Claude Code's source code has been leaked via a map file in their NPM registry

Something went wrong, but don’t fret — let’s give it another shot. Some privacy related extensions may cause issues on x.com. Please disable them and try again.

Axios compromised on NPM – Malicious versions drop remote access trojan

https://cdn.prod.website-files.com/673b71f0790aabf30bd30bf8/69cb2363fdc3f8e8fa0460a5_blog-cover-image.png
StepSecurity identified malicious versions of the axios HTTP client library published to npm, [email protected] and [email protected], which inject a remote access trojan (RAT) dropper. Developers who installed these versions should rotate all secrets and credentials, check network logs, and downgrade to safe versions, and StepSecurity provides end-to-end npm supply chain security across three pillars: ...

The Claude Code Source Leak: fake tools, frustration regexes, undercover mode

Anthropic accidentally exposed Claude Code's source code and internal features, including anti-distillation mechanisms and a companion system. The leak reveals product roadmap details that competitors can now see and react to.

Oracle slashes 30k jobs

Oracle laid off 20,000-30,000 employees, roughly 18% of its workforce, in a single email with no advance notice. The cuts are tied to Oracle's aggressive expansion into AI infrastructure, freeing up $8-10 billion in cash flow.

Artemis II is not safe to fly

https://idlewords.com/images/oig_heat_shield.jpg
NASA's Orion spacecraft has a defective heat shield that could kill the crew on Artemis II due to spalling, impact from heat shield fragments, and bolt erosion. Despite this, NASA is planning to fly the mission with a crew, citing a change in the re-entry trajectory and a new heat shield design for future missions.

Ollama is now powered by MLX on Apple Silicon in preview

https://files.ollama.com/ollama_mlx.png
Ollama now runs faster on Apple silicon with MLX framework, leveraging GPU Neural Accelerators for speedup. Ollama 0.19 sees 1851 token/s prefill and 134 token/s decode with improved memory efficiency and model accuracy.

GitHub backs down, kills Copilot pull-request ads after backlash

https://regmedia.co.uk/2024/05/21/github1_shutterstock.jpg
GitHub removed Copilot's ability to insert ads into pull requests after backlash from developers. The feature was disabled after users complained of unwanted ads in their pull requests.

Microsoft: Copilot is for entertainment purposes only

https://cdn-dynmedia-1.microsoft.com/is/image/microsoftcorp/MSFT-Learn-Hero-Alt2_tbmnl_en-us?scl=1
You agree to these terms by using Copilot, which includes rules for using the service, protecting others, and respecting Microsoft's rights. By using Copilot, you grant Microsoft permission to use your content and agree to their terms, including the Microsoft Services Agreement and the Microsoft Privacy Statement.

A dot a day keeps the clutter away

https://scottlawsonbc.com/static/dot-system-01.jpg
The author created a simple inventory system using colored dot stickers to track usage of electronic components and tools in their lab. The system reveals patterns and helps the author decide what to keep and what to discard, making it easier to manage their collection.

OpenAI closes funding round at an $852B valuation

https://image.cnbcfm.com/api/v1/image/108283004-17744599851774459982-44807450682-1080pnbcnews.jpg?v=1774459984&w=750&h=422&vtcrop=y
OpenAI closed a record-breaking funding round at $852 billion post-money valuation with $122 billion committed capital. The company raised $3 billion from individual investors and generates $2 billion in monthly revenue.

GitHub's Historic Uptime

View GitHub's monthly uptime between 2016 and 2026.

OkCupid gave 3M dating-app photos to facial recognition firm, FTC says

https://cdn.arstechnica.net/wp-content/uploads/2026/03/okcupid-640x427.jpg
OkCupid and Match Group settled with the Trump administration over sharing 3 million user photos with a facial recognition firm without consent. They agreed to a permanent prohibition on misrepresenting data use and sharing, without paying a fine.

Universal Claude.md – cut Claude output tokens

https://opengraph.githubassets.com/51e61dfbcd98b9faca0cb7e47d57dfdbf9b19326ffa23c3c1c377eea914ef093/drona23/claude-token-efficient
A CLAUDE.md file reduces Claude output verbosity by ~63% without code changes, targeting sycophancy, verbosity, and formatting noise. It's most beneficial for high-output use cases, and users can customize it to target specific failure modes and compose multiple files for different project types.

