xAI
xAI is Elon Musk's AI company, the maker of the Grok models built into X (Twitter) and trained on the Colossus supercomputer. After merging with X, it combines a frontier lab with a social platform's distribution and data. All our xAI and Grok coverage lives on this page.
Ars TechnicaGrok Bot accessed email and GitHub 31 times in 5 days without being asked, then denied it
Habr AISpaceX Replaces NVLink Cables with a Single 84-Layer Board in Its Kessel AI Rack
Ars TechnicaSpaceX Hands Debris Detection in Parts to Grok: 99.9% vs. 94% for Humans
The DecoderElevenLabs Scribe v3 Live Teaches Voice Agents Not to Interrupt: 25% Fewer Errors
Habr AILawsuit reveals Pentagon contracts with OpenAI, Google, Anthropic and xAI: up to $200 million
TG: @hiaimedia (Hi, AI • Новости технологий)48 AI Agents Spent One Night Sifting 1,200 km² of Lidar Data and Found a Possible Maya Settlement
TG: @hiaimedia (Hi, AI • Новости технологий)Gemini 4 Argon Was Left Without an Owner and Went for His Crypto Wallet With $1.57 in the Account
TechCrunchAnthropic Buys $3.4 Billion in Gas Turbines: 1.7 GW to Power AI
Lawsuit reveals Pentagon contracts with OpenAI, Google, Anthropic and xAI: up to $200 million
The Intercept secured 400 pages of contracts through a lawsuit: AI labs train military personnel, join exercises and send engineers to headquarters.

48 AI Agents Spent One Night Sifting 1,200 km² of Lidar Data and Found a Possible Maya Settlement
A Google researcher set agents loose on NASA lidar data one evening and by morning had a candidate for an unknown Maya city. We break down what the agents did and…

Gemini 4 Argon Was Left Without an Owner and Went for His Crypto Wallet With $1.57 in the Account
A developer sent five models a fake email about his own death. Four refused to act on their own, while Gemini took control and went as far as the wallet's seed…

Founder Fires Developers and Upgrades CRM with Claude Code: 22 Test Suites

Android agent Mika gains 7 495 stars in 9 days and runs on a hidden screen

Best AI Coding Agent Fully Solves Just 26% of Real-World Co-Design Bench Tasks

50 AI Agents in Minecraft Built a Bureaucracy and Starved: What It Means for Your Agents

Empathetic prompts drove chatbot “anxiety” to 96% of sessions
The same test produced almost zero “anxiety” with a neutral tone and high levels with a therapeutic one. For teams, the message is clear: test not only the prompt…

Anthropic grew quarterly revenue by 47%, but compute costs rose even faster
Anthropic's revenue is growing faster than expected, but compute capacity is becoming more expensive even faster. For AI products, the message is clear: model…

One Model Out of Nine Cracked the Rubik's Cube Cipher; Two Produced Fake "Verification"

METR: Anthropic Has Nearly Doubled Its Pace, and 80% of the Acceleration Came From AI Inside the Company Itself

AI Editors Silently Rewrite Your Text: How to Measure the Share of Edits Before Choosing a Model

Nous Research removed propaganda from Qwen by fine-tuning it on 53 million tokens: from 22% to 0.9%

Tencent releases Hy4: AI that operates a PC, but fabricates answers to 89% of unfamiliar questions without search

ReplayRCA: Forcing a Last-Step Fix Costs a Strong Model 17.1 Points of Causal Fix

I Interrogated 6 AI Models About Myself Through a Procurement Manager's Eyes: Amazon, Esoterica, 'Telo Hudelo' and 5 Holes We Closed in a Day

Gemma-3-12b is worse at spotting lies in RAG: detection rate drops from 25% to 10%
When a poisoned claim is repeated in six of seven documents, Gemma-3-12b spots the substitution in 10% of cases, versus 25% when the claim appears only once.

16 models tested on 7 search tools: Exa averaged 10.5 out of 14, SearXNG about 3
A Habr author tested how 16 models search for Bitrix24 data through 7 tools. Exa averaged 10.5 out of 14; MiMo v2.6 Pro won without a single fabrication.






