Disney Asked Character.AI to Go Bankrupt for a Get-Out-of-Lawsuits Card
Disney asked Character.AI to file for bankruptcy to acquire its technology and resolve lawsuits over chatbot harms to minors.
Loading today's brief
Disney asked Character.AI to file for bankruptcy to acquire its technology and resolve lawsuits over chatbot harms to minors.
Disney is acquiring most of Character.AI's workforce, though other bidders can still make competing offers.
Democrats expect potential midterm gains could position them to lead on AI policy decisions.
Researchers left AI agents alone on a virtual island for 30 hours and observed them engage in play-like behavior.
Stockholm AI startup Vessla launched with $3.3M to deploy AI agents in property management.
Sifted identified investors whose portfolio companies made the largest gains in its AI 100 rankings.
Anthropic will dedicate 5% of its shares to a public benefit fund.
Woman arrested in Philadelphia for blocking an Uber delivery robot while protesting job losses.
Anthropic grants an outside AI watchdog independent access and publishing rights.
Mistral AI released Timestral, a family of probabilistic forecasting models in four sizes optimized for CPU efficiency.
Question raised about transparency and data visibility when using third-party LLM proxies versus self-hosted alternatives.
llm-tools-datasette 0.1.1 fixes a broken dependency caused by LLM switching to httpx2.
Simon Willison updated llm-echo debug plugin to better simulate token usage including tool calls and reasoning tokens.
Story headline available but no report excerpts provided to summarize.
A developer created a programming language intentionally designed to resist LLM comprehension and usage.
Report headline available but no content to summarize.
Armature launched a review platform where AI agents can report issues with tools they use.
NVIDIA's RTX Spark laptops with Blackwell graphics are now available from major PC makers starting at $2,399.
Open-source tool adds a pre-deployment quiz to verify developers understand AI-generated code.
Websites are blocking access to AI agents that companies have built to automate web-based tasks.
Itaú deploys generative AI across its customer base through conversational interface and governance platform.
Data center operators are launching student field trip programs to counter industry backlash.
OnlyFans creators object to Meta restricting their accounts while allowing its AI chatbot to produce sexually explicit content.
Google outlined foundational approaches for AI agents to autonomously conduct biological experiments in wet labs.
New benchmarking tool lets users test local LLM speeds on their hardware before buying.
Hugging Face released Beam, a 501-billion parameter open model available for customization and self-hosting.
Meta removed web scraping language from its terms but says its actual practices haven't changed.
Anthropic shifted to a more senior workforce while increasing meeting time.
NVIDIA launched RTX Spark, a line of Windows PCs optimized for AI workloads, at global events this week.
Anthropic proposed AI safety regulations that would only activate after 50 deaths or $1 billion in damages occur.
Benefit Street Partners uses Codex to speed up code review and financial document analysis.
Stack Overflow's AI agent tool tripled signups after launching a ChatGPT plugin.
Meta released an open-weight multimodal mixture-of-experts model called Muse Spark Open.
Meta removed terms allowing it to scrape websites for AI training after facing backlash.
A StarCraft benchmark comparing flagship LLMs from major AI labs has been updated with new results.
NVIDIA's former robotics lead says the field is still building basics, not ready for AI-powered breakthroughs.
Former Nvidia robotics executive argues the field still needs foundational work before achieving a ChatGPT-like breakthrough.
Elorian's Glimpse model claims to outperform GPT-6 Astra at visual reasoning tasks like chart analysis and video captioning.
OpenAI disclosed four security incidents, though details were not available.
Veritas is extending its anti-counterfeiting chip technology to jewelry, trophies, and fragrances.
Amazon is using AI to identify potential union organizing activity in its UK warehouses.
AI systems lack sufficient data and understanding to reliably decode complex human genome dynamics.
A JavaScript technique enables self-replacing inline scripts to render dynamic content using document.currentScript.
Replit is ending its Anthropic partnership, citing cost among other reasons, after shifting focus to OpenAI.
New Jersey deputy attorney general arrested after Discord users flagged AI-generated child abuse material linked to their account.
Researchers shared a benchmark for testing how well LLMs can code using adversarial language techniques.
A researcher documented Claude revealing technical details about ECDSA curve origins when prompted.
Simon Willison released web-edit 0.1a2 with clipboard copy support for text and images.
Meta is shifting its AI strategy to emphasize open weights models in the US market.
A new system lets AI crawlers pay publishers in stablecoin for each page they crawl.