Why the US Navy won't blast the Iranians and 'open' Strait of Hormuz

https://responsiblestatecraft.org/media-library/image.jpg?id=65428245&width=1800&height=900&quality=74&coordinates=0%2C71%2C0%2C54
The US Navy's power projection is limited by shore-based anti-access and area denial systems, making it vulnerable to Iranian and Chinese missiles. This shift in naval warfare requires a reevaluation of investment in expensive instruments of national power.

Show HN: 1-Bit Bonsai, the First Commercially Viable 1-Bit LLMs

https://cdn.prod.website-files.com/697a3312d33c2cc715ec3899/69cae04a19115963ea13d12d_prism-og-img%20(2).png
PrismML's ultra-dense intelligence models, like 1-bit Bonsai, offer significant memory and energy savings. They achieve high performance and accuracy while requiring much less memory and energy than full-precision models.

Open source CAD in the browser (Solvespace)

SolveSpace has a web version that runs in the browser with some speed penalty and bugs. It's experimental and can be hosted locally like static web content.

Claude Code users hitting usage limits 'way faster than expected'

https://regmedia.co.uk/2016/03/11/empty-gauge.jpg
Users of Claude Code are experiencing high token usage and early quota exhaustion due to a combination of factors including reduced quotas during peak hours, bugs, and unclear usage limits. Anthropic is actively investigating the issue and users are negotiating with providers over acceptable pricing and usage models for AI development.

Google's 200M-parameter time-series foundation model with 16k context

https://opengraph.githubassets.com/3a715ab5ed97409698fa19e1f50846332c191dbd18b04dbc7566243837cc8897/google-research/timesfm
TimesFM is a pretrained time-series model by Google Research for forecasting. It can be installed via pip and used for point and quantile forecasting.

MiniStack (replacement for LocalStack)

https://ministack.org/logo.png
AWS services run on real infrastructure, including databases, containers, and caching. LocalStack offers a Pro version with real infrastructure support for AWS services.

Tell HN: Chrome says "suspicious download" when trying to download yt-dlp

Google's browser warns of a "Suspicious Download" for yt-dlp, a tool to download files from Google's servers. This is seen as browser monopoly abuse and misleading people, as Chrome also downloads files from various servers.

Slop is not necessarily the future

https://www.greptile.com/blog/ai-slopware-future/unnamed.png
AI models will write good code due to economic incentives, as good code is cheaper to generate and maintain. Economic forces will drive AI models to generate simpler, good code as it will be cheaper overall.

Fast and Gorgeous Erosion Filter

https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgeWotRn8KxGEbVTCIl1-znAH0ZDNSwFUf7hkp9GrJ4u5jfK8zS8-sA1SHtWyBaHSLx3Lskzcr0z_XzW9y4847lBQo8e_1TIHa49EQXT1XFbv0lui_ZKX-hNlpDChD3Wq3Rz9n4HcB3MwEosYU5Xi_IOTzADWDd18u4joJcgUOF94ZN-HNhLdZ_8tvnBpM/s1600/ErodedTerrain.png
The user explains an erosion technique for generating virtual landscapes, which involves using a special kind of noise to produce branching gullies and ridges while allowing fast and GPU-friendly evaluation. The technique has been developed and refined over time, with various approaches and tweaks to address issues such as chaotic distortions, inconsistent gully magnitudes, and non-straight gullies.

What changes when you turn a Linux box into a router

We can turn a Linux system into a router or switch by making 7 changes to the system, including enabling packet forwarding, defining a bridge, and configuring netfilter rules to filter traffic. These changes allow us to create a complete forwarding system, including enabling IP forwarding, defining a bridge, configuring netfilter rules, deploying a DHCP and DNS server, and vending WiFi ...

Cohere Transcribe: Speech Recognition

https://cdn.sanity.io/images/rjtqmwfu/web3-prod/8054a4393c0b87afbde5d6d4de810d08d2c4db26-3140x1420.png?auto=format&fit=max&q=90&w=1570
Cohere Transcribe is an open-source automatic speech recognition model that achieves state-of-the-art accuracy with a low word error rate of 5.42%. It is available for download and can be used for real-world transcription tasks across 14 languages.
https://raw.githubusercontent.com/timescale/pg_textsearch/main/images/tapir_and_friends_v1.0.0.png
pg_textsearch is a PostgreSQL extension for full-text search with BM25 ranking. It supports indexing, querying, and faceting with various language configurations and parameters.

