Your weekly briefing on all things AI

It would be tough to find a truer statement in the world of AI than “This is the worst AI will ever be.”

There has been a lot of hype in the news around bigger LLMs and their emerging capabilities, but it's important to see past the magic tricks and recognize why these developments will change (and already are changing) businesses outside of silicon valley. Accessibility is increasing just as rapidly as capabilities, and whether it’s through developing their own models, or developing their own workflows to leverage existing models, industry leaders will be the ones getting their hands dirty with this stuff first. 

TL;DR: The better these LLMs get, the better the businesses using them get. Early adopters will have the operations in place to capitalize on upgrades faster and penetrate the market sooner.

Let’s dive in:

📰 Why Meta’s New Model Matters For Businesses

📈 Clio's $900M Investment to Bridge the Justice Gap

💡 “Model Collapse”: How AI Can Distort Its Own Reality

 🛠️ Airtable Eliminates Their Learning Curve with AI

📰 AI in the News

Meta’s Llama 3.1 Gives Billions of Users Free Access to Next-Level AI

Meta Platforms has launched Llama 3.1, its most advanced AI model, now powering Meta AI across WhatsApp, Messenger, Instagram, Facebook, and Ray-Ban Meta, with Meta Quest integration planned. This move challenges OpenAI's market dominance, offering advanced AI capabilities to billions of users.

Llama 3.1's advanced capabilities offer transformative use cases across various industries. In customer service, its precise, context-aware responses improve satisfaction and streamline operations. Retail, hospitality, and healthcare sectors can efficiently manage high volumes of interactions with AI-driven chatbots and virtual assistants, reducing costs and enhancing service quality. Additionally, its ability to analyze extensive customer data allows for highly personalized marketing campaigns, boosting engagement and loyalty. Its multilingual features further break down language barriers, facilitating better communication in diverse markets. 

By democratizing access to advanced AI technology, this open-source model empowers even small and medium-sized enterprises to compete more effectively with larger corporations, driving innovation and operational efficiency across various sectors​​​​​​. (Link)

Nvidia's AI Foundry: Customizing Large Language Models for Businesses

Nvidia has launched AI Foundry, a new service to help businesses create and deploy custom large language models. Combining Nvidia's hardware, software, and expertise, AI Foundry enables companies to develop tailored versions of open-source models like Meta’s Llama 3.1. Verticalizing AI has never been this easy before and will never be as hard as it is today.

Nvidia also introduced NIM (Nvidia Inference Microservices), which packages custom models into API-accessible formats for easy deployment. This move positions Nvidia as a comprehensive AI solutions provider beyond its core GPU business, catering to enterprises seeking advanced AI capabilities while controlling their data and applications. (Link)

More Stories

• Amazon Upgrades “Just Walk Out” Checkout Experience (Link)

• ChatGPT is Rolling Out Voice Mode (Link)

• OpenAI Endorsed 3 Senate Bills This Week (Link)

• AirTable Aquires Dopt and Some Top AI Talent Along With It (Link)

• Runway Can Turn Images into Videos Ads  (Link)

• Former Tesla Robotics Chief Launches Warehouse Automation Startup. (Link)

• Zuckerberg's Vision: AI Clones for Content Creators. (Link)

• Meta AI's New 'Imagine Me' Selfie Feature Takes Off. (Link)

• AI Revolution Hits Video Game Industry Jobs. (Link)

• Google Supercharges Gemini Chatbot for Speed and Reach. (Link)

• OpenAI Introduces SearchGPT, the Future of Search Engines. (Link)

📈 Venture Deals 

Clio Secures $900 to Drive Legal Industry Innovation

Clio, a leading cloud-based legal practice management software provider, has successfully raised a substantial $900 million in Series F funding. Clio's platform streamlines legal workflows, enhances productivity, and ensures the security of sensitive client data.

This investment round, led by New Enterprise Associates (NEA) and joined by prominent investors such as Goldman Sachs Asset Management, Sixth Street Growth, CapitalG, and Tidemark, places a $3 billion valuation on the company. With this substantial capital infusion, Clio aims to accelerate its AI initiatives, expand its product offerings, and enter new markets. The company plans to utilize generative AI to further streamline legal workflows, increase productivity, and “address the justice gap” by enabling lawyers to serve more clients.

