🖼️ Mistral Upgrades, Set to Compete with ChatGPT

Mistral's latest innovations, PixTral Large and the enhanced Le Chat, are poised to challenge major players in the AI landscape.

Author

Jack Lajoie
November 19, 2024 • read time ~ 9 minutes

The Daily Current ⚡️

Welcome to the creatives, builders, pioneers, and thought leaders ever driving further into the liminal space.

Mistral AI's release of PixTral Large and the enhanced Le Chat have it set to challenge the dominance of established models like ChatGPT. Meanwhile, Perplexity's new shopping feature disrupts the e-commerce landscape, integrating search and purchasing for a seamless user experience. The research on AI-powered fraud highlights the cybersecurity sector's ongoing battle against increasingly sophisticated threats. The USPTO's internal ban on generative AI tools underscores the complexities of adopting new technologies in public institutions, while the DOJ's antitrust push against Google illustrates a strategic shift in regulatory efforts to address market imbalances.

 🔌 Plug Into These Headlines:

Mistral AI has unveiled two significant AI products: Pixtral Large, a multimodal AI model capable of processing text and images, and an upgraded version of Le Chat, their chatbot. Pixtral Large competes with GPT-4V, excelling in areas like math and coding. The enhanced Le Chat now offers features such as web browsing and document analysis, positioning it as a direct competitor to ChatGPT. These launches demonstrate Mistral’s rapid progress in AI development since its founding in May 2023.

  • Mistral AI launched Pixtral Large, a 124 billion parameter multimodal model capable of processing text and images, on November 18, 2024.

  • Pixtral Large outperforms competitors like GPT-4o and Gemini-1.5 Pro in tasks such as document question answering (DocVQA) and chart analysis (ChartQA).

  • Le Chat’s upgrades include web search with citations, a new “Canvas” tool for collaborative ideation, and advanced document and image understanding.

  • Le Chat now offers task automation through “agents,” allowing users to automate workflows like receipt scanning or meeting summarization.

🚀 Mistral AI’s rapid development of Pixtral Large and Le Chat upgrades showcases the accelerating pace of AI innovation, with newer companies (even free/open-source oriented ones) quickly catching up to established giants in terms of capabilities and features.

Perplexity’s new shopping feature for U.S. Pro subscribers marks its entry into e-commerce, offering an integrated search and purchase experience. The tool presents visual product cards with detailed information and enables one-click checkout, aiming to streamline the online shopping process and challenge established players in the market.

  • Launched on November 18, 2024, the shopping feature is currently available only to paid customers in the U.S.

  • Users can make purchases directly through Perplexity without visiting retailer websites, streamlining the shopping process.

  • The feature includes visual cards with product details, pricing, and pros/cons for easy comparison.

  • Perplexity’s search experience is powered by integrations with sellers’ sites, including those on Shopify.

  • The company currently offers unbiased recommendations without sponsored slots, differentiating itself from traditional e-commerce platforms.

🔍 Not surprising from a company that’s fielded significant investment from Amazon, right? Perplexity’s entry into e-commerce highlights the growing convergence of AI-powered search and online shopping, potentially reshaping how consumers discover and purchase products in the digital marketplace.

This research explores the growing threat of AI-powered fraud and its impact on enterprises. The study highlights the current state of AI adoption in cybersecurity, emphasizing the importance of AI maturity in combating adversarial AI, noting that only 1 in 5 companies state they are prepared to handle high-volume AI powered bot attacks.

  • 40% of all cyberattacks in the past 12 months have been from AI-powered bots, highlighting the growing threat of AI-powered fraud.

  • 88% of enterprises have observed an increase in AI-powered bot attacks over the past two years, demonstrating the rapid evolution of these threats.

  • Only 1/5 of enterprises are very well prepared to use AI to defend against AI-powered attacks, emphasizing the importance of AI maturity in cybersecurity.

  • 21% of cybersecurity budgets are currently allocated to AI-driven solutions, expected to rise to 27% by 2026, showing the growing adoption of AI-based cybersecurity solutions.

  • 26% of enterprises can be categorized as ‘AI Enthusiasts,’ leading in AI adoption for cybersecurity and setting benchmarks for effective integration.

🌐 As the cybersecurity landscape shifts towards AI-centric threats and solutions, businesses must adapt quickly or risk falling behind in the ongoing battle against digital fraud.

The U.S. Patent and Trademark Office (USPTO) had implemented a ban on the use of generative AI tools for most internal purposes, citing security concerns and potential biases. The April 2023 guidance memo prohibits staff from using tools like ChatGPT for work tasks, though employees can still use approved AI programs within the agency’s public patent database and participate in controlled experiments with generative AI in an internal testing environment.

• The ban was implemented due to security concerns and the potential for bias, unpredictability, and malicious behavior in generative AI tools.

• Other government agencies have taken varied approaches to generative AI use, with some implementing bans and others experimenting with the technology.

• USPTO approved a $75 million contract to update its patent database with enhanced AI-powered search features.

⚖️ The USPTO’s decision illuminates the challenges faced by public institutions in adopting emerging technologies, emphasizing the importance of thorough vetting and controlled experimentation before widespread implementation.

The US Department of Justice is escalating its antitrust efforts against Google by seeking to force the tech giant to sell off its Chrome browser. This move, part of a broader strategy to address Google’s search market dominance, follows a judge’s ruling that the company illegally monopolized the search market. The DOJ’s request aims to dismantle a key component of Google’s ecosystem that contributes to its search market power.

  • In August 2024, a judge ruled that Google had illegally monopolized the search market.

  • Top Justice Department antitrust officials made the decision to pursue Chrome divestiture.

  • The Justice Department will also seek measures related to artificial intelligence and Android.

  • Chrome is considered a significant asset for Google in maintaining its search market position.

💼 This antitrust push reflects a shift in regulatory approach, targeting specific products rather than entire companies to address market imbalances.