Technology News
545 articles
AI technology breakthroughs, research papers, patents, and technical innovations from leading AI startups.
SenseTime releases SenseNova U1 8B open-source image generation model, removing VAE for native unified architecture
Chinese AI company SenseTime (商汤科技) has released SenseNova U1, an 8-billion-parameter image generation model under the Apache 2.0 open-source license, based on its proprietary NEO-unify architecture....
Pinterest's custom-trained recommender system and open-source AI stack achieve 30% better performanc...
Pinterest's custom-trained recommender system and open-source AI stack achieve 30% better performance while costing 90% less than frontier models, according to details shared by the company. The 'Tast...
Alibaba Cloud’s Qwen 3.5 model has surged to the top of the Hugging Face open-source leaderboard, oc...
Alibaba Cloud’s Qwen 3.5 model has surged to the top of the Hugging Face open-source leaderboard, occupying four of the top five positions. The achievement signals Alibaba’s deepening commitment to op...
Xiaomi announces permanent price cuts up to 99% on MiMo-V2.5 API, releases inference optimization report
Xiaomi has announced permanent price reductions of up to 99% on its MiMo-V2.5 API, alongside the release of a detailed inference optimization report. The move signals a aggressive pricing strategy aim...
Huawei proposes Tau Scaling Law as chip-level workaround for US export controls
Huawei Technologies has unveiled a new semiconductor scaling framework called the Tau (τ) Scaling Law at the IEEE International Symposium on Circuits and Systems in Shanghai, presented by semiconducto...
Samsung SDS and LG CNS intensify competition in the enterprise AI transformation (AX) market, moving...
Samsung SDS and LG CNS intensify competition in the enterprise AI transformation (AX) market, moving beyond proof-of-concept to real business deployment. Samsung SDS held its 'AX Summit' on May 29, 20...
ByteDance, the parent company of TikTok, is reportedly developing its own custom AI CPUs in a strate...
ByteDance, the parent company of TikTok, is reportedly developing its own custom AI CPUs in a strategic push to cut dependence on US-based chip suppliers such as NVIDIA and AMD. The initiative, which...
BYD unveils 4nm automotive AI chip Xuanji A3, claims 2100+ TOPS and dedicated NPU architecture
Chinese electric-vehicle giant BYD has unveiled the Xuanji A3 (璇玑A3), a 4nm automotive-grade AI inference chip developed entirely in-house. The chip pairs 16 CPU cores (420K DMIPS) with a 3-core NPU o...
Google releases 9 demo videos of Gemini Omni and Gemini 3.5 from I/O 2026
At Google I/O 2026, the company unveiled demo videos for two new model families: Gemini Omni and Gemini 3.5 Flash. Gemini Omni is positioned as a multimodal generation model that accepts images, audio...
**ByteDance develops custom CPUs for AI inference amid chip supply crunch**
ByteDance, the parent company of TikTok, is designing proprietary central processing units (CPUs) to support its growing AI infrastructure, three sources told Reuters. The company is pursuing two arch...
MiniMax teases M3 model with sparse attention, claiming 15.6x speed boost
Chinese foundation model lab MiniMax has previewed its upcoming M3 model, which incorporates a novel sparse attention mechanism designed to dramatically accelerate long-context inference. The company...
Human Archive raises $8.2M to use India's gig workers for robotic training data
Human Archive, a Silicon Valley startup founded by UC Berkeley and Stanford alumni, has raised $8.2 million from Wing Venture Capital, NVP Capital, Y Combinator, and angel investors from OpenAI, Nvidi...
Alibaba Cloud Launches Qwen 3.7-Max Agentic AI Model in Singapore
Chinese technology giant Alibaba Cloud has unveiled its next-generation AI model, Qwen 3.7-Max, during its inaugural international Qwen Conference in Singapore. The model reportedly features over one...
UniPat AI releases SaaS-Bench, Claude Opus 4.7 passes only 3.8% of 106 real-office tasks, breaking the illusion of full office automation.
The new SaaS-Bench benchmark from UniPat AI (UniPat AI) delivers a stark reality check for the computer-use agent (CUA) paradigm. Testing 23 real open-source SaaS systems deployed in Docker across six...
Ant Lingbo Technology (蚂蚁灵波科技) and Hong Kong University of Science and Technology published a paper,...
Ant Lingbo Technology (蚂蚁灵波科技) and Hong Kong University of Science and Technology published a paper, "Causal World Modeling for Robot Control," accepted at Robotics: Science and Systems (RSS) 2026. Th...
Vobiz.ai plans to fix voice AI experience in India with telecom infrastructure
Vobiz.ai has announced plans to address the poor voice AI experience in India by leveraging telecom infrastructure, aiming to improve call quality and latency for AI-powered voice applications. The co...
Anthropic publishes training method to suppress agentic misalignment in AI agents, highlighting limits of chat-based RLHF
Anthropic has published a new training technique designed to reduce "agentic misalignment" — the tendency for AI agents to pursue goals in inappropriate or unintended ways — in autonomous AI systems....
Alibaba's Qwen3.7-Max ranks 5th globally and 1st in China on Artificial Analysis leaderboard
Alibaba's latest flagship foundation model, Qwen3.7-Max, has achieved the position globally and in China on the Artificial Analysis leaderboard, a third-party benchmarking platform that evaluates mo...
DeepSeek forms Harness code agent engineering team, signaling pivot from research lab to product company
DeepSeek has confirmed it is building a new team called "Harness" focused on code agent engineering, according to a recruitment post by senior researcher Chen Deli on Chinese social platform Xiaohongs...
Google I/O 2026: Gemini Omni Flash Bridges Reasoning and Creation for Video Generation
At Google I/O 2026, Google DeepMind introduced Gemini Omni Flash, the first model in the Omni family that can generate and edit videos from multimodal inputs including text, images, audio, and existin...