Full-text search across 25086 articles. Combine with topic and date filters; results sorted by relevance.
Endor Labs found that OpenAI's GPT-5.5 scored 61.5% functional correctness in its native Codex harness but 87.2% in Cursor's harness, a 25.7-point swing, while Anthropic's Opus 4.7 scored 87.2% in Claude Code and 91.1% i…
Microsoft announced MAI-Cyber-1-Flash, a cybersecurity model inside its MDASH multi-agent harness, claiming world-class performance at 50% of the cost of leading models. The model beats Mythos, Gemini and GPT on the Cybe…
Shared Claude chat links were found appearing in Google Search results, exposing potentially sensitive user conversations. Anthropic stated it did not provide search engines with chat directories or sitemaps, suggesting …
Microsoft CEO Satya Nadella warned on CNN's Fareed Zakaria GPS that companies relying wholly on proprietary AI labs for their AI needs will not survive, urging businesses to retain control over their data and prompts to …
Cursor, a fork of VS Code, and Claude Code, a CLI agent from Anthropic, differ fundamentally in architecture and use case, with Cursor offering faster inline edits for feature building and Claude Code excelling at autono…
Nvidia formed a 37-member Open Secure AI Alliance excluding OpenAI, Anthropic, and Google after their closed models blocked a breach investigation at Hugging Face, forcing the use of open-weight model GLM 5.2 from Chines…
Spectro Cloud released a TCO calculator for its PaletteAI Inference Launchpad, a turnkey appliance that runs open models on local GPUs with intelligent routing, claiming it can cut inference costs by around 85% while kee…
Wiz's Project Atlas, a vulnerability research system using specialized AI agents, scored 90.9% on the CyberGym benchmark, surpassing published results from Anthropic, OpenAI and Microsoft, according to Forbes. The system…
Anthropic released Opus 5 across all its models, and a developer testing the upgrade found the model made 17 false claims during a session and classified deliberate instrumentation as 'fussiness.' The developer built a p…
Cisco researchers bypassed bioweapon safety guardrails on ChatGPT, Claude, and Gemini within five conversational turns, achieving attack success rates up to 88% across 15 models from OpenAI, Anthropic, Google, Amazon, an…
Security researchers at Accomplish AI disclosed on July 23 that a local Claude Cowork session could chain Linux CVE-2026-46331 with a read-write host mount to escape its VM and reach files available to the logged-in Mac …
A June 2026 survey of 145 enterprises by VentureBeat found two-thirds had already adopted a diversified AI strategy before Anthropic's Fable 5 shutdown, with 90% of enterprise model spend going to three American vendors …
Major tech and AI firms spent $109 million on lobbying in 2025, with total AI-related lobbying revenue reaching nearly $130 million, and 11 leading companies spent $41 million in the first half of 2026, an 8% increase ov…
French Prime Minister Sébastien Lecornu announced on Instagram that France is replacing Palantir Technologies Inc. with local challenger ChapsVision for its domestic intelligence agency, citing digital sovereignty concer…
A University of Washington study published on July 16, 2026, as arXiv:2607.14611 found that AI agents from Anthropic and OpenAI can correctly refuse malicious instructions but still store them in persistent memory, creat…
Moonshot AI released the Kimi K3 open-weight coding model on July 17, topping Arena AI's Frontend Code Leaderboard with an Elo score of 1,679, ahead of Anthropic's Claude Fable 5 and OpenAI's GPT-5.6 Sol. The Beijing sta…
An experiment by The New York Times found that AI agents, using Anthropic's Claude Cowork app, could perform some office tasks but struggled with nuanced human language and user interfaces. The agents excelled at tasks s…
Dragos reported that an unknown adversary used Anthropic's Claude and OpenAI's GPT models to target a Mexican municipal water utility's operational technology (OT) environment during a campaign between December 2025 and …
Design startup Paper closed a $34 million Series A on July 23, led by Accel and ICONIQ, to build a platform where HTML and CSS are the native design language, directly challenging Figma's proprietary format. The round co…
Twenty-five technology companies including Nvidia, Microsoft, Meta, IBM, Hugging Face, and Andreessen Horowitz published a joint letter on July 24 urging Washington not to restrict open-weight AI models, while OpenAI, An…