Popular：
Virtualization DNS security formal verification reachability analysis compiler errors macro conflict web extension development framework Bitmap Graphics API inconsistencies All Tags

Swiss Canton Buys Typo-Filled Domain to Fix 100k Flyer Error

2025-01-31
Swiss Canton Buys Typo-Filled Domain to Fix 100k Flyer Error

The Basel-Stadt canton in Switzerland accidentally omitted '.ch' from the URL printed on over 100,000 tax filing flyers, resulting in a '.bs' (Bahamas) domain. Instead of reprinting at a cost of roughly $100,000, they opted to purchase the erroneous domain for $1,000 and set up a redirect to the correct URL. The redirect is not yet live, pending domain registration completion.

Read more
(techcrunch.com)
Misc domain typo printing error government efficiency

Netflix's Unified Data Architecture: Model Once, Represent Everywhere

2025-06-14
Netflix's Unified Data Architecture: Model Once, Represent Everywhere

Netflix's exploding content offerings — films, series, games, live events, ads — have created a complex web of supporting systems. To tackle duplicated models, inconsistent terminology, and data quality issues, Netflix built the Unified Data Architecture (UDA). UDA is a knowledge graph enabling teams to define models once and reuse them consistently across systems. Leveraging an internal metamodel called Upper, UDA translates domain models into various technical data structures (GraphQL, Avro, SQL, etc.), automating data movement and transformation between containers. This boosts efficiency and data consistency. Two production systems, Primary Data Management (PDM) and Sphere, showcase UDA's power, handling authoritative reference data and self-service operational reporting respectively.

Read more
(netflixtechblog.com)
Development Data Architecture

The Flawed On-Call System in Big Tech and the Promise of AI

2025-03-18

This article explores the shortcomings of the on-call system prevalent in large tech companies. Drawing on personal experience, the author contrasts the approaches of big tech and startups in handling software issues. Big tech's incentive structures lead to declining software quality, making on-call a permanent fixture. The author argues that AI and machine learning can revolutionize on-call processes, for example, using LLMs to simplify logs, search for similar issues, and reduce the burden on engineers, thus increasing efficiency.

Read more
(reflector.dev)
Development On-Call

Revitalizing Your Mavericks: Giving an Old System New Life

2025-08-21

This guide shows you how to breathe new life into your aging OS X Mavericks system. It involves a simple script to update the system, replacing the outdated browser with a modern Firefox (Firefox Dynasty), using Aqua Proxy to fix HTTPS compatibility issues, adding new emojis, and setting up Time Machine backups. The guide also instructs users on deleting unwanted pre-installed apps such as Chess, iTunes, and iBooks for a personalized experience. The entire process is straightforward, making your Mavericks system run smoother and more tailored to your needs.

Read more
(mavericksforever.com)
Development

Briar: Decentralized Messaging App for Activists and Journalists

2025-03-14

Briar is a messaging app designed for activists, journalists, and anyone needing secure communication. Unlike traditional apps, Briar doesn't rely on central servers; messages sync directly between devices. Offline, it uses Bluetooth, Wi-Fi, or memory cards. Online, it leverages Tor for enhanced privacy. Briar resists surveillance and censorship by employing end-to-end encryption and a decentralized architecture. It offers private messaging, public forums, and blogs, protecting against metadata surveillance, content filtering, takedown orders, and denial-of-service attacks. Briar's long-term vision extends beyond messaging, aiming to support secure, distributed applications for crisis mapping and collaborative work, fostering safe spaces for communication and organization globally.

Read more
(briarproject.org)
Development censorship-resistant

SQLite Page Explorer: A GUI for Peeking Inside Your Databases

2025-02-06
SQLite Page Explorer: A GUI for Peeking Inside Your Databases

A small GUI application, built with redbean, lets you explore your SQLite databases page by page, just as SQLite sees them. It's a single 6.5MB executable running natively on Windows, Linux, macOS, and more, offering insights into how indexes are stored, data compactness, and B-tree structures. While potential virus warnings exist due to the use of a polyglot executable, the project is trustworthy and offers a unique perspective for developers. It's a fun project that may be slow with larger databases.

