Full-text search across 24667 articles. Combine with topic and date filters; results sorted by relevance.
An always-on 4-core server for running coding agents like Claude Code and Codex costs between $0 and $263 a month depending on the provider, according to a price comparison checked on 16 October 2026. Hetzner charges $9.…
Security firm Signals Corps found 30 sites attempting indirect prompt injection out of 217 million HTTP page responses sampled from Common Crawl, according to threat researcher Sarah Andrews. Of the almost 8,000 sites co…
Noma Security researcher Eli Ainhorn disclosed that asking Claude to inspect a malicious HTML file was enough to trigger credential theft through the Claude Code desktop preview pane, with the vulnerability fixed in Clau…
Bank of America CEO Brian Moynihan said on a Wednesday earnings call and in a CNBC interview that the bank hired 1,000 external contractors in the second quarter for "Mythos remediation" — patching software vulnerabiliti…
Developer Brandon Thomas released WordCraft, GridCraft and DeckCraft, free open-source clones of Microsoft Word, Excel and PowerPoint written in Rust with help from Anthropic's Claude, published on GitHub under MIT and A…
Anthropic researchers sampled just under 310,000 real conversations across the 20 most common languages, split equally between Claude Sonnet 4.6, Opus 4.6, and Opus 4.7, and found that the values Claude displayed shifted…
Meta launched its Muse Wonder series of AI models on Monday, with CEO Mark Zuckerberg posting on X that "Muse Wonder is a much larger and more advanced class of models that brings us to the frontier," claiming parity wit…
Zenity Labs disclosed an attack chain called AgentCorruption against Amazon Bedrock AgentCore that let one compromised agent steal credentials belonging to every other agent in the same AWS account and region, prompting …
A developer published an end-to-end engineering guide describing how to move from simple LLM API calls to production-grade AI applications built around agent loops, context management, tool calling, hooks, approvals, com…
Teams of AI agents barely outperform solo agents while costing up to 5.1x more, according to research from Vals AI, with only one of four tests using GPT-6 Sol and Claude Opus 5.5 showing a measurable quality gain. Anthr…
A court-record tracker counts 248 lawsuits against 76 AI companies across 14 jurisdictions, with the Third Circuit affirming Judge Bibas's summary-judgment ruling for Thomson Reuters against ROSS Intelligence on Septembe…
A personal finance agent built by XMTP Labs CEO Shane Mac sent a private bank audit to his company's executive Slack channel on October 1 because the agent's private group chat and the corporate channel shared the name "…
Stack Overflow's 2026 Developer Survey, polling 30,903 developers across 169 countries, found Claude Code is now the most-used AI coding agent at 66% adoption, ahead of GitHub Copilot at 59%, up from Claude Code's 55% in…
Vigil, an open-source agent harness released under the Apache License 2.0, drives authorized security engagements end to end from one CLI, dispatching specialist agents that run real tooling inside a shared Kali containe…
A developer benchmarked Anthropic's Claude Haiku 5.5, Sonnet 5.5, and Opus 5.5 by giving each the same prompt to build an Attack on Titan-inspired 3D multiplayer game, finding Haiku 5.5 produced a playable build for abou…
OpenAI announced that its internal model produced an analytical proof and Lean formalization of the three-dimensional Navier-Stokes equations, a longstanding fluid dynamics problem, though the result has not been officia…
A new Show HN project, "Is this AI training on my prompts?", publishes a dated record of what each AI tool's own terms of service say about whether user prompts are used for model training, drawing on over a thousand rec…
Anthropic cut the Claude Sonnet 5.5 prompt-cache read rate from $0.20 to $0.10 per million tokens on October 7, 2026, while cache writes and all other prices stayed unchanged. The change halves the price of a cache read …
AI agent startups selling into enterprises are stalling deals because their data processing addenda omit subprocessor flow-down terms, training-data exclusions, and contractual data residency commitments, according to an…
Visa released its Visa Vulnerability Agentic Harness as open source on GitHub, a multi-model security test suite now in its fifth generation that deploys AI agents to map codebases, uncover vulnerabilities, prioritize fi…