MetaMask launches AI agent wallet on Grok bot
MetaMask released an AI agent wallet inside a Grok bot, letting users trade crypto autonomously with the wallet handling secure transactions.
MetaMaskClosedApp
Latest news rss
Anthropic PAC spent over $250K in third quarter
Anthropic's political action committee spent more than $250,000 in the third quarter, splitting contributions between the two parties.
Anthropic
Construction robotics startup Origin raises $20M led by Accel
Origin, a construction robotics startup, raised $20 million in funding led by Accel and Zacua Ventures to scale its building-site robots.
Origin
Professor says OpenAI, Anthropic AI lead is 'very perishable'
A professor argued that the current lead held by OpenAI and Anthropic in AI is 'very perishable,' warning investors against assuming durable dominance.
Stripe president says agentic commerce went vertical
Stripe president John Collison said the company's agentic commerce activity went 'vertical' last month, expanding work on AI agents that handle payments and checkout.
Stripe
Cerebras says inference speed is key for agentic workflows
Cerebras argued that agentic workflows make inference speed a requirement, highlighting its speed advantage over GPU-based serving for multi-step agent tasks.
Cerebras
Reuters: market selloff and rising rates test AI data center deal boom
Why it matters: Rising rates threaten the debt-fueled data center buildout underpinning AI compute expansion.
Reuters reports that a broader market selloff and rising interest rates are testing the boom in AI data center deals, raising questions about financing for planned builds.
Bank of America assigns 4,000 people to fix AI-found vulnerabilities
Why it matters: Shows AI-driven vulnerability discovery forcing large-scale remediation in enterprises.
Bank of America brought in 1,000 outside contractors and committed 4,000 people total to patch software flaws surfaced by AI models including Anthropic's Mythos.
Bank of America
GPT-6-Astra solves 49 of 50 physics problems on TheoPhysEval
Why it matters: Shows frontier models approaching expert-level theoretical physics reasoning at low cost.
On the TheoPhysEval benchmark GPT-6-Astra solved 49 of 50 unpublished theoretical physics calculations at about $2.40 each, matching specialists; some reference solutions had errors.
OpenAIClosed
Study finds Korean-language gaps in AI safety tests
A study reports that AI safety evaluations miss or underperform on Korean-language prompts, leaving Korean users less protected.
Report warns federal Anthropic restrictions raise competition concerns
A legal analysis argues federal restrictions on Anthropic could distort AI market competition and disadvantage the company relative to rivals.
Anthropic
Humanoid robot kicks drone across room at conference
A humanoid robot kicked a drone across the room during a demo at a robotics conference, an unscheduled moment captured on video.
Elevatry raises €3M to automate procurement workflows
German startup Elevatry secured €3 million to build AI that automates corporate procurement workflows.
Elevatry
Multiverse Computing launches Comet 60B with Nvidia
Why it matters: Adds a parameter-efficient multimodal competitor tuned for Nvidia hardware.
Multiverse Computing released Comet 60B, a multimodal model built with Nvidia that it says matches larger models at half the parameters.
Multiverse ComputingClosed
More: TradingView
Ubtech Robotics pursues three growth paths at differing maturity
Ubtech Robotics is pursuing three distinct growth paths, each at a different stage of maturity, according to a report on the humanoid and service robotics company.
Ubtech Robotics
Microsoft Maia 300 compared with Nvidia Rubin on specs
TechRepublic compares Microsoft's Maia 300 accelerator with Nvidia's Rubin on specifications, not benchmarks.
Microsoft
Rivian to bring AI agents to its cars
TechCrunch reports AI agents are coming to Rivian vehicles, with more details expected at the company's 'AI & Autonomy Day' event in December.
Rivian
Mindbeam partners with Nvidia on open-source AI speed
Mindbeam announced an Nvidia collaboration aimed at making open-source AI faster and more efficient.
MindbeamOpen
More: Yahoo Finance
Simons Foundation launches $15M effort for mathematicians in AI era
The Simons Foundation launched a $15 million initiative to help mathematicians navigate the AI era, drawing discussion on Hacker News.
Simons Foundation
Germany plans AI partnership with Nvidia on automation and robotics
Why it matters: Signals national-level government alignment with Nvidia for industrial AI and robotics infrastructure.
Germany is reportedly preparing an AI cooperation deal with Nvidia covering automation and robotics, according to Handelsblatt. Scale AI also announced a separate partnership with Nvidia on post-training.
Nvidia
Apple's Joswiak says Apple Intelligence targets mainstream users
In a TIME interview, Apple marketing chief Greg Joswiak said Apple Intelligence is built for 'everyday people' rather than the 'lunatic fringe,' and said Apple does not feel threatened by OpenAI's hardware plans.
