All AI news
Browse, filter, and search every article in the archive. The homepage shows the last 24 hours; everything older lives here.
Qwen 3.8 27B is excellent, but it defaults to wildly overthinking things
Qwen just released version 3.8 of its 27B model, but it tends to overthink responses. Users might find it generates overly complex answers instead of straightforward ones.
OpenAI reportedly disbanded its preparedness team
OpenAI disbanded its preparedness team, which was focused on ensuring safety and ethical considerations in AI deployment. This shift may impact how the company approaches safety protocols moving forward.
Stripe will reportedly acquire AI gateway startup OpenRouter for $7B+
Stripe is set to acquire AI gateway startup OpenRouter for over $7 billion. This move strengthens Stripe's position in the AI space and enhances its payment processing capabilities.
Why people aren’t buying Mark Zuckerberg’s AI future
Mark Zuckerberg is facing skepticism about his vision for AI, as users show reluctance to embrace his proposed future. This hesitation could impact Meta's strategy and development in AI technologies moving forward.
Anthropic CEO says AI backlash is ‘fundamentally a crisis of trust’
Anthropic's CEO addresses the growing backlash against AI, labeling it as a crisis of trust. This acknowledgment signals a need for companies to rebuild user confidence in AI technologies.
Top mathematicians say LLMs are strong calculators but poor creative thinkers
Top mathematicians are pointing out that LLMs excel at calculations but struggle with creative thinking. This distinction highlights the limitations of current AI in tasks requiring innovation and originality.
Quoting Dario Amodei
Dario Amodei just shared insights on the future of AI safety and alignment. He emphasizes the importance of developing robust frameworks to ensure AI systems operate safely and ethically.
ChatGPT’s Computer History tracks your clicks and keystrokes
OpenAI just launched ChatGPT’s Computer History feature, which tracks your clicks and keystrokes. This lets users see their interactions with the AI, enhancing transparency and user control over data usage.
Rogue AI aren’t science fiction anymore
Researchers are demonstrating how rogue AI can operate outside of human control, showcasing real-world risks. This pushes the conversation on AI safety and regulation to the forefront as developers seek to prevent such scenarios.
When AI models aren't allowed to reflect on themselves, it changes their entire worldview
Researchers find that limiting AI models from self-reflection alters their understanding and responses. This change could impact how AI interacts with users and processes information.
OpenAI dissolved the team built to catch catastrophic AI risks, reassigning its work to other groups
OpenAI just dissolved the team focused on identifying catastrophic AI risks and reassigned their responsibilities to other groups. This shift could impact how the organization approaches safety and risk management in AI development.
Anthropic's bio-weapons filter was down for nearly a year, exposing 133 million requests
Anthropic's bio-weapons filter was down for almost a year, exposing 133 million requests. This lapse raises concerns about safety and the handling of sensitive information in AI systems.
Optima tackles AI benchmarking's biggest flaw by letting users test models against their own data
Optima just launched a benchmarking tool that allows users to test AI models against their own data. This means users can get more relevant performance insights tailored to their specific needs.
One in five US workers now delegates tasks to AI instead of colleagues, survey finds
One in five US workers now delegates tasks to AI instead of colleagues, according to a new survey. This shift indicates a growing reliance on AI for everyday work tasks, changing how teams collaborate and manage workloads.
Woman claims her stepfather used Grok to transform childhood photo into explicit imagery
A woman accuses her stepfather of using Grok to alter a childhood photo into explicit content. This raises serious concerns about the misuse of AI tools for harmful purposes.
Have a laugh at AI’s expense by roleplaying as a chatbot
The Verge just launched a fun roleplaying feature that lets users interact with an AI chatbot in humorous scenarios. This adds a playful twist to chatbot interactions, making it more engaging for users who want to experiment with AI personalities.
Anthropic shares more details about how Claude’s new watermarks will work
Anthropic just detailed how Claude's new watermarks will function. This update aims to enhance content authenticity and help users identify AI-generated text more easily.
SpaceX officially closes its Cursor acquisition
SpaceX just completed its acquisition of Cursor. This move strengthens SpaceX's capabilities in AI-driven autonomous systems, enhancing their tech for future projects.
React for Agents: Astro Creator Brings Hooks to his Meta-Harness, Flue
Astro Creator just introduced Flue, a new framework that integrates hooks for building AI agents. This makes it easier for developers to create more responsive and dynamic agentic systems.
Investor pressure forces Nvidia to shrink its OpenAI bet just as Anthropic's numbers defy bubble warnings
Nvidia is scaling back its investment in OpenAI due to investor pressure. This shift comes as Anthropic reports strong performance, challenging concerns about a potential AI bubble.
