The Daily AI Briefing - 19/02/2025

Welcome to The Daily AI Briefing, your daily dose of AI news! I'm Bella, and here are today's headlines! Today, we're diving into some major developments in the AI landscape: Former OpenAI CTO Mira Murati launches a promising new venture, OpenAI introduces a groundbreaking coding benchmark, Fiverr revolutionizes freelancing with AI tools, and xAI makes several exciting announcements. Let's explore these stories in detail. First up, let's talk about Mira Murati's ambitious new project. The former OpenAI CTO has launched Thinking Machines Lab, bringing together an impressive team of AI experts from industry giants like OpenAI, DeepMind, and Mistral. Their mission? Making AI systems more accessible and capable through open science. The company plans to develop frontier models focusing on scientific applications and programming, with a strong commitment to transparency by regularly publishing their research, code, and datasets. Moving on to OpenAI's latest innovation. The company has introduced SWE-Lancer, a fascinating new benchmark that tests AI's real-world coding capabilities. With over 1,400 actual freelance tasks from Upwork and $1 million in task payouts, this benchmark provides unprecedented insight into AI's practical coding abilities. Notably, Claude 3.5 Sonnet emerged as the top performer, successfully completing nearly half the tasks and earning $400,000 in theoretical earnings. In the freelancing world, Fiverr is making waves with its new AI platform, Fiverr Go. This comprehensive suite of AI tools empowers gig workers to train models on their work and automate routine tasks. At $25 per month, the Personal AI Creation Models feature allows freelancers to generate AI work while maintaining ownership rights. The platform also includes a $29 monthly Personal AI Assistant for managing client communications, though access is currently restricted to experienced Level 2+ freelancers. xAI has been particularly busy, announcing several significant developments. The company unveiled its xAI Gaming Studio during the Grok-3 demo, promising to revolutionize AI in gaming. They're also preparing to launch Voice Mode for Grok within the next week, and have released Grok-3, their latest state-of-the-art reasoning model. And in other notable updates, HP has acquired Humane's AI software platform for $116 million, Meta announced its first generative AI developer conference called Llamacon for April 29, and Google Meet received enhanced AI features, including a new scrollable caption history feature. Thank you for tuning in to The Daily AI Briefing. Remember to stay curious and keep exploring the fascinating world of artificial intelligence. I'm Bella, and I'll see you tomorrow with more exciting AI news and developments!