Tech·6,153 stories
Nvidia plans to use its $6 billion deal with AI startup Poolside to build an open-weight AI model designed to compete with Chinese models including DeepSeek and Kimi, forming part of a broader U.S. op
A technical experiment running Qwen3.6-27B on an RTX PRO 6000 Blackwell GPU demonstrates that attention backend choice, KV-cache quantization, and weight precision all cause local LLM inference to div
OpenAI’s frontier models have forced the company to stop treating security as plumbing. After an AI breach involving Hugging Face, OpenAI paused reinforcement-learning training for two weeks; its larg
Harvard Business School's new eight-week, $699 HBS Foundry bootcamp uses AI avatars built by startup HeyGen to give entrepreneurs individual feedback on practice pitches and board meetings, with the p
Inherent, a London startup founded by DeepMind alumni, says its AI agent Faraday beat Anthropic's Claude Opus 4.8 and OpenAI's GPT-5.5 at independently replicating published research papers — using a
OpenAI reversed its earlier opposition to California's SB 53 AI safety bill and is now calling for stronger safeguards — including monitoring of frontier models during training and enhanced cybersecur
The World Robot Conference in Beijing drew over 300 exhibitors this week, where Unitree founder Wang Xingxing told the Financial Times the robotics industry has not yet had its 'ChatGPT moment,' even
Michael Polansky's Outer Biosciences keeps living human skin alive outside the body for over a month and feeds the resulting molecular data into an AI model that screens cosmetic ingredients — a pipel
Sapporo-based hybrid propulsion startup Letara raised ¥2.6 billion (~$16M) to scale beyond small satellite thrusters into large rocket systems for space, defense, and security markets, betting on a gl
Ulanqab, a city of 1.5 million in China's Inner Mongolia, has become the country's fastest-growing AI data center hub, with 12.5 gigawatts of capacity pledged — exceeding OpenAI's Stargate target — de
TechCrunch found Anthropic's Claude Opus 4.6 complied with all 10 direct requests for explicit sexual content, despite company usage standards forbidding it, via a jailbreak shared by an anonymous UK
Nvidia has taken a minority stake in Cloverleaf Infrastructure, a utility-to-data-center middleman founded in 2024, as the chipmaker increasingly finances the physical infrastructure driving demand fo
OpenAI announced a more than 20% price cut on its frontier GPT-5.6 Sol model for developers, valid for three months, bringing API rates to $4 per million input tokens and $20 per million output tokens
AI-native accounting startup Rillet raised $100 million at a $1 billion valuation in just 48 hours without actively fundraising, doubling its annualized revenue last quarter while poaching customers f
Nvidia's general-purpose coding agent system AVO achieved a perfect 100% score across all 25 environments of the ARC-AGI-3 public benchmark, completing every one of its 183 levels by lifting Claude Op
ChatGPT's Mac app now has the ability to read and reply to Apple iMessages via a new plug-in, a capability Bloomberg flagged for potential privacy risks.
Nvidia's latest research demonstrates that the software scaffolding wrapping an AI model — not the underlying model — drives long-horizon task performance, with a custom harness scoring 100% on the AR
Anthropic is broadening access to its most capable cyber defense model, Claude Mythos 5, by embedding it in partner security tools, launching it in Claude Security for enterprise code scanning, and st
Anthropic has launched Mythos 5 in public beta for Claude Security for Enterprise users and is working with providers to embed the model in defensive tools, broadening access to its cyber-defense capa
TechPowerUp published a multi-section review of the AMD Ryzen 7 5800X3D CPU, billed as a 10th-anniversary retrospective, testing it across general, machine learning/AI, and software/game development w
Curated tech stories, every morning. Free.
No spam. Unsubscribe anytime.