AppleClosed
More: 9to5Mac
Apache Ranger creators launch AI agent security guide series
The creators of Apache Ranger published a guide series for CISOs on securing AI agents, covering governance, access control and risk management for enterprise deployments.
Ramp reaches $60B valuation as revenue tops $2B
Why it matters: Signals investor appetite for AI-embedded fintech at near-decacorn scale.
Fintech Ramp hit a $60B valuation as annualized revenue crossed $2B with 85,000 customers, pushing deeper into AI-powered finance tools.
Ramp
Mod brings Nvidia DLSS frame-gen performance mode to RTX 30 GPUs
Club386 reports a community mod enables Nvidia's new DLSS frame generation performance mode on GeForce RTX 30-series GPUs ahead of official support.
NvidiaOpenCodeFree
Chosun: superintelligent AI risk prompts calls for kill switch
Chosun Ilbo reports warnings about dangerous superintelligent AI, with experts urging kill-switch controls as a safeguard against loss of human oversight.
Report: AI companies aim to become frontier scientific researchers
AI labs are increasingly pursuing frontier scientific research directly, with better AI models seen as the payoff from those efforts.
Maximum signs Emprise Bank as design partner
AI-native core banking startup Maximum announced a design partnership with Emprise Bank to overhaul its core banking system.
Maximum
Sipeed MaixStation pairs AX8860P with 128GB, targets DGX Spark
Why it matters: A cheaper Chinese rival to DGX Spark pressures Nvidia's desktop AI workstation pricing.
Sipeed revealed MaixStation, a Chinese-made AI workstation pairing Axera's AX8860P chip with 128GB of LPDDR5X at 273 GB/s, expected to cost $3,999 against Nvidia's DGX Spark and AMD's Ryzen AI Halo.
Sipeed
Scientific American: OpenAI math divides cryptographers on encryption
Why it matters: Touches whether AI progress could eventually threaten deployed cryptography.
Scientific American reports OpenAI's new mathematical results have cryptographers split over whether AI can break encryption, with no demonstrated attack on deployed schemes.
OpenAI
CNBC: Nvidia, Apple, Broadcom compete for TSMC capacity
Why it matters: TSMC capacity rivalry shapes allocation of leading-edge AI chips across major buyers.
CNBC reports Nvidia, Apple and Broadcom are in fierce competition for limited chip manufacturing space at TSMC as AI demand strains capacity.
TSMC
Nvidia joins The Earthshot Prize as founding member
Nvidia became a founding member of The Earthshot Prize, the environmental innovation award founded by Prince William.
Nvidia
Global Times ties China robot boom to eldercare demand
Global Times argues eldercare deployments show genuine commercial momentum behind China's robot industry, pointing to real paying demand rather than hype.
Futura reviews robots breaking human athletic records
A French science outlet surveys robots beating records previously thought exclusive to humans, including running and other physical feats.
AWS, Anthropic, OpenAI, Google race to build forward-deployed engineer teams
Why it matters: Forward-deployed engineers are becoming the main channel for turning enterprise AI pilots into production revenue.
CNBC reports cloud and AI firms are competing to hire large teams of forward-deployed engineers who embed with customers, as labs look to prove AI spending delivers returns.
New Zealand regulators set out cyber and AI risk expectations
The Reserve Bank of New Zealand and the Financial Markets Authority outlined supervisory expectations for regulated firms on cyber risk linked to frontier AI.
Reserve Bank of New Zealand
FIS makes AML Investigations Agent commercially available
Why it matters: Brings bank-grade agentic anti-money-laundering investigations to any US bank.
FIS said its AML Investigations Agent, developed with Anthropic and announced in May 2026, is now commercially available to any US-based bank, including non-FIS clients.
FISClosedPaid
Wix's Symphony multi-agent system connects to Stripe Link
Why it matters: Links a small-business agent platform into a major wallet, a concrete step for agentic checkout.
Wix said Symphony, its standalone multi-agent system for small businesses and solopreneurs, now connects to Link, Stripe's digital wallet, letting those agents interact with Link at checkout.
WixClosedAppPaid
Pineapple Financial rebrands as Pentiv, launches mortgage AI platform
Pineapple Financial became Pentiv and launched an intelligence platform for mortgage agents and brokerages, its next phase after the rebrand.
Pentiv
EU questions OpenAI, Anthropic over AI agent incidents
Why it matters: Signals EU scrutiny of autonomous agent safety as agentic products proliferate.
European Union regulators have questioned OpenAI and Anthropic about security incidents involving their AI agents, according to TechRepublic.
Cohere publishes guide on measuring AI ROI
Cohere outlines methods for enterprises to measure return on investment from AI deployments, covering cost and value metrics.
CohereOfficial
