Hacker NewsTuesday, November 18

Supercookie is a tracking method that uses favicons to assign unique identifiers to website visitors that persist even in incognito mode, after clearing cache, or using VPNs and ad blockers. Unlike traditional cookies, this tracking exploits browser favicon caching and works across all major browsers including Chrome, Firefox, Safari and Edge.

DarkPattern.Games is a new website that identifies gaming dark patterns - deliberately designed negative player experiences that benefit developers. Starting with iOS and Android games, they're working to expand reviews across platforms and encourage user submissions to build their database.

The author argues that Google's deprecation of XSLT support in Chrome is part of a deliberate strategy to destroy the open web rather than addressing legitimate security concerns. They claim Google could have fixed the security issues or adopted safer alternatives but instead chose to remove the feature and offer only an inadequate JavaScript polyfill that users must manually implement. The author advocates for resistance by continuing to use XSLT and flooding Google's issue tracker with complaints, viewing this as part of a broader pattern where tech corporations are undermining web standards to increase their control over the internet.

The article contrasts childhood among the BaYaka hunter-gatherers, where children roam freely with machetes and form independent peer groups, with Western children who are heavily supervised but spend hours online. It argues that declining childhood mobility since the 1970s due to car-dependency, urbanization, and parental fears has eliminated children's traditional spaces for independent play, forcing them to seek autonomy in digital worlds where they face new dangers like predators and harmful content.

Samsung's AppCloud app, preloaded on Galaxy phones in India and other markets, is facing scrutiny over privacy concerns and ties to Israeli company ironSource. The app forces users to install third-party apps during setup and lacks an accessible privacy policy. Consumer advocates want Samsung to provide clear opt-outs and stop preloading the app in sensitive regions.

The author explains how they switched from Tailscale to Cloudflare Zero Trust with Warp after NAT/firewall issues. They detail the key concepts including tunnels (exit points for traffic), routes (traffic direction), and targets (infrastructure pointers) that work together to create secure networks. The system allows exposing private services publicly, creating fully private networks, and adding granular access policies without NAT concerns since all traffic goes through Cloudflare's edge network.

The author reflects on how AI has made small utility libraries like their popular blob-util package obsolete, as developers can now generate similar code with LLMs instead of using dependencies. While this reduces supply-chain risks, it eliminates educational opportunities that well-documented libraries provided. The author concludes that the era of small utility libraries is ending, but larger, more innovative open source projects that LLMs can't easily replicate still have value and purpose in software development.

This content presents a humorous fictional premise where pop star Britney Spears is portrayed as an expert in semiconductor physics who will guide readers through the fundamentals of semiconductor laser components used in digital music formats. The site also mentions being a physics education website with sections on mechanics, optics, electricity, magnetism, thermodynamics, nuclear physics and cosmology.

Replicate, an AI platform that provides tools for developers to use AI models, is joining Cloudflare while maintaining its distinct brand. The acquisition aims to leverage Cloudflare's network infrastructure to make Replicate faster and better integrated with Cloudflare's developer platform, with existing APIs and models continuing to work unchanged.

The author explains building a custom search engine that uses existing databases instead of external services like Elasticsearch. The system works by tokenizing content into searchable pieces using multiple tokenizers (word, prefix, n-gram), storing tokens with weights in database tables, and scoring search results based on field importance and token relevance to provide accurate search functionality.