AI News
Page 2 of 5 · 149 articles
-
Nvidia optimizes data center design for thermal efficiency at scale
Nvidia revealed new data center architecture optimizations that reduce cooling overhead while maintaining performance density, addressing power consumption concerns.
-
OpenAI launches bug bounty program to strengthen open source security
OpenAI introduced a structured bug bounty initiative to identify and address vulnerabilities in open source AI projects, strengthening ecosystem security.
-
2026 tech layoffs accelerate as AI economics reshape hiring
As AI capabilities mature, major tech firms are restructuring workforces, citing automation and AI-driven efficiency gains as drivers of layoff decisions.
-
Fine-tuning small LLMs with Qwen delivers category-specific AI
Researchers demonstrate effective fine-tuning of Qwen's smaller models for specialized tasks, achieving strong performance with minimal computational overhead.
-
Sakana releases Fugu: optimized language models for edge deployment
Sakana AI introduces Fugu, a family of optimized language models designed for efficient deployment on edge devices with minimal computational overhead.
-
Apple brings practical AI to iOS devices beyond Siri
Apple's upcoming iOS updates will introduce AI-powered features across devices, moving beyond traditional Siri voice assistance to integrate conversational AI more deeply into core apps.
-
Trump administration scrutinizes Anthropic's AI compliance
The Trump administration's regulatory focus shifts to Anthropic, questioning AI safety practices and governance standards in enterprise LLM deployment.
-
Can you really tell human from AI text? Deep statistical dive
A technical analysis challenges the common claim that LLM-generated text is indistinguishable from human writing, diving into statistical foundations and practical detection approaches.
-
The Atlantic reveals music tracks used in AI training datasets
The Atlantic released a searchable database revealing the vast scale of unlicensed music used to train AI models, exposing gaps between what artists licensed and what AI companies actually used.
-
New AI search engine lets you find yourself online in the LLM era
In the Weights lets users find what LLMs know about them, addressing gaps when traditional web search no longer surfaces personal identity data.
-
Signal chief warns: AI chatbots are not your friends
Signal's Meredith Whittaker pushes back on the notion that AI chatbots are consciousness-like companions, warning users to maintain critical distance from AI systems.
-
Sanders proposes $7T plan for public AI ownership
Senator Bernie Sanders unveiled a sweeping $7 trillion economic proposal that would give ordinary Americans ownership stakes in AI companies, challenging Silicon Valley's concentration of AI wealth.
-
AI inference startup Baseten raising $1.5B at record valuation
Baseten, an AI inference platform, is raising $1.5B in Series D funding at record valuation, signaling intense investor appetite for AI infrastructure.
-
Elastic acquires DeductiveAI to boost enterprise AI capabilities
Elastic, the search and analytics platform, has agreed to acquire DeductiveAI for up to $85 million, adding AI-powered reasoning capabilities to its enterprise product lineup.
-
OpenAI's research chief exits after five months
Barret Zoph, VP of Research at OpenAI, departed after just five months, raising concerns about leadership stability and internal alignment disputes at the AI research lab.
-
AI agents now deploy hardware independently—robots installing GPUs
Research shows AI coding agents can autonomously install hardware, performing physical tasks previously thought to require human intervention.
-
WakeCap acquires Frontline to build construction intelligence platform
WakeCap, a MENA-focused startup, acquires Frontline to build an AI-powered construction intelligence platform for the Gulf region.
-
Google lets users disable AI features in Google Docs
Google is giving users the ability to turn off AI-powered writing features in Google Docs and other productivity apps.
-
Midjourney expands beyond art into medical imaging generation
Midjourney's latest capability expands AI image generation into medical imaging, potentially enabling faster diagnostic tool development.
-
Pentagon deploys AI to draft congressional reports, saving staff time
Pentagon uses AI to draft congressionally-mandated reports, reducing writing overhead while maintaining full human review of accuracy.
-
Anthropic pauses Claude token billing due to pricing confusion
Anthropic pauses its token-based pricing for Claude, reverting to simpler billing to improve cost transparency and user clarity.
-
DeepMind AI cuts UK building permit review time by 50%
DeepMind AI system cuts UK building permit review time in half, enabling faster housing approval and addressing supply bottlenecks.
-
Anthropic faces Trump administration conflict over AI export rules
Anthropic publicly challenged the Trump administration's AI export restrictions, arguing they harm U.S. innovation and competitiveness against China.
-
CNTXT AI raises $60M Series A to scale AI infrastructure in MENA
CNTXT AI, an AI infrastructure startup focused on MENA, raised $60M in Series A funding. The raise signals investor confidence in AI adoption across the Gulf region.
-
Nvidia raises $25B in bonds to fund AI chip expansion
Nvidia is raising $25 billion through a corporate bond offering to fund massive expansion of AI chip production. The bond sale signals continued confidence in demand for Nvidia's GPUs.
-
Sundar Pichai faces pushback over Google's AI ethics record
Google CEO Sundar Pichai's Stanford commencement speech on AI safety was met with criticism from audience members who questioned Google's internal AI ethics practices.
-
White House legal fight over Claude Mythos 5
The White House is challenging Anthropic's proprietary claims to the Claude Mythos 5 model, arguing government funding entitles public access. Anthropic is defending its trade secrets in court.
-
Respond.io's $62.5M Series B signals AI agents ready for enterprise
Malaysian AI startup Respond.io raised $62.5M Series B to scale AI-powered customer service agents globally. The funding signals institutional confidence in agentic AI as a commercial category.
-
Deep reinforcement learning tames transformer complexity
New research combines deep reinforcement learning with transformer architectures to solve complex reasoning tasks more efficiently than previous approaches.
-
AI companies rush toward public markets
As AI reaches mainstream adoption, frontier companies are preparing for IPOs. OpenAI, Anthropic, and others are signaling private market exit timelines, reshaping tech's funding landscape.