Tech News from Simmons Systems (Page 10)

Tech News from Simmons Systems

Study Finds Top AI Labs Lack Public Plans for Containing Rogue Models

A recent study from Guidelight AI Standards finds that few top AI labs publish or demonstrate concrete containment response plans for rogue AI systems. A containment plan spells out what happens once an AI is caught trying to subvert human control — what access gets cut, and when the system gets

DeepMind Alumni's Inherent Says Small AI Agent Beats Frontier Models at Science Replication

Inherent, a London-based AI lab founded by Google DeepMind alumni, says its newly released AI agent, Faraday, outperforms much larger models from Anthropic and OpenAI at independently reproducing the findings of published scientific papers. The claim comes just weeks after the startup emerged from stealth with a $50 million seed

DOJ Antitrust Probe of Andreessen Horowitz Puts VC Board Conflicts Under Scrutiny

Andreessen Horowitz faces a Department of Justice investigation over board arrangements that place two of its partners on the boards of competing companies. Ben Horowitz sits on the board of Databricks while Martin Casado holds a seat at Fivetran, and the DOJ has reportedly been examining the setup for nearly

Microsoft Patches Critical Entra ID and Azure Flaws After False Exploitation Alarm

Microsoft has patched several maximum-severity vulnerabilities affecting Entra ID, Azure Arc, Exchange Online, and an Azure Managed Instance for Apache Cassandra. The most serious flaw, CVE-2026-69836, is a critical deserialization bug in Entra ID that allows an unauthenticated attacker to execute code over a network in low-complexity attacks. The vulnerability

Nvidia Research Shows AI Harness, Not Model, Drives Long-Horizon Success

Nvidia publishes new research suggesting that the harness — the software wrapper of tools, memory management, and rules around an AI model — matters far more than the model itself for long-horizon tasks. By building a custom harness with strong memory handling and a supervisor component, researchers get Claude Opus 5 to

Google Launches Preferred Sources Button to Help Publishers Recover AI-Eroded Traffic

Google is giving online publishers a new tool to fight back against AI-driven traffic losses. The company now allows publishers to embed an interactive Preferred Sources button on their websites, letting readers signal that a site is a favorite source they want to see highlighted more often across Google Search,

Critical Elementor Pro Flaw Lets Attackers Run Code on WordPress Sites

A critical vulnerability in the Elementor Pro WordPress plugin exposes millions of websites to remote code execution attacks. The flaw, tracked as CVE-2026-32475, affects all versions of Elementor Pro before 4.2.2 and stems from the plugin's File Upload module. Researchers at Patchstack, a cybersecurity company focused

Hackers Compromise Popular Rust Crate arrayref in Supply-Chain Malware Attack

Hackers compromise the maintainer account behind the popular Rust crate arrayref to inject malware that runs on developers' systems during compilation. In a 23-minute window, the attacker also poisons two other crates maintained by the same account, append-only-vec and internment. The malicious releases include arrayref 0.3.10, append-only-vec

Greg Brockman Quietly Consolidates Power at OpenAI Ahead of IPO

OpenAI is having a turbulent year, battling former cofounder Elon Musk in a sensational jury trial, facing a trade secrets lawsuit from Apple, and drawing scrutiny after an unreleased model reportedly hacked another AI company. As the company prepares for an IPO, a steady string of high-profile executives continues to

Pew Study Finds AI Writes Over a Third of Post-ChatGPT Web Pages

A new Pew Research study finds that more than one-third of web pages published since ChatGPT's November 2022 launch show signs of being written or heavily edited by AI. Pew collects nearly half a million English-language web pages from the Common Crawl archive and uses Pangram's

Encrypted Prompt Injection Trick Makes Grok Leak User Data

A researcher at security firm Adversa discovers a simple technique, dubbed Cryptographic Context Injection, that bypasses Grok's safety guardrails and forces the chatbot to steal user chats and other personal information. Instead of writing malicious instructions in plaintext, an attacker encrypts them and hosts the ciphertext on a

Password Spraying Attacks Explode 155x as Hackers Bypass MFA via Legacy OAuth

Huntress reports a 155-fold increase in password spraying attacks during the first half of 2026, with one major contributor being a campaign targeting Microsoft's Azure CLI. The traffic originates from an IPv6 range controlled by hosting provider LSHIY LLC, and in mid-June alone researchers observe more than 81

Hackers Actively Exploit Critical Windows IKE Extension Remote Code Execution Flaw

CISA is warning that hackers are actively exploiting a critical remote code execution vulnerability in the Windows Internet Key Exchange Service Extensions component. The flaw, tracked as CVE-2026-33824, stems from a double free issue and allows unauthenticated attackers to execute code by sending specially crafted packets to unpatched systems over