What’s interesting here is that Clio’s last valuation in 2021 was at $1.6B… Companies that intelligently integrate AI into their core offerings, like Clio has, are likely to get substantially more attention from investors. (Link)

More Venture Deals

• Beijing’s Baichuan Secures $687M, Reaching $2.8B Valuation to Build LLMs. (Link)

• Cohere Raises $500M to Outpace Generative AI Competitors. (Link)

• Vanta Acquires $150M to Redefine Tech Compliance with AI. (Link)

• IntelePeer Raises $140M to Revolutionize Customer Contact Centers. (Link)

• OpenAI-Backed Harvey Secures $100M for Cutting-Edge Legal Tech. (Link)

• Lakera Raises $20M to Shield Enterprises from LLM Security Risks. (Link)

• ZoomInfo Alum Raises $15M to Build AI Sales Engineers. (Link)

• London’s Fractile Raises $15M in Seed Funding for Next-Gen AI Chips. (Link)

💡 Industry Impacts

🏩 Healthcare

VeriSIM Life's BIOiSIM: AI at the Forefront of Drug Development Innovation

VeriSIM Life, based in San Francisco, is transforming drug discovery through its AI platform, BIOiSIM. By utilizing advanced machine learning, BIOiSIM analyzes molecular and biological data to design, test, and refine drug candidates virtually, cutting development time by up to 2.5 years and boosting prediction accuracy by 82%. A standout feature, AtlasGEN, simulates biological responses to predict drug efficacy and minimize animal testing. With successful client outcomes and significant funding, VeriSIM Life is positioning itself as a key player in accelerating drug discovery and addressing high costs and failure rates in the pharmaceutical industry. (Link)

🏛️ Law

Disabled Lawmaker Delivers First AI-Assisted Speech on House Floor

Virginia Congresswoman Jennifer Wexton made history on Thursday by addressing the House using an AI-generated clone of her voice. Diagnosed a year ago with progressive supranuclear palsy, a condition that impairs her ability to speak, Wexton turned to artificial intelligence to recreate her voice. This technology utilized past recordings of her congressional appearances and speeches. Wexton is believed to be the first individual to deliver a speech on the House floor with an AI-replicated voice. This groundbreaking use of AI in governmental proceedings opens new possibilities for accessibility in politics and public speaking, potentially paving the way for similar applications in various fields. (Link)

⚙️ Tech

Artificial Inception: How AI Distorts Its Own Reality

A new study led by Ilia Shumailov at Oxford, published in Nature, reveals a potential risk in AI model training known as "model collapse." This occurs when AI systems learn from data generated by other AI systems, creating a feedback loop that amplifies inaccuracies and biases. For example, image generation models might overproduce popular dog breeds, neglecting rarer ones.

To prevent this, researchers stress the importance of maintaining diverse, high-quality, human-generated data in AI training. They recommend developing methods to differentiate between AI and human-created content and establishing benchmarks for data variety.

This study highlights the critical need for careful data management in AI development to ensure long-term effectiveness and accuracy, emphasizing the ongoing importance of human input in the evolving field of machine learning. (Link)

🛠️ Tool of the Week

Airtable is leveraging AI to enhance their customer experience in big ways. Airtable is a popular cloud-based platform that blends the best features of spreadsheets and databases. What has set Airtable apart are its customizable fields, multiple data views, and the ability to link related information across different tables- allowing for a diverse range of use-cases ranging from project management to customer relationship management (CRM), and even inventory tracking. An obstacle to adoption though, was the learning curve for implementation.

With new AI updates, branded as CoBuilder, Airtable is set to become even more powerful and user-friendly. 

Here’s what’s new:

• Natural Language Querying: Ask questions in plain English and get answers.

• Personalized User Experience: A tailored experience that adapts to your needs.

• Enhanced Data Quality: AI-powered checks for accurate data.

• Automated Data Entry: Save time with AI-driven input.

• Improved Data Visualization: Enhanced visual tools for better understanding.

• AI-Powered Insights: Deeper, actionable insights from your data.

These enhancements make Airtable more efficient and accessible, expanding its usefulness for various applications. (Link)

More Tools:

• Topview: Turn links or media into viral videos with realistic avatars in just one click.

• Tempest AI: Build your dream game in just 10 minutes without coding. 

• ChartFast: Never waste time on spreadsheets again. Let your AI Data Analyst handle the data and create graphs in seconds!

• Study Map: Your AI-powered study companion. Learn anything, anytime with personalized courses.