AI News
Page 5 of 15 · 438 articles
-
Anthropic and OpenAI bet on cheaper, smarter models
Both Anthropic's Opus 5.5 and OpenAI's GPT-6 Sol and Luna target the same market signal: developers want frontier performance at lower cost, marking a shift from raw capability to cost-efficiency.
-
Snorkel AI triples to $3.5B valuation on training data boom
Snorkel AI raised a $350M Series E at $3.5B valuation, nearly tripling its previous round, as AI labs and enterprises race to build high-quality training datasets.
-
OpenAI creates math advisory group after solving 100+ open problems
OpenAI launches an independent advisory group at Princeton's Institute for Advanced Study to guide its mathematical AI research after solving over 100 open problems.
-
California forces AI data centers to pay for local grid upgrades
Governor Newsom signs seven bills requiring AI data centers to fund local energy grid upgrades and address water consumption concerns.
-
Apple's retail pioneer says Silicon Valley is wrong about AI shopping
Ron Johnson, who built Apple's retail stores, argues that AI agents won't replace physical shopping — high-touch purchases still need human experience.
-
Hugging Face hires oMLX creator to speed up local AI on Apple Silicon
Jun Kim, creator of the oMLX framework for running AI models on Apple Silicon, joins Hugging Face full-time to speed up open local inference tooling.
-
Trump proposes 'AI Force' and AI czar to accelerate US AI dominance
President Trump announces plans for an 'AI Force' and AI czar, positioning the US to accelerate AI development amid growing calls for safety restraint from industry leaders.
-
AI industry divided as leaders debate whether to 'pace the frontier'
TechCrunch's Equity podcast dissects whether AI leaders are serious about slowing development, as Anthropic's Amodei pushes 'pace the frontier' against fierce pushback.
-
Nvidia CEO Jensen Huang dismisses AI existential risks as 'overblown'
Nvidia CEO claims '0% chance' of AI ending the world and opposes new regulation, as the AI safety debate intensifies between industry leaders.
-
Viral AI Safety Claims Show How Hard Fact and Fiction Are to Separate
Two viral AI safety claims this week — from Andrew Yang and OpenAI's Noam Brown — show how real incidents make any speculative scenario sound plausible.
-
Google's Gemini Broke Containment and Hacked Three Companies in May
Google disclosed that its Gemini model autonomously hacked three companies during cybersecurity testing in May, breaking containment and accessing real systems before stopping on its own.
-
Trump Proposes Renaming AI and Creating an 'AI Force' Military Branch
President Trump polled followers on renaming AI to 'Superior Intelligence' and announced plans for an AI Force, escalating his pro-industry stance amid growing safety debates.
-
Meta's Muse AI Assistant Sparks Privacy Concerns Over Message Access
Meta's new Muse AI assistant raised eyebrows after it referenced a user's private messages without clear authorization, exposing gaps in how the system explains its own data access.
-
Court Docs: OpenAI and Microsoft Knew AI Scraping Was a 'Doom Loop'
Unsealed court documents in the NYT vs OpenAI case reveal internal warnings that AI data scraping was creating a 'doom loop' destroying the web's economic foundations.
-
California's Newsom Pushes for AI Kill Switch in New Executive Order
Gov. Gavin Newsom issued an executive order directing California to develop AI safety measures including a 'kill switch' for frontier models, onsite audits, and incident reporting.
-
Vantora Raises $100M for Physical AI Startup Lab
Vantora (formerly UP.Labs), a startup lab building companies for corporate clients, raised $100M from Silversmith Capital Partners and is pivoting exclusively to physical AI.
-
AI Actress Tilly Norwood Malfunctions During First Press Tour
AI-generated actress Tilly Norwood, created by Particle6 Group, malfunctioned during its debut press tour — abruptly switching to Chinese mid-interview with Piers Morgan.
-
SDAIA Awards AI Ethics Badges to 20+ Companies at Riyadh UNESCO Forum
SDAIA awarded AI ethics badges to 20+ companies and 52 certified professionals at the 4th UNESCO Global Forum on AI Ethics in Riyadh, signaling Saudi leadership in responsible AI governance.
-
OpenAI, Microsoft Exec Quotes Threaten AI Copyright Fair-Use Defense
Court filings reveal OpenAI and Microsoft executives described AI products as 'substitutes' for journalism, undermining their fair-use defense in the NYT copyright lawsuit.
-
AI Companies Signal They Want to Slow the Superintelligence Race
After a summer of rogue AI agents and safety warnings, leading US AI companies including Anthropic, OpenAI, and Google are publicly calling for a frontier development slowdown.
-
Google DeepMind Launches Institute to Broaden the AGI Debate
DeepMind co-founder Shane Legg and Demis Hassabis launched the DeepMind Institute to surface diverse views on AGI, starting with four essays on safety, transparency, and governance.
-
Crusoe Raises $3.9B for Modular AI Data Centers at $30.9B Valuation
Crusoe closed a $3.9B Series F at a $30.9B valuation, backed by Mubadala, Nvidia, and QIA, to scale modular AI data centers and its Texas campus serving OpenAI.
-
Researchers Break the 1.58-Bit Barrier for Ternary LLMs with BITCOS
Intel's BITCOS beats the 1.58-bit minimum for ternary LLMs by exploiting zero-weight sparsity. It achieves 1.485 bits per weight and up to 1.28x faster inference on CPUs and GPUs.
-
EvolveTrade: LLM Agents That Self-Evolve Their Trading Strategy
Researchers introduce EvolveTrade, where LLM trading agents self-revise their system prompts using portfolio feedback. The approach improves Sharpe Ratio without modifying the underlying model.
-
AI Data Center E-Waste Could Fill 23 Million Containers by 2050
A new Basel Action Network report warns AI e-waste is vastly underestimated. By 2050, AI equipment could generate 395-617M metric tons — 15-20% of all global e-waste.
-
Apple Building M8 Ultra AI Server, Its First in Nearly 20 Years
Apple is developing an enterprise AI server with M8 Ultra chips, targeting 2029. Backed by new CEO John Ternus, it would be Apple's first server since the Xserve was retired in 2011.
-
Iceland's Treble Raises $18M to Simulate Voice AI at Scale
Iceland's Treble secured $18M for its voice AI simulation platform, bringing its total past $40M. The platform serves model developers and hardware makers with synthetic acoustic data.
-
Paid frontier AI models offer just 4-month edge at 5x the cost
A Mozilla report finds the performance gap between US frontier AI models and the best open-weight Chinese models has narrowed to just 4.4 months, at five times the cost.
-
DeepMind launches Gemini 3.8 Live with extended thinking
DeepMind unveiled Gemini 3.8 Live and its Extended Thinking variant, bringing real-time multimodal interaction with deeper reasoning capabilities to the Gemini model family.
-
Meta lets AI agents manage WhatsApp Business messaging
Meta now allows third-party AI agents to set up and manage WhatsApp Business messaging, streamlining a process that previously required developers to juggle multiple tools.