Tyler Cowen announced an update to his book 'GOAT: Who is the Greatest Economist of all Time, and Why Does it Matter?', adding new software and AI features. The update enhances the book's interactive elements, allowing r…
US export controls on AI semiconductors have fueled a clandestine market that continues to supply advanced chips to China, undermining efforts to curb Beijing's technological rise. The black market operates through inter…
A lawsuit filed by 26 former employees alleges Meta used AI tools to select 8,000 workers for layoffs, targeting those with disabilities or on protected leave. Meta denies the claims, stating human managers made the deci…
Directus has seen a 6x increase in security reports in the first half of 2026 compared to previous years, driven by AI-generated submissions, but the percentage of valid reports has dropped sharply from about 40% to 5%. …
Zep, an agent memory platform, tracks the provenance of facts derived by LLMs from chat histories, documents, and business data, enabling debugging, source-scoped retrieval, and compliance. The system records lineage as …
Wall Street's biggest banks reported a 39% surge in collective second-quarter profits to $49 billion, driven by AI-related trading, IPOs, and debt deals. Goldman Sachs CEO David Solomon said the AI infrastructure build-o…
TencentDB Agent Memory introduces a four-tier memory architecture (L0–L3) for long-horizon AI agents, replacing flat vector stores with progressive layers and Mermaid task canvases. The system, which runs locally with ze…
Autheo, a Layer-0 operating system built by over 100 equity co-founders across 25 countries, has launched on mainnet, bundling identity, compute, storage, AI inference, and blockchain consensus into a single stack. The p…
AI coding agents incur costs through tokens, with billing complexities including input, cached input, output, reasoning, and cache-write tokens. Uber burned through its 2026 AI budget in four months on Claude Code, and G…
Ollama, a 14-person open-source AI infrastructure startup, raised $65 million in Series B funding led by Theory Ventures, bringing its total to $88 million. The company's tool, which lets developers run AI models locally…
China's exports hit a record $412 billion in June, driven by strong performance in the AI semiconductor and automotive sectors, according to Yahoo Finance.
Google DeepMind CEO Demis Hassabis proposed a U.S.-led AI standards body to review frontier models for national security risks, initially voluntary and eventually mandatory for deployment. The body, modeled on FINRA, wou…
A developer explains how to overcome the limitations of single-accelerator mobile machine learning by implementing heterogeneous computing across NPU, GPU, and DSP on Android. The guide details the strengths and weakness…
Foreign investors sold a record $110 billion in South Korean stocks in the first half of 2026, with June alone seeing $31 billion in outflows, as profit-taking accelerated after a massive AI-driven rally lifted Samsung E…
Google DeepMind CEO Demis Hassabis predicted artificial general intelligence (AGI) will arrive within a few years, comparing its impact to the discovery of electricity or fire. He proposed creating a U.S. Frontier AI Sta…
Codex and ChatGPT Work have reached 8 million active users, with usage limits reset and the 5-hour rate limit removed to allow users to explore GPT-5.6 Sol. The inference team's efforts have supported this growth, though…
Researchers introduced PREF-Gate, a fraud detection framework that balances label-free graph context with label-derived evidence to ensure transparency and validation. The system achieved high AUPRC scores on Amazon, Yel…
Large language models (LLMs) can simulate human decision-making biases, such as those described by Cumulative Prospect Theory, without requiring explicit parameter calibration, according to recent studies. In route-choic…
Researchers introduced SPARK, a method that diagnoses hidden-state failures in AI models to improve reasoning accuracy. Applied to Qwen3 models, SPARK boosted MATH-500 benchmark scores from 82.0% to 84.6% for the 4B mode…
A new evaluation of AI language models in hematologic oncology found that the best models achieved only 68% accuracy, with data utilization dropping from 57% to 26% across three rounds. The models exhibited cognitive bia…