AI News
Simple Prompting Hack for OpenAI’s New Open-Weight Models
TechnologyAdvice's Grant Harvey shares a "dead-simple prompting hack" for OpenAI's gpt-oss models, as well as several related developer guides.
Black Hat 2025: Security Researcher Unpacks Cybercrime’s Evolution… and How AI Is Changing the Game
From prank viruses to profit-driven cybercrime, Mikko Hypponen explains how today’s malware is targeted, professional, and all about money.
NVIDIA Takes Firm Stance Against Kill Switches & Backdoors in AI Chips
The statement comes as both US and Chinese authorities probe the lucrative global AI chip business, which NVIDIA dominates.
Claude Opus 4.1: Anthropic Delivers Better Coding, Debugging, Analytics Abilities
Claude Opus 4.1 scores 74.5% on the SWE-bench Verified benchmark, indicating major improvements in real-world programming, bug detection, and agent-like problem solving.
OpenAI’s New Open Models Are Available on Microsoft’s Azure AI Foundry
Developers can now build, test, and deploy applications powered by OpenAI’s gpt-oss models within the AI development platform.
Google to Reduce AI Data Center Power Use During Peak Demand
In new agreements with two US utilities, Google has committed to rescheduling or pausing non-urgent AI workloads when the power grid is strained.
Helping data storage keep up with the AI revolution
Storage systems from Cloudian, co-founded by an MIT alumnus, are helping businesses feed data-hungry AI models and agents at scale.
AI Beats Hackers to a Zero-Day Cybersecurity Discovery, Twice
AI prevented real-world cyberattacks before they began. Can AI continue to beat human threat actors to zero-day vulnerabilities?
ChatGPT’s On Track For 700M Weekly Users Milestone: OpenAI Goes Mainstream
With nearly one in 10 people on Earth using ChatGPT, OpenAI is at a turning point.
ISC2 AI Certificate Offers Upskilling for Cybersecurity Professionals
The total cost of the AI security certificate is approximately $640 and counts as 16 continuing education credits.
Google DeepMind’s Genie 3 Simulates 3D Worlds for AI Training
Google says its latest AI model called Genie 3 takes us closer to artificial general intelligence.
OpenAI Launches GPT-OSS, Its First ‘Open’ AI Models Since 2019
OpenAI’s new open-weight models are gpt-oss-120b and gpt-oss-20b. The smaller model, gpt-oss-20b, can be run on a consumer laptop or powerful phone.
Cloudflare Accuses AI Startup of ‘Stealth Crawling Behavior’ Across Millions of Sites
Cloudflare is accusing Perplexity of using stealth crawlers to bypass site restrictions, triggering fresh concerns over how AI firms access web content.
Claude API Controversy: Anthropic Pulls the Plug on OpenAI
An OpenAI spokesperson said the API access allowed for industry-standard benchmarking and safety improvements.
MIT tool visualizes and edits “physically impossible” objects
By visualizing Escher-like optical illusions in 2.5 dimensions, the “Meschers” tool could help scientists understand physics-defying shapes and spark new designs.
Elon Musk Weighs In On Consulting’s ‘Existential Transformation’
McKinsey deploys 12,000 AI agents as consulting faces an “existential” shift. Elon Musk says consultants aren't obsolete yet.
Elon Musk Weighs In On Consulting Industry’s ‘Existential Transformation’
McKinsey deploys 12,000 AI agents as consulting faces an “existential” shift. Elon Musk says consultants aren't obsolete yet.
Cisco Talos Researcher Reveals Method That Causes LLMs to Reveal Training Data
In this TechRepublic interview, researcher Amy Chang details the decomposition method and shares how organizations can protect themselves from LLM data extraction.
Cisco Talos Researcher Reveals Method That Causes LLMs to Expose Training Data
In this TechRepublic interview, researcher Amy Chang details the decomposition method and shares how organizations can protect themselves from LLM data extraction.
NVIDIA Triton Vulnerabilities Could Let Attackers Hijack AI Inference Servers
Three NVIDIA vulnerabilities allow unauthorised users to obtain the IPC memory key and use it to craft malicious inference requests.