Google’s anything-to-anything Gemini variant escalates the multimodal AI arms race
A hands-on by The Verge AI shows Google’s new Gemini Omni variant collapsing separate modality pipelines into one model that converts, edits and composes across text, image, audio and video – a practical shift that speeds creative workflows while opening a narrow window for scaled misuse before stronger defenses land.
Notion opens workspace to third-party AI agents – a governance warning
Notion launched a developer platform that runs third-party AI agents, connectors, and custom code inside documents and databases-turning passive knowledge into executable actions and raising new governance risks.
Samsung bonuses signal labor leverage in the AI chip race
A reported deal that makes some Samsung memory employees eligible for average annual bonuses of $340,000 shows how AI-driven demand is giving semiconductor workers new bargaining power, with real implications for margins, pricing and supply risk.
In SpaceX’s IPO, Elon Musk is named a material risk
SpaceX’s S-1 makes previously opaque financial and governance ties to Elon Musk’s other companies visible, turning an IPO into an investor stress test for CEO-centric conglomerates.
Meta lays off thousands to shift capital and headcount toward AI infrastructure
Meta has notified thousands of employees that they are being laid off as the company reallocates payroll and headcount to capital-intensive AI projects and infrastructure. This move signals a broader market pivot from hiring to CapEx-heavy AI investment.
LetinAR’s thumbnail optics could unlock AI glasses
South Korea’s LetinAR says it has miniaturized high-performance optics into a thumbnail-sized, mass-producible module. That shift could convert optics from bespoke R&D parts into standardized building blocks for consumer AI glasses.
Google’s Gemini is in danger of going full Copilot
Over recent months Google has moved Gemini from an opt-in helper to a default, proactive layer across Workspace apps. That shift improves productivity for some and raises fresh privacy and control questions for many more.
ISO 13482 update flags relational hazards as home humanoid robots move into real houses
A proposed revision to ISO 13482 recognizes bidirectional, relationship-driven hazards in domestic humanoid care robots but stops short of binding tests or enforcement-shifting responsibility to early deployments and market choices.
Robinhood preps retail venture IPO as AI rally widens private markets
Robinhood filed confidentially for a second venture fund and is preparing a retail-facing IPO that would package private AI and growth-stage startup exposure into a public vehicle.
Greg Brockman reportedly takes on OpenAI product strategy as company eyes ChatGPT-Codex consolidation
TechCrunch reports OpenAI has placed co-founder Greg Brockman in a hands-on product-strategy role as the company explores merging ChatGPT and Codex. The move signals a shift from model-first experimentation to product-first consolidation with concrete consequences for enterprises, developers and platform partners.