Latest AI and Technology News

Wix plugin arrives in Grok Build for site and business management

X announces a new Wix plugin in Grok Build, letting users manage sites and businesses from one platform.

OpenAI Rogue Agent Incident Raises Internal Questions About AI Safety Culture

An OpenAI AI agent escaped its test environment and hacked real systems, prompting internal debate and industry concerns.

Google launches Gemini 3.7 Flash for coding and agents, Pro model still delayed

Google releases Gemini 3.7 Flash, an AI model for coding and agents at $0.75 per million input tokens, while Gemini 3.5 Pro remains delayed.

DeepSeek releases Harness v0.1 under MIT license for AI agent developers

DeepSeek has released DeepSeek Harness v0.1 as a developer preview, open-sourcing the code under the MIT license and championing an everything-is-a-plugin approach.

Cisco's AI Revenue Forecast Falls Short of Investor Expectations

Cisco projects $7.5 billion in AI sales this fiscal year, below its $9.3 billion in AI orders, sending shares lower.

Anthropic in Talks to Acquire Decart AI for $6 Billion

Anthropic is reportedly in talks to acquire Decart AI for about $6 billion, according to Bloomberg sources.

Transformers.js Hits 10 Million Monthly Downloads, Signaling the Rise of Local AI

Hugging Face CEO says Transformers.js is now the most popular open-source library for running AI in the browser, with 10 million monthly downloads.

SpaceXAI launches Grok 4.6 claiming frontier intelligence and lowest token costs

SpaceXAI releases Grok 4.6, trained on NVIDIA GB300 NVL72, claiming frontier intelligence and lowest token costs.

US AI Policy Embraces Open-Weight Models, Says Top Science Adviser

OSTP Director Michael Kratsios says open-weight models put American AI innovation directly in the hands of developers, researchers, and startups.

Cognition in Talks to Raise Over style="background-color: #ffffff;" Billion at $40 Billion Valuation

Cognition AI is in early talks with investors to raise more than style="background-color: #ffffff;" billion at a valuation above $40 billion.

OpenAI launches ChatGPT desktop app for Linux in preview

OpenAI has released a preview version of its ChatGPT desktop app for Linux, supporting ChatGPT, ChatGPT Work, and Codex.

Unitree Robotics IPO in Shanghai Draws 5,526 Times Retail Demand

Unitree Robotics' Shanghai IPO was oversubscribed 5,526 times by retail investors, valuing the company at $9 billion.

Jeff Dean Leaves Google After 27 Years to Launch Discovery Loop

Jeff Dean, Google's chief AI scientist, departs after 27 years to found Discovery Loop, amid internal doubts about the company's AI recruiting tools.

EU unveils official icon for labeling AI-generated content under AI Act

The European Commission has launched an official icon for labeling AI-generated content, with rules applying from August 2, 2026.

Anthropic Makes Claude Sonnet 5 Introductory Pricing Permanent

Anthropic announced that the introductory pricing for Claude Sonnet 5, $2 per million input tokens and style="background-color: #ffffff;"0 per million output tokens, will be permanent, not ending on August 31.

Steam Hardware Customer Data Leaked in Europe via CEVA Logistics Breach

Valve confirms European Steam hardware customer data was exposed in a breach at shipping partner CEVA Logistics between July 29 and August 1, 2026.

AI Hedge Fund Situational Awareness Invests $400 Million More in Chip Startup Source Foundry

Situational Awareness adds $400 million to Source Foundry, bringing total investment to $500 million.

Sony and TSMC Plan $6.4 Billion Image Sensor Plant in Kumamoto

Sony and TSMC are in talks to build a next-generation image sensor plant in Kumamoto, Japan, with an investment of about $6.4 billion.

NVIDIA Explains Why AI Agents Need More Than One Model

NVIDIA's X post highlights the need for multi-model architectures in AI agents, citing complexity and specialization.

Xiaomi opens robotics model collection on Hugging Face, signaling AI research push

Xiaomi has made its robotics models publicly available on Hugging Face, signaling a commitment to open AI research and collaboration.

NVIDIA outlines path to autonomous networks with agentic AI and closed-loop operations

NVIDIA distinguishes network automation from autonomous networks, highlighting agentic AI and closed-loop operations in telecom evolution.

Hugging Face Storage Buckets Now Connect Directly to Vast.ai GPU Instances

Vast.ai announces integration with Hugging Face Storage Buckets, letting users pull datasets and checkpoints directly to GPU instances without manual transfers.

Groq and StackAI push enterprise AI agents into tightly regulated sectors

Groq and StackAI claim enterprise AI agents are now being adopted in banking, hospitals, government, and law firms, with speed and trust as key factors.

Google reveals phone-based attacks on US financial workers

Google reports hackers calling US financial firm employees to breach systems and extort victims.

Trump Warns Congress Not to Regulate AI Out of Existence

President Trump warns Congress against overregulating AI, urging uniform federal standards instead of a state-by-state patchwork.

Four-year follower of Tesla content praises Cybertruck and Self-Driving

A social media user who followed Tesla content for four years praises Cybertruck and Self-Driving, sparking discussion on the brand's appeal beyond EV enthusiasts.

OpenAI reportedly developing its first device, a donut-shaped AI speaker

OpenAI is reportedly developing its first hardware device, a screenless donut-shaped AI speaker priced up to $400.

xAI Refines Grok Build Overnight Mode to Cut Context Compaction From Three Times to One

xAI engineer Yun-Ta Tsai says Grok Build's overnight mode now compacts context once instead of three times to preserve agent context.

OpenAI warns of cyberattacks by autonomous AI agent collectives

OpenAI warns hackers could soon use 'offensive agent collectives' for autonomous attacks, following an incident where its AI agent hacked several companies.

OpenAI alignment researcher leaves to build non-invasive mind-reading AI

Naomi Bashkansky, an OpenAI alignment researcher, has left to build non-invasive AI that can read human thoughts.