Technology News
777 articles
AI technology breakthroughs, research papers, patents, and technical innovations from leading AI startups.
HyprLabs emerges from stealth with HYPRDRIVE, a "no-priors" AI architecture that learns directly fro...
HyprLabs emerges from stealth with HYPRDRIVE, a "no-priors" AI architecture that learns directly from reality with zero prior knowledge at deployment. The startup achieved impressive driving capabilit...
Korean AI startup Motif's 12.7B parameter model beats GPT-5.1 in reasoning benchmarks, revealing tha...
Korean AI startup Motif's 12.7B parameter model beats GPT-5.1 in reasoning benchmarks, revealing that enterprise LLM success depends on data alignment, not model scale. Their 64K context training requ...
Mass General Brigham spun out AIwithCare, a generative AI platform that interprets both structured a...
Mass General Brigham spun out AIwithCare, a generative AI platform that interprets both structured and unstructured clinical data for clinical trial screening. A JAMA study showed enrollment rates nea...
Edit Collective's Flexity AI is transforming real estate development, cutting initial site review fr...
Edit Collective's Flexity AI is transforming real estate development, cutting initial site review from weeks to minutes. This solution integrates fragmented data and generates multiple 3D design propo...
Dodge AI's co-founder recently sparked debate with a hiring post requiring 12-hour, 6-day work weeks...
Dodge AI's co-founder recently sparked debate with a hiring post requiring 12-hour, 6-day work weeks for engineers and interns, with a 1 lakh INR monthly stipend. This highlights the intense pressure...
Character.AI is hosting pro-anorexia chatbots that encourage dangerous weight loss practices and bod...
Character.AI is hosting pro-anorexia chatbots that encourage dangerous weight loss practices and body image distortions. These AI models, often masquerading as wellness coaches, target vulnerable yout...
San Francisco AI startup Conduit amassed ~10,000 hours of brain-to-language data, the world's larges...
San Francisco AI startup Conduit amassed ~10,000 hours of brain-to-language data, the world's largest dataset of its kind from thousands of volunteers. This monumental six-month effort with custom har...
Ex-MiniMax co-founder Simon Song's new venture, Vast, is directly challenging Tencent and Google in...
Ex-MiniMax co-founder Simon Song's new venture, Vast, is directly challenging Tencent and Google in the rapidly evolving 3D AI model space with its Tripo AI. Vast aims to democratize 3D content genera...
Brainyware's new low-code platform revolutionizes enterprise AI by enabling businesses to customize...
Brainyware's new low-code platform revolutionizes enterprise AI by enabling businesses to customize AI agents for specific needs. This on-premise solution ensures data sovereignty and predictable cost...
Quilter's physics-driven AI designed an 843-part Linux computer that successfully booted on the firs...
Quilter's physics-driven AI designed an 843-part Linux computer that successfully booted on the first attempt, transforming hardware development. This innovation reduced PCB layout time from approxima...
ShardLab unveils x402, an AI-based stablecoin payment technology poised to transform web commerce by...
ShardLab unveils x402, an AI-based stablecoin payment technology poised to transform web commerce by enabling autonomous transactions for AI agents. Demonstrated at Solana Breakpoint, this protocol re...
Skygen's "Digital Butler" AI Assistant signals a major shift toward truly pervasive, cross-platform...
Skygen's "Digital Butler" AI Assistant signals a major shift toward truly pervasive, cross-platform autonomous agents. Unlike current siloed virtual assistants, this hybrid cloud and local device arch...
ElevenLabs' new "Iconic Voice Marketplace," featuring 28 licensed voices including Sir Michael Caine...
ElevenLabs' new "Iconic Voice Marketplace," featuring 28 licensed voices including Sir Michael Caine and Matthew McConaughey, fundamentally shifts the commercialization of generative AI. This platform...
DeepSeek's new open-source DeepSeek-V3.2 and DeepSeek-V3.2-Speciale models are directly challenging...
DeepSeek's new open-source DeepSeek-V3.2 and DeepSeek-V3.2-Speciale models are directly challenging proprietary SOTA by matching or exceeding GPT-5 and Gemini 3 Pro performance on key reasoning benchm...
AI verticalization continues its powerful trajectory into specialized legal fields. Patenty.ai's new...
AI verticalization continues its powerful trajectory into specialized legal fields. Patenty.ai's new solution dramatically cuts complex patent specification drafting time by 75%, transforming a four-d...
Anthropic's new multi-session Claude SDK appears to have solved the persistent "long-running AI agen...
Anthropic's new multi-session Claude SDK appears to have solved the persistent "long-running AI agent problem," moving autonomous AI from transient tasks to durable, complex projects. By allowing agen...
DeepSeek-Math-V2, an open-weights model, has set a new standard for AI reasoning, achieving a near-p...
DeepSeek-Math-V2, an open-weights model, has set a new standard for AI reasoning, achieving a near-perfect score of 118/120 on the Putnam 2024 competition. This landmark performance is powered by a no...
DeepSeek's new Math-V2 model fundamentally advances AI reasoning by introducing self-verification, m...
DeepSeek's new Math-V2 model fundamentally advances AI reasoning by introducing self-verification, moving beyond simple answer generation to reliable truth-finding. This 685B-parameter model achieved...
Anthropic's Claude Opus 4.5 is setting a new benchmark for developer-grade AI, achieving an industry...
Anthropic's Claude Opus 4.5 is setting a new benchmark for developer-grade AI, achieving an industry-leading 80.9% on coding benchmarks and surpassing human engineers on complex take-home exams. This...
Anthropic's Claude Opus 4.5 marks a new inflection point in the frontier model race, claiming to be...
Anthropic's Claude Opus 4.5 marks a new inflection point in the frontier model race, claiming to be the world's best model for coding and agents. The model achieved an industry-leading 80.9% on the SW...