Read more
(github.com)
Development

A Year After Ditching React: Scholarly's Server-Side Rendering Success

2025-01-22

Scholarly, a rapidly growing company, has maintained a remarkably simple tech stack: Rails, Stimulus, and MySQL, supplemented by Turbo and ActionCable for enhanced interactivity. Their experience after a year of abandoning React demonstrates that this server-side rendering approach significantly improves testing efficiency, application speed, and developer productivity. Compared to complex JS frontends, this lightweight approach reduces code maintenance costs, minimizes risks, and enables the team to deliver value faster, focusing on core business goals.

Read more
(kellysutton.com)
Development

Quantum Leap: Monolithic Integration of Photonic Quantum System on a Chip

2025-07-20
Quantum Leap: Monolithic Integration of Photonic Quantum System on a Chip

Scientists from Northwestern University, Boston University, and UC Berkeley have achieved a breakthrough: integrating a miniature photonic quantum system onto a conventional electronic chip. This 1mm² chip generates quantum light and incorporates a smart electronic system for stabilization, reliably producing photon pairs for light-based quantum communication, sensing, and processing. Fabricated by a commercial semiconductor foundry, the chip demonstrates scalability potential, representing a crucial step towards larger quantum photonic systems and opening doors for applications in computing, sensing, and communication.

Read more
(news.northwestern.edu)
Tech chip integration

2100-Year-Old Alexander the Great Mosaic Undergoes Non-Destructive Analysis

2025-01-22
2100-Year-Old Alexander the Great Mosaic Undergoes Non-Destructive Analysis

Researchers have used non-invasive techniques to analyze a 2,100-year-old mosaic depicting the Battle of Issus, housed at the National Archaeological Museum of Naples. Multispectral imaging, X-ray fluorescence, and spectroscopy revealed the mosaic's intricate composition and signs of degradation. The study identified various materials used in its construction, including Italian and Mediterranean marbles and volcanic rocks, as well as gypsum and wax from past restorations. The findings also highlighted thermal instability and structural issues, providing valuable data for future restoration efforts.

Read more
(phys.org)
Archaeology conservation Alexander the Great

Kim Jong-il, Kidnapped Filmmakers, and a Godzilla Rip-off: The Bizarre Story Behind Pulgasari

2025-05-05
Kim Jong-il, Kidnapped Filmmakers, and a Godzilla Rip-off: The Bizarre Story Behind Pulgasari

A seemingly abandoned building in Brooklyn houses the Spectacle theater, screening cult and cutting-edge cinema. Recently, it showcased Pulgasari, a bizarre Godzilla knockoff allegedly produced under the direction of Kim Jong-il. The film's story is even more outlandish than the movie itself. To elevate North Korea's film industry, Kim Jong-il kidnapped renowned South Korean director Shin Sang-ok and actress Choi Eun-hee, forcing them to create films. Their final collaboration, Pulgasari, became their ticket to escape. The film, a mix of political allegory and low-budget camp, is a uniquely demented spectacle.

Read more
(www.vanityfair.com)
Misc Kim Jong-il North Korean Cinema

Nova Act SDK: A Crucial Step Towards Reliable Agents

2025-03-31
Nova Act SDK: A Crucial Step Towards Reliable Agents

The Nova Act SDK simplifies the development of intelligent agents by allowing developers to break down complex workflows into atomic commands (like search, checkout, answering on-screen questions), add more detailed instructions to these commands (e.g., "don't accept the insurance upsell"), and call APIs, thus improving reliability. As intelligent agents are still in their early stages, the Nova Act SDK represents a crucial advancement.