Italy blocks US use of Sicily air base for Middle East war

https://www.politico.eu/cdn-cgi/image/width=1160,height=772,quality=80,onerror=redirect,format=auto/wp-content/uploads/2026/03/31/GettyImages-1394407486-scaled.jpg
EU countries consider alternatives to Hungary if Orbán wins, while progressive voters abandon center-left parties in Europe. Transatlantic tensions rise as the US rejects a Russian proposal, affecting Italy's leader and other European nations.

Herbie: Automatically improve imprecise floating point formulas

https://herbie.uwplse.org/logo-car.png
Herbie is a tool that improves the accuracy of floating-point expressions. It helps find and fix inaccuracies in code.

Gone (Almost) Phishin'

https://i0.wp.com/ma.tt/files/2026/03/CleanShot-2026-02-07-at-19.36.31%402x-1-1024x996.png?resize=604%2C587&quality=80&ssl=1
A scammer targeted the user's Apple account by sending fake password reset prompts and contacting Apple Support, pretending to be the user. The scammer's convincing tactics included a fake Apple Support call and a pixel-perfect replica website to steal the user's password.

I traced my traffic through a home Tailscale exit node

https://tech.stonecharioteer.com/images/posts/tailscale/mesh-connection.png
You set up a Tailscale exit node on your Proxmox box to route internet traffic through your home network, achieving full-tunnel VPN-like behavior. This setup shifts trust from your current network to your controlled exit node machine.

Combinators

https://tinyapl.rubenverg.com/combinators/same.svg
Combinators are functions that refer to their arguments without modifying them. They are often represented by bird names in APL, such as Kestrel and Cardinal.

GitHub Monaspace Case Study

https://lettermatic.com/_next/image?url=https%3A%2F%2Fcdn.sanity.io%2Fimages%2Fblwjvcya%2Fproduction%2F095a628b998a2e81d05e90e387b87b9ca3eeac08-1733x1229.png&w=1920&q=75
GitHub and Lettermatic collaborated to create Monaspace, a superfamily of five interchangeable typefaces for code editors. Monaspace offers high personalization and accessibility features, including Texture Healing, which improves legibility in monospace typefaces.

U.S. stocks are set to deliver their worst quarter in nearly four years

Please enable JS and disable any ad blocker

Ordinary Lab Gloves May Have Skewed Microplastic Data

https://lede-admin.nautil.us/wp-content/uploads/sites/70/2026/03/Currie_HERO_009052.png?resize=2880%2C1920
Scientists researching microplastics may have skewed data by wearing gloves that shed particles mimicking microplastics. Clean-room gloves are a safer option, with 100 false positives per millimeter squared, compared to 2,000 from other gloves.

Nobody is coming to save your career

https://images.unsplash.com/photo-1620416265040-cc777cad1883?crop=entropy&cs=tinysrgb&fit=max&fm=jpg&ixid=M3wzMDAzMzh8MHwxfHNlYXJjaHwxfHxtaXJyb3J8ZW58MHx8fHwxNzc0Mzg0MzEyfDA&ixlib=rb-4.1.0&q=80&w=1080
Your manager is not responsible for your career growth, you must take ownership and initiate conversations about your goals. To start, tell your manager you want to grow and discuss what's needed for a promotion, then take action to make progress toward your goals.

RubyGems Fracture Incident Report

https://rubycentral.org/assets/images/gem-logo--badge.svg?v=f673e7d640
Ruby Central's Open Source Committee faced a crisis in September 2025 known as the "RubyGems Fracture" due to a poorly communicated process to offboard two engineers, André Arko and Samuel Giddens, from the RubyGems.org service. The incident led to the removal of paid contributors and highlighted the need for better policies, procedures, and communication in managing access and offboarding in ...

Mr. Chatterbox is a Victorian-era ethically trained model

https://static.simonwillison.net/static/2026/chatterbox.jpg
Trip Venturella released Mr Chatterbox, a language model trained on 28,000 Victorian-era British texts. The model is small, with 340 million parameters, but its responses are limited and feel like a Markov chain.

