Perplexity has launched Deep Research, a free AI-powered research tool that can analyze hundreds of sources in minutes to create comprehensive reports across various domains, promising to save users significant research time.
Founded by Former DeepMind scientist Simon Kohl, Latent Labs has emerged from stealth with $50M in funding to develop AI models for protein design. Latent Labs aims to accelerate drug discovery through partnerships with biotech and pharmaceutical companies.
Generative AI infrastructure startup fal has raised a $49M Series B led by Notable Capital and Andreessen Horowitz to scale its video-focused AI platform. fal already serves over 50 enterprise customers including Quora and Canva, and processes over 100 million inference requests daily.
BBC research reveals that leading AI tools including ChatGPT, Perplexity, Microsoft Copilot, and Google Gemini produced significant factual errors in about 20% of responses and altered or fabricated quotes in over 10% of cases when answering news questions using BBC articles as sources.
A recent experiment by LangChain tests the performance of five leading LLMs by Anthropic, OpenAI, and Meta within single-agent architectures for single-domain tasks. The findings reveal that performance degrades as the domains to manage, context complexity, and available tools increase.
Thousands of artists have signed a letter protesting Christie's first major AI art auction, claiming the models used to create many artworks were illicitly trained on copyrighted material. Christie's and some featured artists have defended the auction, suggesting the criticism is misguided.
Google has integrated NotebookLM Plus, the wildly popular AI-powered notebook and research assistant, into its Google One AI Premium subscription plan at no additional cost. In addition, the company also introduced a 50% discount on Google One AI Premium for US-based students under 18.
DeepSeek released Janus-Pro, a new iteration of its image generation models; OpenAI strikes strategic partnerships in Asia; Mistral AI has enhanced its Le Chat assistant with new features; Eleven Labs raised a $180M Series C; Tülu 3 showcases Ai2's open source post-training recipe; and more.
A new benchmark based on the NPR's Sunday Puzzle riddles aims to test LLMs' general reasoning skills. The findings are remarkable: reasoning models like OpenAI's o1 do best at the benchmark, and some replicate behaviors such as "giving up" or showing "frustration" when stuck on difficult problems.
AlphaGeometry2 (AG2) is an enhanced neuro-symbolic system for solving geometry problems, which Google DeepMind claims outperforms IMO gold medalists. Although AG2 has some limitations, it undoubtedly has taken DeepMind one step closer to "solving geometry".
Mistral is expanding its AI assistant Le Chat with new iOS and Android apps, a Pro subscription tier, and enterprise features while competing with established players through fast performance and unique capabilities like on-premises deployment.
Tana, an AI-powered productivity app founded by ex-Google employees, has emerged from stealth after gathering $25 million in funding and a 160,000-person waitlist for its platform that automatically organizes user information through a knowledge graph system.