Read more
(www.aboutamazon.com)
AI intelligent agents

Snow Signs: A Cross-Cultural Journey Through Time

2024-12-25
Snow Signs: A Cross-Cultural Journey Through Time

This article explores the diverse ways different cultures around the world have represented 'snow' in writing and symbolism, from the Shang oracle bone script in ancient China to the 'wind, flowers, snow, and moon' motif on ancient Chinese wine jugs, and from the Naxi Dongba script to the Inuktitut language's detailed descriptions of various snow types. The article also delves into snow-related words and symbols in ancient Greek, Egyptian, Hebrew, and Mayan civilizations, showcasing the unique understandings and expressions of snow across cultures, highlighting their cultural contexts and historical origins.

Read more
(viewsproject.wordpress.com)
Misc writing systems cross-cultural comparison snow

AMD Zen Chief Architect Interview: Unpacking the Secrets of Low-Power x86 Design

2025-03-26
AMD Zen Chief Architect Interview: Unpacking the Secrets of Low-Power x86 Design

This article presents a transcript of an interview between Casey and Mike Clark, the chief architect of AMD's Zen. The discussion centers on low-power design in x86 architectures. Clark dispels the myth that the x86 ISA inherently hinders low-power design, emphasizing the role of market strategy and design priorities. He explains how AMD improves energy efficiency through microarchitectural optimizations (like TLBs and uop caches), balancing bandwidth and power consumption. The interview delves into instruction set size, cache line size, scatter/gather operations, non-temporal stores, CPU pipeline diagrams, and how software developers can better leverage hardware features, offering invaluable insight into modern CPU design.

Read more
(www.computerenhance.com)
Hardware low-power design

100 Billion Pixels on the Moon: A Permanent Record of Humanity

2025-01-16
100 Billion Pixels on the Moon: A Permanent Record of Humanity

An international project called 'Sanctuary on the Moon' aims to create a lunar time capsule containing 100 billion pixels of information, offering a detailed guide to our civilization. Backed by NASA, UNESCO, and the French government, the project seeks to leave a lasting legacy for future humanity. The capsule consists of 24 sapphire discs, each engraved with information on specific aspects of human knowledge, including the human genome. The project focuses on preserving information in a way that will be easily decipherable by future humans, even millions of years from now, ensuring the survival of crucial knowledge and artistic achievements.

Read more
(www.aljazeera.com)
Tech Time Capsule Human Civilization

Proton Launches Lumo: A Privacy-First AI Assistant to Challenge Big Tech

2025-07-24
Proton Launches Lumo: A Privacy-First AI Assistant to Challenge Big Tech

In response to Big Tech's use of AI to fuel surveillance capitalism, Proton introduces Lumo, a privacy-first AI assistant. Lumo keeps no logs, employs zero-access encryption for all chats, and ensures users retain complete control of their data, never sharing, selling, or stealing it. Lumo offers a secure alternative, allowing users to enjoy AI benefits while protecting their privacy. Built on open-source language models and operating from Proton's European datacenters, Lumo features unique privacy tools like 'Ghost Mode'. This launch represents Proton's commitment to building a European sovereign tech stack and underscores its dedication to data privacy and user rights.

Read more
(proton.me)
AI privacy AI

Clarification on Alleged ESP32 Backdoor

2025-03-11
Clarification on Alleged ESP32 Backdoor

Recent media reports claimed an ESP32 chip backdoor. Espressif clarifies that the reported functionality is internal debug commands for testing, not remotely accessible via Bluetooth, radio, or internet. These commands pose no security risk by themselves, though Espressif will provide a software fix to remove them. Only ESP32 chips are affected; ESP32-C, ESP32-S, and ESP32-H series are not. Espressif thanks the security researchers for their responsible disclosure.

Read more
(www.espressif.com)
Hardware

Docker Hub Usage Limits and Abuse Rate Limiting Explained

2025-02-21
Docker Hub Usage Limits and Abuse Rate Limiting Explained

Docker Hub will enforce new usage limits starting March 1, 2025. Free users have limitations on the number of image pulls per month, while paid users enjoy higher quotas and on-demand options. The article details the quota limits for different user types, including monthly pulls, hourly pull rate limits, and public and private repository storage. Additionally, Docker Hub has abuse rate limiting to prevent malicious requests. Exceeding limits will result in throttling or additional charges.

