Discover quality AI tools
Bring together the world’s leading artificial intelligence tools to help improve work efficiency
Articles published by RobertGreen
US Judge Mark Pittman ordered X and SpaceX AI to disclose settlement terms with Apple in their antitrust case. The companies previously accused Apple and OpenAI of colluding to limit competing AI apps via App Store rankings. While X and SpaceX AI dismissed claims against Apple without revealing terms, OpenAI sought transparency. The judge mandated the release of any agreements by September 17. Specific terms and Apple's commitments remain undisclosed. This legal battle highlights tensions over AI integration in Siri and Apple Intelligence, with all parties denying allegations of anti-competitive behavior. The outcome may impact future AI app visibility on iOS.
US Judge Mark Pittman ordered X and SpaceX AI to disclose settlement terms with Apple in their antitrust case. The companies previously accused Apple and OpenAI of colluding to limit competing AI apps via App Store rankings. While X and SpaceX AI dismissed claims against Apple without revealing terms, OpenAI sought transparency. The judge mandated the release of any agreements by September 17. Specific terms and Apple's commitments remain undisclosed. This legal battle highlights tensions over AI integration in Siri and Apple Intelligence, with all parties denying allegations of anti-competitive behavior. The outcome may impact future AI app visibility on iOS.
AI company MiniMax launched its productivity tool MiniMax Design on August 20, advancing multimodal content creation by shifting from pixel manipulation to semantics handling. Built on its H3 multimodal video model, the tool allows users to describe creative goals in natural language rather than adjusting each pixel manually, with the model automating task breakdown and output generation using its strong understanding and reference control capabilities for project context. It has shown practical value in commercial use (such as mass marketing material creation, lesson plan to video conversion, and shot preview for filmmakers) and supports integration with existing workflows like ComfyUI. Industry analysts say it addresses controllability and coherence issues in multimodal model deployment, offering a new approach for industrialized video content production. The tool is available for download at design.minimaxi.com.
AI company MiniMax launched its productivity tool MiniMax Design on August 20, advancing multimodal content creation by shifting from pixel manipulation to semantics handling. Built on its H3 multimodal video model, the tool allows users to describe creative goals in natural language rather than adjusting each pixel manually, with the model automating task breakdown and output generation using its strong understanding and reference control capabilities for project context. It has shown practical value in commercial use (such as mass marketing material creation, lesson plan to video conversion, and shot preview for filmmakers) and supports integration with existing workflows like ComfyUI. Industry analysts say it addresses controllability and coherence issues in multimodal model deployment, offering a new approach for industrialized video content production. The tool is available for download at design.minimaxi.com.
Anthropic shared lessons from building three AI products: claude.ai uses gVisor temporary containers for general users, Claude Code uses OS-level sandboxing reducing permission prompts by 84%, and Claude Cowork uses VM-level isolation for enterprises. Key principle: prioritize environment-layer isolation. Prompt injection via phishing achieved 96% success in tests, driving ongoing security improvements.
Anthropic shared lessons from building three AI products: claude.ai uses gVisor temporary containers for general users, Claude Code uses OS-level sandboxing reducing permission prompts by 84%, and Claude Cowork uses VM-level isolation for enterprises. Key principle: prioritize environment-layer isolation. Prompt injection via phishing achieved 96% success in tests, driving ongoing security improvements.
OpenAI shifts ChatGPT ads to cost-per-click billing from CPM aiming to compete with Google and Meta It introduces action-oriented ads and improves conversion tracking to address advertiser concerns Early ad CPMs are low at 15-25 dollars with revenue projected to hit 11 billion by 2027 The company has launched a self-service platform lowered spending thresholds but faces challenges in brand safety and precise targeting
OpenAI shifts ChatGPT ads to cost-per-click billing from CPM aiming to compete with Google and Meta It introduces action-oriented ads and improves conversion tracking to address advertiser concerns Early ad CPMs are low at 15-25 dollars with revenue projected to hit 11 billion by 2027 The company has launched a self-service platform lowered spending thresholds but faces challenges in brand safety and precise targeting
OpenAI reportedly considered a controversial strategy to leverage geopolitical tensions and create a prisoner's dilemma for governments, pushing them to fund the company to avoid falling behind in AI. The plan, described internally as crazy, faced severe employee backlash and threats of collective resignation, leading to its abandonment. OpenAI has denied the allegations, calling them absurd. The report emerges amid intense AI industry competition and ethical scrutiny.
OpenAI reportedly considered a controversial strategy to leverage geopolitical tensions and create a prisoner's dilemma for governments, pushing them to fund the company to avoid falling behind in AI. The plan, described internally as crazy, faced severe employee backlash and threats of collective resignation, leading to its abandonment. OpenAI has denied the allegations, calling them absurd. The report emerges amid intense AI industry competition and ethical scrutiny.
Google DeepMind CEO Demis Hassabis expressed surprise at OpenAI’s early move to test ads in its chatbot, stating Google is considering monetization very carefully. He noted there are no current plans for ads in Google’s AI chatbot, emphasizing the need for user trust. Hassabis questioned how ads fit the assistant model, contrasting it with search.
Google DeepMind CEO Demis Hassabis expressed surprise at OpenAI’s early move to test ads in its chatbot, stating Google is considering monetization very carefully. He noted there are no current plans for ads in Google’s AI chatbot, emphasizing the need for user trust. Hassabis questioned how ads fit the assistant model, contrasting it with search.
Researchers launch the Massive Sound Embedding Benchmark MSEB an open source platform to evaluate AI sound intelligence It unifies eight core audio capabilities like retrieval and classification across diverse datasets MSEB reveals significant performance headroom in current models and aims to accelerate the development of universal sound embeddings for AI systems
Researchers launch the Massive Sound Embedding Benchmark MSEB an open source platform to evaluate AI sound intelligence It unifies eight core audio capabilities like retrieval and classification across diverse datasets MSEB reveals significant performance headroom in current models and aims to accelerate the development of universal sound embeddings for AI systems
Tinders parent company Match Group is testing an AI feature called Chemistry to analyze user photos and interests aiming to boost subscriptions after nine straight quarters of paying user declines The feature piloted in Australia and New Zealand may access private camera rolls and has contributed to a 14 million revenue impact in Q4 testing
Tinders parent company Match Group is testing an AI feature called Chemistry to analyze user photos and interests aiming to boost subscriptions after nine straight quarters of paying user declines The feature piloted in Australia and New Zealand may access private camera rolls and has contributed to a 14 million revenue impact in Q4 testing
