AI

Machine learning, LLMs & the AI revolution

Solos Cuts Smart Glasses Weight in Half With Camera-Free AirGo A6
AI

Solos Cuts Smart Glasses Weight in Half With Camera-Free AirGo A6

Solos has unveiled the AirGo A6, a redesigned pair of smart glasses that weighs just 19 grams — roughly half the weight of its predecessor. The glasses drop the camera entirely in favor of a voice-driven AI assistant and a slimmer, more wearable form factor.

The Verge AI21h ago
Anthropic Brings Its AI Collaboration Tool to Phones and Browsers, Starting With Max Tier
AI

Anthropic Brings Its AI Collaboration Tool to Phones and Browsers, Starting With Max Tier

Anthropic is expanding Claude Cowork beyond its desktop-only roots, making the AI platform accessible on iOS, Android, and web browsers for the first time. Max plan subscribers get access first, with broader rollout to other users expected over the coming weeks.

The Verge AI21h ago
Microsoft Quietly Swaps OpenAI and Anthropic for Homegrown MAI Models in Word and Excel
AI

Microsoft Quietly Swaps OpenAI and Anthropic for Homegrown MAI Models in Word and Excel

Microsoft has begun routing a portion of Office 365 user prompts through its in-house MAI models rather than relying exclusively on OpenAI and Anthropic. The shift reflects a growing industry-wide push to rein in skyrocketing AI infrastructure costs.

TechCrunch AI21h ago
Frontier AI Labs Hold Premium Pricing Ground Even as Open Source Models Flood Enterprise Pipelines
AI

Frontier AI Labs Hold Premium Pricing Ground Even as Open Source Models Flood Enterprise Pipelines

Despite surging adoption of open source models like DeepSeek, frontier providers such as Anthropic continue to dominate overall AI spending on major platforms. A new theory from Decagon CEO Jesse Zhang suggests frontier and open source models are actually complementary phases of the same deployment lifecycle rather than direct rivals.

TechCrunch AI21h ago
Meta Superintelligence Labs Debuts Muse Image, Letting Users Drop Friends Into AI-Generated Scenes
AI

Meta Superintelligence Labs Debuts Muse Image, Letting Users Drop Friends Into AI-Generated Scenes

Meta has unveiled Muse Image, the first AI model from its Superintelligence Labs division, now live across Instagram, WhatsApp, and the Meta AI app. The tool introduces an 'agentic' generation pipeline and a controversial feature that can incorporate other Instagram users into AI photos.

The Verge Tech21h ago
Meta's Muse AI Can Remix Strangers' Public Photos — And Won't Tell Them About It
AI

Meta's Muse AI Can Remix Strangers' Public Photos — And Won't Tell Them About It

Meta has launched Muse Image, a free AI image generator that controversially lets users manipulate public Instagram profiles' photos without notifying the people pictured. The opt-out-by-default policy has ignited fresh privacy concerns given the company's troubled history with user data.

TechCrunch AI22h ago
Nearly 5,000 Jobs Gone as Microsoft Restructures Around AI at Fiscal Year Start
AI

Nearly 5,000 Jobs Gone as Microsoft Restructures Around AI at Fiscal Year Start

Microsoft has begun cutting approximately 4,800 employees — around 2.1 percent of its total workforce — as the company enters a new financial year. The reductions hit Xbox and commercial sales hardest, with leadership citing AI-driven transformation as the driving force behind the reorganization.

The Verge AI1d ago
Wearing Smart Glasses Feels Like Spying — And TV Is Finally Showing Why That Matters
AI

Wearing Smart Glasses Feels Like Spying — And TV Is Finally Showing Why That Matters

A Netflix comedy about an elderly undercover operative accidentally captures the most honest critique of smart glasses culture yet seen on screen. The show highlights the deeply uncomfortable reality of walking around recording strangers without their knowledge or consent.

The Verge AI1d ago
JadePuffer's 'Autonomous' AI Ransomware Attack Had a Human Pulling Strings All Along
AI

JadePuffer's 'Autonomous' AI Ransomware Attack Had a Human Pulling Strings All Along

Security researchers at Sysdig made waves claiming their documented JadePuffer attack was the first fully autonomous AI-driven ransomware operation, but key details reveal a human still set up targets, provisioned infrastructure, and supplied stolen credentials. The clarification raises important questions about what truly 'agentic' cybercrime looks like — and how far away the fully automated version really is.

TechCrunch AI1d ago
SK Hynix Eyes $28B US Listing as AI Memory Shortage Fuels Historic Chip Demand
AI

SK Hynix Eyes $28B US Listing as AI Memory Shortage Fuels Historic Chip Demand

South Korean memory giant SK Hynix is preparing to launch American depositary receipts on US markets, potentially raising $28 billion amid surging AI-driven demand. The move gives American investors direct exposure to one of the world's top producers of high-bandwidth memory powering the AI infrastructure boom.

TechCrunch AI1d ago
Vercel Wants to Be AI's AWS — and That Means Taking On the Labs Directly
AI

Vercel Wants to Be AI's AWS — and That Means Taking On the Labs Directly

Vercel CEO Guillermo Rauch says the critical battle in AI infrastructure is keeping models and agents separate, giving developers the freedom to mix and match providers. With 6 million daily deployments and 1 trillion tokens flowing through its gateway, Vercel is positioning itself as the neutral backbone of the agentic era.

TechCrunch AI2d ago
iOS 27 Beta 3 Unlocks Siri Speech Tuning — But ChatGPT Still Has the Edge
AI

iOS 27 Beta 3 Unlocks Siri Speech Tuning — But ChatGPT Still Has the Edge

Apple's third iOS 27 developer beta activates long-awaited voice controls that let users adjust how fast and emotively Siri speaks. The feature, first teased at WWDC 26, marks a meaningful step in Apple's push to make its AI assistant feel more human.

TechCrunch AI2d ago