A.I. Tracker 🤖

Meta is leading in AI development with Llama 4, using a cluster of 100,000 H100 chips for training. Their open-source approach with Llama models is gaining popularity among startups and researchers. Other competitors like OpenAI and Google are also developing advanced AI models with new capabilities. Energy consumption and costs are significant challenges.

AI models struggle to accurately answer election-related questions in Spanish, with a higher error rate compared to English. This bias poses potential harm.

Panelists at TechCrunch Disrupt 2024 discuss combating disinformation, highlighting the dangers of generative AI in creating and spreading fake content on social media. Calls for stronger regulation and oversight while acknowledging the beneficial aspects of AI.

The article warns of potential future AI culture wars that could further deepen political divides in the United States. It suggests that current divisions may pale in comparison to the potential conflicts arising from differing views on AI technology.

TechCrunch's AI newsletter previews content from their Disrupt conference featuring speakers from the AI industry and discussions on disinformation, legal ethics, and new AI capabilities. Google's AI models convert handwriting to digital ink, while Cohere releases multilingual text-generating models.

Betaworks' recent Camp focused on AI-native applications. Startups included Unternet, Ursula, ESAI, Autoplay, Alice Camera, Dessn, Sarama, Tato, and Hopscotch Labs, aiming to improve user experiences and efficiency with AI technology.

OpenAI's Whisper transcription tool is widely used in healthcare, but experts are concerned about its accuracy. Precision is crucial in medical settings, raising alarms about potential risks.

Alphabet, parent company of Google, reports 15% revenue growth, exceeding analyst expectations with strong performance in ads and cloud services. Legal battles continue, including antitrust cases and ongoing expansion of Waymo's self-driving car business. CEO Pichai remains optimistic about future innovation and engagement with regulators.

Ashton Kutcher’s VC firm Sound Ventures is investing in AI startups like Fei-Fei Li’s World Labs and exploring AI hardware. They believe in the potential of AI technology and are looking to disrupt existing players like Apple.

Ashton Kutcher believes every company will become AI-driven. He doesn't think there will be a single winner in the AI space. Sound Ventures invests in foundational AI companies. Kutcher encourages founders to focus on people and market insight, not just their pitch deck.

Google's Project Astra aims to build AI apps and agents for real-time understanding. CEO Sundar Pichai plans to ship experiences by 2025, delaying consumer-focused agent launch until next year at the earliest. Anthropic faces challenges with generative AI models.

Google relies heavily on AI for its products and revenue, with a quarter of new code generated by AI. Strong earnings were reported for Google Services and Google Cloud, with AI playing a key role in the company's success. Despite challenges, such as a ruling of monopolistic behavior, Google continues to innovate with AI-powered features and products.

OpenAI's ChatGPT has become widely used for various tasks but faces controversies like copyright issues and misinformation. New updates and features aim to improve user experience and safety.

OpenAI is partnering with TSMC and Broadcom to develop an in-house AI chip, and will use AMD chips alongside Nvidia's for training due to shortages and costs.

Elon Musk's AI company, xAI, is in talks to raise $40 billion in funding, nearly double its current valuation. xAI has developed Grok model for X network and offers enterprise API. It's training next-gen models on 100,000 Nvidia GPUs.

OpenAI plans to use AMD chips in Microsoft Azure for now, but is working with Broadcom to develop custom silicon for AI workloads, potentially starting production in 2026. The company aims to compete with tech giants in AI server hardware.

Narada AI is a startup creating an AI assistant for enterprise tasks, based on new research from UC Berkeley. The assistant can execute tasks across multiple work apps and uses a unique LLM Compiler technique. It can draft emails, make calendar invites, and interact with websites without APIs. The company emphasizes data privacy and is focused on enterprise applications.

Cyril Gorlla taught himself to code, founded CTGT to improve AI safety by identifying biases, has Fortune 10 clients, and aims to grow in the $16.2 billion explainable AI market.

Amazon extends AI shopping assistant Rufus to more markets in Europe and the Americas, offering product search, comparison, and recommendations. Customers can access Rufus through the Amazon Shopping app, but it is still in early stages and will continue to improve over time.

LinkedIn launches Hiring Assistant, an AI tool for recruiters. It automates tasks like job descriptions, candidate sourcing, and interaction. Aimed at B2B market, it enhances user experience with future updates planned.