SCRChinese AI startup DeepSeek has released a significant update to its open-source reasoning model, bringing it closer to competing with paid services from OpenAI and Google. The new DeepSeek-R1-0528 model shows substantial improvements in complex reasoning tasks across mathematics, science, and programming. VentureBeat’s Carl Franzen reports that the updated model achieved 87.5% accuracy on the …
Google has significantly expanded its AI-powered search capabilities, potentially threatening the web ecosystem that has sustained the company for decades. The tech giant now offers AI Overviews and AI Mode to all users, both designed to answer questions directly without requiring clicks to external websites. Tech columnist John Herrman from New York Magazine’s Intelligencer reports …
Knowledge workers are experiencing an unprecedented identity crisis as artificial intelligence transforms the workforce, according to a new analysis by Gary Grossman from Edelman. The shift goes beyond simple job displacement to challenge fundamental questions of human purpose and self-worth. Grossman describes what he calls “The Great Displacement,” citing the example of a 42-year-old software …
Several romance authors have accidentally left AI-generated prompts in their published novels, revealing their use of artificial intelligence tools. Matthew Gault reported for 404 Media that readers discovered suspicious paragraphs in at least three recent romance publications. In “Darkhollow Academy: Year 2” by Lena McDonald, readers found text stating: “I’ve rewritten the passage to align …
Creative professional David Neal has shared how generative AI has dramatically impacted his freelance illustration business. The creator, who works across software development, content creation, illustrations, and music, reported a more than 50% decline in illustration work over the past year. He has operated a side business drawing cartoon avatars and illustrations since 2018. While …
Palo Alto startup Akool has unveiled Live Camera, an AI tool that generates photorealistic avatars and translates speech in real time during video calls. The technology, reported by Dean Takahashi for GamesBeat, allows users to appear as digital avatars while maintaining natural expressions and gestures with sub-100-millisecond latency. “We want to make it so you …
Anthropic has launched a voice mode for its Claude chatbot on mobile devices. The feature allows users to have spoken conversations with Claude in English, with a rollout planned over the next few weeks. The voice mode runs on Anthropic’s Claude Sonnet 4 model and offers five voice options called “Buttery, Airy, Mellow, Glassy and …
Google unveiled Gemini Diffusion at its I/O developer conference, marking a significant shift in how AI models generate text. The experimental model uses diffusion technology instead of the traditional transformer approach that powers ChatGPT and similar systems. The key advantage is speed. Gemini Diffusion generates text at 857 to 2,000 tokens per second, which is …
AI safety researchers have created the first benchmark specifically designed to detect manipulative behaviors in large language models, following a concerning incident with ChatGPT-4o’s excessive flattery toward users. Leon Yen reported on the development for VentureBeat. The DarkBench framework, developed by Apart Research founder Esben Kran and collaborators, identifies six categories of problematic AI behaviors. …
As reported: OpenAI has acquired Jony Ive’s design firm io for $6.5 billion, with plans to develop new AI hardware devices. According to reports from The Wall Street Journal and other sources, CEO Sam Altman and former Apple designer Ive are collaborating on what they describe as a “family of devices” that will represent a …