Read more
(docs.docker.com)
Development usage limits quotas

UK Online Safety Act Forces Closure of Numerous Online Communities

2025-02-23

The UK's Online Safety Act 2023 has led to the closure or geo-blocking of numerous online communities and forums. From a Mastodon instance for AWS discussions to a long-standing green living forum with nearly 500,000 posts, and even a cycling forum, many have succumbed to the pressures of the new legislation. This raises concerns about freedom of speech online and the survival of online communities, highlighting the tension between online safety legislation and the thriving of online spaces.

Read more
(onlinesafetyact.co.uk)
Tech Online Communities Freedom of Speech

Rules to Avoid Common Extended Inline Assembly Mistakes

2024-12-21

This article isn't an inline assembly tutorial, but rather a summary of six rules to avoid common mistakes. The author emphasizes that inline assembly is treacherous and should be avoided whenever possible, as modern compilers offer intrinsics and built-ins for most use cases. If unavoidable, the rules are: use the `volatile` qualifier; use a `memory` clobber if accessing memory; never modify input constraints; never call functions from inline assembly; don't define absolute assembly labels; and use the assembler's local label feature to avoid label conflicts. The article concludes by encouraging readers to practice applying these rules by reviewing online tutorials and even LLM-generated code.

Read more
(nullprogram.com)
Development inline assembly compiler

TinyCompiler: A Weekend Compiler Project

2025-02-20

This project details the creation of TinyCompiler, a minimalist compiler built in a weekend. It translates the esoteric Wend programming language (created by the author) into GNU assembly. Wend is a simple language, omitting pointers, arrays, and other complexities, focusing on core compiler concepts. The entire project is under 500 lines of Python and includes test programs like fixed-point square root calculation, Mandelbrot set rendering, and simple games. It's a great resource for learning about compiler theory.

Read more
(ssloy.github.io)
Development

Deep Dive into Hygienic Macros in Scheme

2025-05-08

Scheme's macro system employs a 'hygiene' mechanism to prevent variable name clashes during macro expansion. This article delves into the concept of identifiers in Scheme, which encompass not only a symbolic name but also a lexical context and a historical context. The predicates `bound-identifier=?` and `free-identifier=?` compare identifier equivalence; the former focuses on the interchangeability of bound identifiers after macro expansion, while the latter focuses on free identifiers. The article uses multiple examples to illustrate the differences between these two equivalence relations and the role of historical context in the hygienic macro mechanism. Ultimately, it explains how Scheme uses the historical information of identifiers to ensure macro hygiene and prevent variable name conflicts.

Read more
(mnieper.github.io)
Development

The Ethical Quandary of LLMs: Why I've Stopped Using Them

2025-02-19

This post delves into the ethical concerns surrounding Large Language Models (LLMs) and explains the author's decision to stop using them. The author explores five key issues: energy consumption, training data sourcing, job displacement, inaccurate information and bias, and concentration of power. High energy usage, privacy concerns related to training data, the potential for job displacement, the risk of misinformation due to biases and inaccuracies, and the concentration of power in the hands of a few large tech companies are highlighted as significant ethical problems. The author argues that using LLMs without actively addressing these ethical concerns is unethical.

Read more
(ntietz.com)
AI Ethics

arXivLabs: Experimental Projects with Community Collaborators

2025-04-03
arXivLabs: Experimental Projects with Community Collaborators

arXivLabs is a framework enabling collaborators to develop and share new arXiv features directly on our website. Individuals and organizations partnering with arXivLabs embrace our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners who adhere to them. Have an idea for a project that will benefit arXiv's community? Learn more about arXivLabs.

