TOGETHER WITH PERPLEXITY AI

Perplexity AI is an AI-powered search engine that utilizes advanced technologies such as natural language processing (NLP) and machine learning to provide accurate and comprehensive answers to user queries.

The Key Features:

  • Accuracy: Perplexity AI uses a large language model (LLM) trained on extensive text and code data to give accurate answers.

  • Transparency: Users can see the sources of information, aiding in evaluating accuracy and reliability.

  • Citation: It provides source citations for users to verify information and learn more.

  • Summaries: Perplexity AI offers text summaries for a quick topic overview.

  • Personalization: It tailors answers based on users' interests and search history.

Now you can access Perplexity AI for free from their website.

OPENAI

OpenAI is restarting its robotics team.

Image Credit: AI Secret

OpenAI is rebooting its robotics team, aiming to develop AI models that can be integrated into robots built by other companies. This move marks a significant shift after the company previously abandoned its robotics efforts in 2020.

The Details:

  • Previous Efforts: OpenAI initially aimed to build a general-purpose robot and published research on training neural networks to solve tasks like solving a Rubik's Cube with a robotic hand.

  • Abandonment in 2020: The initial robotics team was disbanded due to a lack of training data, and members were reassigned to other projects within OpenAI.

  • Rebuilding the Robotics Team: OpenAI is hiring research engineers to rebuild its robotics team, disbanded in 2020. The new team has existed for two months. The company will collaborate with external partners, providing AI models to robotics companies.

  • Recent Investments: OpenAI has put money into several robotics companies like Figure AI, 1X Technologies, and Physical Intelligence. OpenAI's models have helped Figure AI build robots that can talk and think.

Why It Matters:

  • Strategic Shift: OpenAI's decision to reboot its robotics team reflects a strategic shift towards providing foundational AI models for other companies' robotics systems.

  • Collaborative Approach: By focusing on collaborations rather than direct competition, OpenAI aims to integrate its AI advancements more broadly across the robotics industry.

  • Enhanced Capabilities: The integration of OpenAI's advanced models could significantly enhance the capabilities of robots developed by its partner companies.

  • Driving Innovation: OpenAI's investments and technological support are likely to drive innovation and accelerate the development of practical and versatile robotic systems.

AI TOOLS

Trending AIs.

📓 Sudowrite helps users write novels or screenplays faster.

⚙️ BuildTheKeyword finds ideas that will bring you customers on auto-pilot.

📜 Airstrip AI accurates legal documents for your business.

📲 ImagineQr creates stunning and fully customizable AI QR codes.

🗂️ AITabGroup utilizes AI to assist you in automatically categorizing your tabs.

Featured AIs.

GOOGLE

Google states that AI Overviews has been enhanced.

Image Credit: AI Secret

Google has recently made "technical improvements" to its AI-generated answers in Search. This comes after user reports highlighted instances of misinformation within these AI-generated summaries, known as "AI Overviews." In a detailed blog post, Liz Reid, Google’s head of search, explained the measures taken to enhance the accuracy and reliability of these summaries.

The Details:

  • AI Overviews Launched in the U.S.: This month, Google introduced AI-generated summaries at the top of search results in the U.S., which are powered by a custom language model and integrated into Google's core web ranking system.

  • User Feedback and Satisfaction: According to Liz Reid, user feedback indicated that AI Overviews increased satisfaction and assisted with complex queries. Their accuracy was compared to that of Search’s "featured snippets."

  • Instances of Misinformation: Despite these benefits, there have been reports of "odd and erroneous overviews." Examples included advice to use glue to keep cheese on pizza and to eat at least one rock per day. Some search queries were designed to produce incorrect results, and in certain cases, the overviews linked to unique, misleading answers from specific sites.

  • Technical Improvements: Google has implemented updates to address broad sets of queries and enhance detection of nonsensical queries. The company has also restricted user-generated content that could mislead and limited the presence of satire and humor in responses. Additional triggers were added to reduce the frequency of less helpful AI Overviews for particular queries.

Why It Matters:

  • Enhancing Search Reliability: Improvements to AI Overviews are crucial for maintaining the reliability and credibility of Google Search, ensuring users receive accurate and useful information.

  • Addressing Misinformation: By fixing issues related to misinformation, Google aims to prevent the spread of false or misleading content, which can have significant real-world implications.

  • User Trust and Satisfaction: Continued enhancements based on user feedback are essential for building and maintaining trust among Google’s user base, ensuring that the platform meets their information needs effectively.

1 Million+ AI enthusiasts are eager to learn about your product.

AI Secret is the world’s #1 AI Newsletter, boasting over 1 million readers from leading companies such as OpenAI, Google, Meta, and Microsoft. We've assisted in promoting Over 500 AI-Related Products. Will yours be the next?

What We Can Offer:

  • Launch an Advertising Campaign

  • Conduct a Survey

  • Introduce New Product or Features

  • Other Business Cooperation

Email our co-founder Mark directly at [email protected] if the button fails.