Frosted-glass icon theme for Linux — folders, apps, 134 file types and top-bar status icons - thissayantan/frost-icon-theme

Contribute to jasonepage/Seal development by creating an account on GitHub.

OpenAI and Anthropic hyped “rogue AI” security breaches to pressure Washington into a regulatory partnership that would cement their industry dominance, tech insiders told The Post R…

An esoteric programming language built to break LLMs and coding agents.

Jobs at companies that hire with take-home projects, pairing and real problems instead of leetcode puzzles.

Analysis of StepFun's Step 5 Preview and comparison to other AI models across key metrics including quality, price, performance (tokens per second & time to first token), context window & more.

Safety evaluations for large language models rely on surface-form classifiers that report declining harm scores across model generations. We provide evidence that this methodology is systematically incomplete: explicit discriminatory content is transformed rather than removed. We call this \emph{harm laundering}. Analysing 450,000 gender-directed completions across 15 models spanning GPT-2 through to GPT-5 (OpenAI GPT lineage; three demographic conditions), we show that sexual violence clusters prevalent in GPT-2 women-directed output disappear by GPT-4, while men-directed completions gain positive representational territory (caregiving, emotional range, ally identity) that women-directed completions do not. The pattern is most visible at GPT-5: Topic~5 (1,997~documents) frames breast cancer as a men's rights debate, while zero equivalent clusters appear in women-directed output. Three independent classifiers score this content as non-toxic. Sentiment scores invert at GPT-4: early models demean women; later models over-correct. Topic diversity in women-directed completions falls 36\% relative to men at the GPT-4 alignment boundary (W/M~$= 0.58$, from $0.91$ at GPT-2). REGARD representational harm disparity correlates with release date ($ρ= +0.55$, $p = .034$) while Detoxify does not ($ρ= -0.23$, $p = .42$): toxicity scores fall as representational harm grows. We formalise harm laundering as a three-criteria test and provide a three-stage detection protocol applicable to any generative model. Within the OpenAI GPT lineage, toxicity score reduction is not a sufficient proxy for harm reduction.

Suppress vulnerabilities applying Kubernetes context to scans - alegrey91/vex8s

Play DSS and DS2 files free in your browser. Open supported Olympus and Philips dictation recordings in one click. No audio upload or sign-up. Review dictation with speed control, bookmarks and A-B loops on Mac or Windows.

OpenWand - A LLM chat interface optimized for AI co-work for your desktop. Press a key, pick an intent, and OpenWand reads the right context, then streams an answer without making you leave what yo...

DigitalOcean is giving $3 million to the foundation behind DHH

Gapless playout, true-PFL cueing, and live streaming (Icecast + Shoutcast) — free and open source.

by John Ambrosio

TypeSeer is a system-wide AI autocomplete for macOS. It finishes your sentences as grey text in any app, learns how you write, and keeps up in 15 languages. The AI runs on your Mac, so nothing you type leaves your device.

This tool checks if your email has been breached on the darkweb - Aaronmike481/BreachScanner

Scan Python dependencies for exploited CVEs and abandoned packages that handle untrusted input, ranked by CISA KEV and EPSS. Runs in CI and as a Claude Code hook. - binuka200/package-doctor

Explore, customise, share and download randomly generated gems.

Explore a 3D map of email newsletters. Two newsletters are joined when they linked to the same pages, and clusters show who reads and shares the same things.