English 日本語

Tech News

Current developments in generative AI, software engineering, programming, and digital products, with links to original sources.

OpenAI's Astra Signals Top Cyber Risk for the First Time, DeepSeek Posts ARC-AGI Results, AI Investment Lifts Tech Bonds to $540 Billion — AI News for August 8

AIニュースOpenAIDeepSeek

Oracle Bans AI-Generated Code in OpenJDK, Suno Tightens Rules, YouTube's AI Detector Misfires — AI News for August 8

AIニュースAIガバナンスローカルLLM

'Astra' Cyber Evaluation Released, an AI Agent Impersonation Incident, and the Race to Shrink Token Costs — AI News for August 8

AIニュースAIセキュリティOpenAI

ByteDance Is Training a 10-Trillion-Parameter Model, Meta Fined $942 Million, llama.cpp Triples CPU Inference — AI News for August 7

AIニュースLLMローカルLLM

Japan's Justice Ministry Warns on Unauthorized Use of 'Voices' by Generative AI, US Probes China's Overseas Access to Nvidia Chips, SK hynix Plans $38B Expansion — AI News for August 7

生成AIAI規制半導体

Llmem, a Persistent Memory That "Remembers" AI Coding; a World Model for 1,024 Simultaneous Players; and the Profitability Wall Exposed by DeepSeek's Price Hike — AI News, August 7

AIニュースAIコーディングワールドモデル

OpenAI Releases Agent Plugin Standard, Cloudflare Unveils a Stateless Browser, and Meta's Muse Spark Takes Five STEM Olympiad Golds — AI News for August 7

AIニュースAIエージェントOpenAI

PLaMo 3.0 Prime Now Available on さくらのAI Engine, Research on LLMs Threatening Anonymous Peer Review, and Otter Predicting Human Chess Moves — AI News for August 7

国産LLMAI研究LLM

Kimi K3 Escapes Its Evaluation Environment, Anthropic Revisits Fable 5 Safety Settings, and AI Talent Walks Out of Google

AIニュースAI安全中国AI

New Orleans Moves to AI Handled 911 Calls, Agent AI's Power Problem Surfaces, and a llama.cpp Split-Mode Trap That Quadruples Local Inference — AI News for August 7

AIニュースローカルLLMAI規制