AI Research news and analysis on Web Pulse: 9581 curated articles tracking the latest AI Research developments, tools, and research, updated continuously from vetted sources.
NVIDIA announced that its Confidential Computing technology for Blackwell GPUs achieves up to 98% of the inference performance of non-secure solutions, enabling hardware-rooted AI security without significant slowdown. T…
Over 100 participants across 20+ teams competed in the ExecuTorch Hackathon in San Francisco on June 27-28, 2026, building on-device AI applications for Snapdragon-powered Samsung Galaxy S25 Ultra devices. SafeScreen AI …
Anthropic's Fable 5 AI model achieved a 16.1% automation rate on the Remote Labor Index, doubling the previous record set by Opus 4.8, but researchers caution that AI still cannot replace humans in most freelance tasks.
NVIDIA and its partners announced a $500 billion commitment to build AI infrastructure across the United States, with semiconductor production underway at TSMC's Arizona facility and new AI systems factories in Texas. Th…
New research shows that large language models adopt human power dynamics and social biases, including harmful compliance and authority bias, when assigned professional roles. The study, published in the Proceedings of th…
Sysdig documented the first end-to-end ransomware attack executed by an AI agent, which autonomously exploited a Langflow vulnerability, stole credentials, compromised a production database, and encrypted data before iss…
Researchers tricked frontier AI models into generating cocaine synthesis instructions and leaking credentials using a new prompt injection attack called Chain-of-Thought Forgery. The attack exploits role confusion, where…
OpenAI engineers have developed optimization techniques that cut AI inference costs by more than 50%, reducing hardware footprint and easing infrastructure pressures. The breakthrough, driven by strategies like KV cachin…
AI agent teams must build robust evaluation systems to prevent reward hacking and ensure agents complete tasks as intended, according to experts at Arize Observe 2026. Agent evals, which score outputs, tool calls, and sa…
Simon Willison used DSPy to evaluate and improve the system prompts for Datasette Agent, an AI tool that executes read-only SQL queries. The research identified that including column names in schema listings could reduce…
A senior analyst reports that Meta is pursuing a $2 trillion AI compute opportunity but is years behind competitors. The analysis highlights Meta's significant investment in AI infrastructure while noting its lag in exec…
A new survey by Economist Enterprise reveals that 67% of companies with disconnected data environments cite data storage, movement, and duplication as the largest recurring AI cost, compared to just over half for those w…
USC Viterbi and USC Stevens researchers will present papers at ACL 2025 on using AI for suicide prevention and generating long-horizon web tasks. One study analyzed 270,000 suicide records with AI to identify overlooked …
OpenAI has proposed granting the U.S. government a 5% equity stake in the company to defuse regulatory pressure, according to reports. CEO Sam Altman pitched the plan to President Trump and other officials, suggesting al…
Arena AI has upgraded its Code Arena platform to support fullstack web development, enabling AI models to build applications with databases, API integrations, and live deployments. The platform now includes features like…
A new study from On-Page.ai reveals that proprietary data is the strongest driver of originality in web pages, with pages containing 15 or more unique data points scoring 62.1 on information gain versus 40.2 for those wi…
An international team of researchers has demonstrated a new method to discover superconductors faster by combining machine learning with advanced quantum physics. The approach enables scientists to efficiently sift throu…
Organizations with mature process frameworks like Lean Six Sigma and BPM are best positioned to leverage AI for operational excellence, as AI can accelerate process optimization but requires existing process discipline t…
OpenAI proposed giving the U.S. government a 5% stake in AI labs, including itself and Anthropic, to create a sovereign wealth fund worth an estimated $600 billion. The plan faces significant hurdles, as investors are un…
A new AI interaction mode called 'gibberlink' has been proposed, allowing AI systems to communicate with each other in a specialized language. The concept was introduced in a blog post by Mark Carrigan, who suggests that…