OpenAI scraps release of new AI model (GPT-6.1 Astra) after internal safety tests
The pull of an unreleased frontier model over misbehavior and safety concerns comes as lab leaders and UK officials debate slowing frontier development.
TechCrunch · +13 outlets
Nvidia adds $150B to buyback, its largest ever, lifting authorization to $235B
The record repurchase is framed partly as a response to competitive pressure on Nvidia's stock amid the AI chip race.
The Next Web · +6 outlets
AMD to acquire Fei-Fei Li's World Labs in $8.2B all-stock deal
World Labs, founded in 2024, focuses on spatial intelligence; Li will join AMD as executive vice president and chief scientist.
TechCrunch · +13 outlets
Nvidia unveils security platform to contain rogue AI agents
The toolkit aims to detect and contain misbehaving autonomous agents, amid a broad debate over whether rogue agents are an engineering problem or a step toward
TechCrunch · +13 outlets
Anthropic releases Claude Sonnet 5.5, a cheaper and faster mid-range model
The release, Anthropic's second model in under a week, comes after CEO Dario Amodei's calls to slow frontier development.
TechCrunch · +13 outlets
Florida AG seeks injunction to halt OpenAI development without independent safety guardrails
The state motion escalates legal pressure on OpenAI over safety as Congress weighs federal AI regulation.
The Next Web · +10 outlets
Apple ordered to pay $5.7B in haptics patent case over iPhone and Apple Watch
A jury found Apple infringed Taction Technology patents on vibration-based tactile feedback; the verdict lifted litigation funder Burford Capital's shares.
CNET · +12 outlets
Meta launches enterprise AI unit, hires MongoDB CEO CJ Desai to lead it
The new Meta Enterprise Platform division, reporting to Zuckerberg, is aimed at justifying Meta's massive AI spending; MongoDB shares fell on the news.
CNBC · +14 outlets
Meta taps Australia's Firmus for AI computing capacity in Southeast Asia
The deal adds data center capacity in the region as Meta expands AI infrastructure spending.
Reuters · +5 outlets
Khanna to introduce AI safety bill barring 'recursive' technology until safeguards exist
The proposal adds to a growing stack of congressional measures on how to regulate frontier AI.
CNBC · +2 outlets
Trump to host OpenAI, Anthropic, Google, Meta executives at White House meeting
The meeting comes as AI safety debate intensifies and executives push for pacing frontier model development.
CNBC · +13 outlets
Twenty-plus researchers warn AI research automation could trigger an 'intelligence explosion'
Signatories include OpenAI's chief scientist and an Anthropic co-founder, who argue the scenario could compress years of progress into a short period.
The Next Web · +4 outlets
Nvidia faces scrutiny over AI chip sales in China and influence in Washington
Experts raise concerns about the company's China business and its sway over Trump administration policy as AI chip competition intensifies.
Ars Technica · +14 outlets
OpenAI's early Hugging Face offer sparked bidding before Nvidia's $13B deal
Hugging Face drew fresh deal interest after an OpenAI investment approach, ahead of Nvidia's agreement to buy the open-source platform.
CNBC · +3 outlets
Elon Musk and SpaceXAI subpoenaed by NYC in AI safety investigation
The New York City Council is requiring testimony from Musk or a SpaceXAI representative.
CNBC · +4 outlets
SpaceX launches Starship into orbit for the first time
The flight delivered 26 Starlink satellites for the largest launch vehicle ever built.
CNET · +13 outlets
Apple releases iOS 27.0.1 to fix iPhone 18 Pro freezing and restart issues
The first iOS 27 update addresses a Face ID related bug that froze or restarted iPhone 18 Pro and Pro Max devices.
ZDNET · +1 outlets
Apple patches CoreGraphics flaw possibly exploited in targeted attacks
iOS 26.7.1, iPadOS 26.7.1, and macOS updates fix an out-of-bounds issue Apple believes was used against a small number of people.
MacRumors · +1 outlets