Read more
(arxiv.org)
Development

Tesla Recalls Over 376,000 Model 3 and Y EVs Over Power Steering Issue

2025-03-02
Tesla Recalls Over 376,000 Model 3 and Y EVs Over Power Steering Issue

Tesla is recalling over 376,000 Model 3 and Model Y electric vehicles in the U.S. due to a potential power steering issue. The problem stems from an "overstress condition" in the electronic power-assisted steering (EPAS) printed circuit board, potentially causing a loss of power steering assist when the vehicle stops and accelerates again. Tesla started rolling out a free over-the-air software fix in October 2023. The company claims the issue won't affect steering while the vehicle is in motion. No injuries, deaths, or crashes have been reported. Owners won't need to visit a dealership and will receive notification by mail next month.

Read more
(www.foxbusiness.com)
Tech

Offline Reinforcement Learning Boosts Multi-Step Reasoning in LLMs

2024-12-23
Offline Reinforcement Learning Boosts Multi-Step Reasoning in LLMs

Researchers introduce OREO, an offline reinforcement learning method designed to enhance the multi-step reasoning capabilities of large language models (LLMs). Building upon maximum entropy reinforcement learning, OREO jointly learns a policy model and value function by optimizing the soft Bellman equation. This addresses limitations of Direct Preference Optimization (DPO) in multi-step reasoning, specifically the need for extensive paired preference data and the challenge of effective credit assignment. Experiments demonstrate OREO's superiority over existing offline learning methods on benchmarks involving mathematical reasoning and embodied agent control.

Read more
(arxiv.org)
AI Multi-step Reasoning

We Built Loneliness Machines and Called Them Smart

2025-06-14
We Built Loneliness Machines and Called Them Smart

Since the advent of smartphones in 2010, they've become ubiquitous, yet this pervasiveness comes at a steep cost. This article argues that excessive smartphone use leads to addiction, loneliness, depression, and damage to mental and physical health. Furthermore, smartphones exacerbate social divisions and political polarization. While an outright ban is unrealistic, the author suggests collective action to mitigate their negative effects, including promoting a "right to disconnect", to regain freedom and well-being.

Read more
(jacobin.com)
Tech digital well-being

The Frankfurt Kitchen: A Modernist Icon and its Controversies

2025-06-11
The Frankfurt Kitchen: A Modernist Icon and its Controversies

Designed in 1926 by Margarete Schütte-Lihotzky, the Frankfurt Kitchen served as a standard prototype, widely implemented in the "New Frankfurt" housing project of the 1920s. Inspired by industrial efficiency and assembly-line production, it prioritized functionality and minimized space, its layout resembling a railway dining car kitchen. This aimed to 'industrialize' housework. However, the design also sparked controversy; while improving hygiene, it didn't challenge gender roles and was later criticized for neglecting individual needs. Today, the Frankfurt Kitchen stands as a significant chapter in modern design history, displayed in museums, showcasing both the brilliance and limitations of modernist ideals.

Read more
(museumderdinge.org)
Design Frankfurt School

Best-of-N Jailbreaking: A Novel Attack on AI Systems

2024-12-15
Best-of-N Jailbreaking: A Novel Attack on AI Systems

Researchers have developed a new AI attack algorithm called Best-of-N (BoN) Jailbreaking. This black-box algorithm repeatedly modifies prompts—randomly shuffling or capitalizing text, for example—until it elicits a harmful response from the AI system. BoN achieved impressively high attack success rates (ASRs) on closed-source language models like GPT-4o (89%) and Claude 3.5 Sonnet (78%), effectively circumventing existing defenses. Furthermore, BoN seamlessly extends to vision and audio language models, highlighting the vulnerability of even advanced AI systems to seemingly innocuous input variations. This research underscores significant security concerns in the field of AI.

Read more
(arxiv.org)
AI AI security AI attack Best-of-N Jailbreaking
1 2 … 490 491 492 494 496 497 498 … 596 597