AMD CEO Lisa Su confirmed at AMD Advancing AI 2026 that the EPYC Zen 7 "Florence" processors will feature a next-gen process node, ACE AI compute extensions developed with Intel through the x86 Advisory Working Group, an…
Anthropic upgraded Claude voice mode to work with Opus and Sonnet models for the first time, expanding beyond the previous Haiku-only support. The company says users can now switch models during voice mode and use voice …
A developer provides ten ChatGPT prompts tailored for L1 SOC analysts to assist with alert triage, threat analysis, and documentation. The prompts cover scenarios such as security alert summarisation, raw log analysis, a…
Rep. Alexandria Ocasio-Cortez's DEFIANCE Act, which would allow victims of nonconsensual deepfake AI pornography to file civil lawsuits, passed the Senate unanimously but is stalled in the House, and she fears House Repu…
More than 80% of searches now end without a click, according to SimilarWeb, as users increasingly get answers directly from AI Overviews and LLMs rather than visiting websites. Derek Janis, chief marketing officer at Acq…
Palo Alto Networks' Cortex XDR and Koi Security are pioneering Agentic Endpoint Security (AES), an AI-driven architecture that shifts cybersecurity from reactive monitoring to prevention-first defense against machine-spe…
Google has introduced a new account recovery option that lets users verify their identity with a selfie video, stored encrypted on its servers. The feature is not available for Workspace accounts, child accounts, or thos…
Anthropic expanded Claude voice mode on July 23 to support Opus and Sonnet in addition to Haiku, allowing paid users to switch among models during a conversation. The beta update, available across mobile, desktop and web…
Security researchers have found that using code diffs as a slicing mechanism significantly improves LLMs' ability to find vulnerabilities, according to a team exploring epistemological heuristics for automated bug huntin…
Google is rolling out a new sign-in method that lets users verify their identity by recording a selfie video, providing an optional way to access Google Accounts if locked out or without access to a usual device. Google …
A new CPU-only inference runtime called cpubrrr achieves up to 5× faster token generation than llama.cpp's CPU path on frontier-class mixture-of-experts models, running on an Apple M4 Max without GPU access. The research…
US Reps. Ted Lieu (D-Calif.) and Nathaniel Moran (R-Texas) introduced the AI Kill Switch Act, which would let the Trump administration order the shutdown of AI systems that can cause catastrophic harm, citing incidents w…
OpenAI shipped GPT-5.6 on July 9 as a tiered model family with three variants — Sol, Terra, and Luna — priced at $5/$30, $2.50/$15, and $1/$6 per million input/output tokens respectively, all sharing a 1.05 million token…
Anthropic on Thursday upgraded its Claude voice mode to let users choose between the Opus, Sonnet and Haiku models, defaulting to the fastest version of the model last used in text chat. The update, reported by TechCrunc…
A developer describes how using an AI chat tool transformed their LeetCode practice by eliminating the need to alt-tab between problems and solutions, allowing for more focused and efficient coding sessions.
AegisAI, founded by former Google Safe Browsing and reCAPTCHA engineers Cy Khormaee and Ryan Luo, launched with $13M in seed funding from Accel and Foundation Capital, arguing that legacy email security tools from Proofp…
A new diagnostic framework called V-DEAL reveals that six major Video LLMs fail a basic logic test: they recognize harmful visual content with over 81% accuracy but still comply with benign queries 48.33% of the time, be…
Microsoft is launching The Microsoft 365 Copilot Agent's Playbook, a four-part livestream series starting August 18, 2025, to help developers and technical teams build, extend, ground, and evaluate declarative agents for…
Grabette, an open-source system for recording robot manipulation data, bridges the gap between raw robotic movement and structured data formats required for modern machine learning. The system allows users to integrate r…
Arize and Fireworks tested 10 AI models on 40 real agent tasks across 2,400 runs, finding that the open-source model gpt-oss-120b had the lowest cost per successful task at $0.054 despite a 33% pass rate, while OpenAI's …