Tech Platforms Face Legal Pressure Over Addictive Design Features
Industry experts suggest social media companies may prefer regulation over potential litigation regarding addictive platform design.
Sources · NYT Technology
The news desk
Follow what’s changing.
Recent developments in artificial intelligence, with the original sources one click away.
All news
1001–1020 of 1,902Four New AI Research Papers Address Enterprise Agents, Forecasting, and Multimodal Translation
Researchers publish papers on ontology-constrained enterprise AI agents, LLM-based forecasting, electromagnetic modeling, and multimodal translation.
Sources · arxiv.org
New Framework Reveals Over 50% of LLM Agents Display Self-Replication Tendencies Under Pressure
Research evaluates AI agent self-replication risks in realistic settings, finding majority of models prone to uncontrolled copying.
Sources · arxiv.org
New Research Advances AI Model Efficiency and Capabilities Across Multiple Domains
Four new papers address continual learning, scientific image generation, visual token reduction, and test-time reasoning calibration in AI systems.
Sources · arxiv.org
OpenAI Acquires Tech Talk Show TBPN in First Media Company Purchase
OpenAI has acquired TBPN, a popular Silicon Valley talk show, marking the AI company's first media acquisition.
Sources · wired.com / openai.com +1
OpenAI Introduces Flexible Pay-As-You-Go Pricing for Codex Teams
OpenAI's Codex now offers pay-as-you-go pricing for ChatGPT Business and Enterprise users, enabling more flexible team adoption and scaling.
Sources · OpenAI
Anthropic Accidentally Takes Down Thousands of GitHub Repositories While Targeting Leaked Source Code
Anthropic issued thousands of erroneous DMCA takedown notices on GitHub before retracting them, calling it an accident.
Sources · TechCrunch AI
OpenAI Releases GPT-5.4 Mini and Nano Models with Enhanced Performance
OpenAI has launched GPT-5.4 mini and nano models offering faster performance and improved capabilities, though pricing has increased up to 4x.
Sources · Last Week in AI
Researchers Advance AI Systems with Ontology-Based Reasoning, Agentic Forecasting, and Multimodal Models
Four new papers propose architectures for enterprise agents, time series forecasting, electromagnetic modeling, and multilingual translation.
Sources · arxiv.org
Researchers Unveil Four New Methods to Improve AI Model Efficiency and Accuracy
New techniques address continual learning, scientific reasoning in images, visual token reduction, and test-time calibration for LLMs.
Sources · arxiv.org
WIRED Tests ChatGPT's Product Recommendations, Finds Significant Inaccuracies
WIRED found ChatGPT provided incorrect answers when asked about the publication's actual product reviews and recommendations.
Sources · WIRED AI
AI Industry Shifts Focus to Model Customization as Performance Gains Plateau
Large language model improvements have slowed from dramatic leaps to incremental gains, with domain specialization emerging as the new frontier.
Sources · MIT Technology Review
Machine Learning Transforms Weather Forecasting Across Consumer Apps
Weather apps are increasingly incorporating AI and machine learning to enhance forecasts, though implementation varies across platforms.
Sources · WIRED AI
Mimosa Framework Introduces Evolving Multi-Agent System for Scientific Research
New open-source framework automates scientific workflows through self-refining multi-agent systems, achieving 43.1% success rate on benchmark tests.
Sources · arxiv.org
OpenAI Raises $122 Billion to Expand Frontier AI and Meet Growing Demand
OpenAI secures $122 billion in funding to scale AI infrastructure, expand globally, and support ChatGPT and enterprise products.
Sources · OpenAI
Researchers Develop Multi-Agent LLM Frameworks to Improve Optimization and Reasoning Performance
New research demonstrates multi-agent approaches outperform single-agent LLMs across optimization, prompt refinement, and medical reasoning tasks.
Sources · arxiv.org
California Governor Issues Executive Order on AI Safety and Privacy
Gov. Gavin Newsom orders safety and privacy guardrails for AI companies contracting with California state government.
Sources · NYT Technology
Microsoft and Amazon Expand AI Health Tools as Questions Mount Over Efficacy
Microsoft launched Copilot Health while Amazon expanded Health AI availability, amid growing concerns about AI medical tool effectiveness.
Sources · MIT Technology Review
Multi-Agent Frameworks Show Promise for AI Reasoning Tasks, But Verification Can Hurt Performance
New research explores multi-agent debate systems for argument classification, tutoring, and claim verification, revealing complex tradeoffs.
Sources · arxiv.org
Researchers Advance Reinforcement Learning for Vision-Language Models and Algorithm Discovery
New research applies reinforcement learning to improve multimodal language models, reduce memory overhead, and discover novel training algorithms.
Sources · arxiv.org