Technology News
545 articles
AI technology breakthroughs, research papers, patents, and technical innovations from leading AI startups.
DeepSeek's DSpark architecture delivers 85% single-user speedup, 4x throughput boost in high-concurrency inference
DeepSeek has published a paper, co-authored by founder Liang Wenfeng (梁文锋), detailing DSpark — a speculative decoding architecture that combines a parallel backbone (DFlash) with a lightweight sequent...
Prompt injection attacks exploit design flaws in enterprise AI agents and RAG pipelines
Security researchers are documenting active exploitation of prompt injection vulnerabilities in enterprise AI systems, specifically targeting large language model agents, retrieval-augmented generatio...
China's Zhipu AI (智谱) releases open-weight GLM-5.2, claiming cybersecurity bug-finding capabilities...
China's Zhipu AI (智谱) releases open-weight GLM-5.2, claiming cybersecurity bug-finding capabilities on par with Anthropic's Mythos, while still lagging in general tasks. Why it matters: This release s...
UnconventionalAI unveils oscillator-based architecture claiming 1,000x AI power reduction.
UnconventionalAI, founded by former Databricks AI head Naveen Rao, has introduced its first AI model, Un-0, based on a novel computing architecture that the company claims can reduce AI inference powe...
Microsoft introduces Agentic Resource Discovery specification for AI agents, MCP servers, and API workflows.
Microsoft, in collaboration with Cisco, Databricks, GitHub, GoDaddy, Google, Hugging Face, Nvidia, Salesforce, ServiceNow, and Snowflake, has released the Agentic Resource Discovery (ARD) specificatio...
Shopify AI distillation pipeline slashes production costs up to 30x, smaller models outperform larger ones.
Shopify has developed an internal distillation pipeline that reduces AI production costs by up to 30x, with smaller distilled models occasionally outperforming their larger teacher models in specific...
OpenAI previews GPT-5.6 series with three-tier 'Sol/Terra/Luna' lineup, delays full release at US government request
On June 26, OpenAI began previewing its next-generation AI model series, GPT-5.6, structured in three tiers: the flagship 'Sol', the balanced 'Terra', and the low-cost 'Luna'. The preview is initially...
**Om AI launches VLX, world’s first end-side streaming multimodal model series for physical world.**
At CVPR 2026, Hangzhou-based Om AI unveiled VLX, a family of three streaming multimodal models (VLX-Flow, VLX-Seek, VLX-Go) designed to run entirely on edge devices — phones, drones, and robots — with...
Anthropic export ban spurs Asian AI labs to launch rival frontier models
Singapore-based Vertex AI and South Korean unicorn Mindforge have launched frontier models — Phoenix-7 and Atlas, respectively — claiming Mythos-level reasoning capabilities, as Anthropic's flagship m...
Trump Administration Permits Anthropic to Release Mythos to Select US Organizations
The Trump administration has eased restrictions on Anthropic's most advanced AI model, Claude Mythos 5, allowing the company to grant access to over 100 US organizations, including large corporations...
Anthropic Mythos 5 Gets White House Clearance for 100+ US Companies and Agencies
The Trump administration has authorized the redeployment of Anthropic’s powerful cybersecurity model, Claude Mythos 5, to over 100 specific U.S. government agencies and companies, including non-Americ...
OpenAI debuts GPT-5.6 Sol, Terra, Luna models in US government preview
OpenAI has introduced the GPT-5.6 family of frontier models, comprising three variants: Sol, Terra, and Luna. The release is initially limited to a set of preview partners as requested by the US gover...
OpenAI delays GPT-5.6 rollout at White House request, following Anthropic model takedown
OpenAI confirmed on Friday that it is delaying the public release of its next-generation GPT-5.6 models at the request of the Trump administration, two weeks after Anthropic was forced to take its mos...
Liquid AI Releases LFM2.5-230M, Outperforms 4x-Larger Models on Data Extraction
Liquid AI has released LFM2.5-230M, its smallest model to date at 230 million parameters. The model outperforms models four times its size on data extraction tasks and is designed specifically for edg...
OpenAI and Broadcom debut custom AI inference chip 'Jalapeno' in nine-month design cycle
On June 24, 2026, OpenAI and Broadcom announced the first custom-designed AI inference chip, named 'Jalapeno,' optimized for large language model inference. The chip, described as the first step in a...
Trump administration asks OpenAI to stagger new model release over security concerns
The Trump administration has requested that OpenAI delay or stage the release of a new cutting-edge AI model due to security concerns. This marks the first known instance of the current administration...
OpenAI and Broadcom have jointly unveiled a custom chip designed specifically for AI inference, mark...
OpenAI and Broadcom have jointly unveiled a custom chip designed specifically for AI inference, marking a significant milestone in OpenAI's strategy to diversify its hardware supply chain and reduce r...
Alibaba Cloud releases HappyHorse 1.1 AI video model, climbing to No. 2 in global rankings with full enterprise API access.
Alibaba Cloud has launched HappyHorse 1.1, an enterprise AI video model that has risen to the No. 2 position in global rankings. The model is available via full API access, positioning it for broad en...
TrendForce: AI Agent Growth Drives Enterprise SSD Supply Crunch, Prices Spike 80%.
Research firm TrendForce reported on June 11, 2026 that Q1 2026 enterprise SSD revenue hit a record $18.46 billion, up 86.1% QoQ, driven by surging cloud provider demand from AI agent deployments. The...
xAI launches Grok 4.20 Beta with focus on factual reliability
xAI has released Grok 4.20 Beta, a new version of its flagship foundation model, emphasizing an ultra-low hallucination rate and prioritizing factual reliability. The launch was reported by Chinese te...