Costco sued for seeking refunds on tariffs customers paid

https://cdn.arstechnica.net/wp-content/uploads/2026/03/GettyImages-855684128-640x427.jpg
Costco customers have filed a class action lawsuit accusing the company of unjust enrichment by passing on tariff costs to consumers and planning to keep the refunds. The lawsuit claims Costco will recover the same tariff payments twice, potentially forcing other businesses to refund customers nationwide.

Maze Algorithms (1997)

https://www.astrolog.org/labyrnth/bower/cartoon3.gif
Mazes can be classified into seven categories: Dimension, Hyperdimension, Topology, Tessellation, Routing, Texture, and Focus, with various subcategories and algorithms for creating each type. Algorithms for creating Mazes include Perfect, Braid, Unicursal, Sparseness, 3D, Weave, Crack, Omega, Hypermaze, and Template, each with its own characteristics and methods for generating the Maze.

Safeguarding cryptocurrency by disclosing quantum vulnerabilities responsibly

https://storage.googleapis.com/gweb-research2023-media/original_images/Quantization-hero.gif
Google researchers warn that future quantum computers may break elliptic curve cryptography used in cryptocurrency and other systems with fewer qubits and gates than previously thought. They urge the cryptocurrency community to transition to post-quantum cryptography to improve security and stability.

Linux Running in a PDF (2025)

Sony halts memory card shipments due to NAND shortage

https://www.techzine.eu/wp-content/uploads/2026/03/shutterstock_2573233579-768x512.jpg
Sony is temporarily halting orders for CFexpress and SD memory cards due to a NAND flash shortage driven by AI data center demand. Resumption is expected in late 2027 or 2028 when NAND production increases.

Accidentally created my first fork bomb with Claude Code

The user created a hook that caused a fork bomb, bricking their computer due to excessive memory usage, but luckily the computer's caching saved it from further damage. The user had been experimenting with Claude, a large language model, and built various tools to help with their work, but the experience was imperfect and costly, resulting in a $3800 API bill.

Zml-smi: universal monitoring tool for GPUs, TPUs and NPUs

https://zml.ai/img/posts/zml-smi/1.png
zml-smi is a diagnostic tool for GPUs, TPUs, and NPUs providing real-time performance and health insights. It supports NVIDIA, AMD, Google TPU, and AWS Trainium devices with various metrics.

Ask HN: Distributed data centers in our basements

The idea of using home basements as mini data centers to reduce energy consumption and increase efficiency is unrealistic due to various technical, security, and regulatory challenges. However, it could be feasible for personal use or small-scale services, especially with advancements in hardware and software security, and if bandwidth and connectivity issues can be resolved.

Incident March 30th, 2026 – Accidental CDN Caching

https://s3-us-west-2.amazonaws.com/public.notion-static.com/535761c1-ecdb-4bed-b7c5-91f7eeb44bd4/Screen_Shot_2021-06-08_at_11.08.11_AM.png
Railway experienced a 52-minute incident where CDN caching was accidentally enabled for some domains, potentially serving unauthenticated data to authenticated users. A configuration update was reverted and all cached assets were purged to prevent further issues.

Scotty: A beautiful SSH task runner

https://raw.githubusercontent.com/spatie/scotty/refs/heads/main/docs/images/scotty-run-deploy.jpg
Scotty is a new SSH task runner that lets you define deploy scripts and run them from your terminal with real-time output. It supports both Blade and plain bash formats, and offers features like pause and pretend modes.

7,655 Ransomware Claims in One Year: Group, Sector, and Country Breakdown

https://ciphercue.com/img/og-card.png
Ransomware groups posted 7,655 victim claims to public leak sites from March 2025 to March 2026, with Qilin being the most active group posting 1,179 claims across 74 countries. The top 5 groups accounted for 40% of the claims, and the remaining 124 groups collectively posted 4,628 claims, suggesting that disrupting any single group is unlikely to reduce the overall total significantly.

Securing Elliptic Curve Cryptocurrencies Against Quantum Vulnerabilities [pdf]

Researchers from Google Quantum AI and other institutions have published a whitepaper warning of the potential risks of quantum computers to blockchain security, particularly in cryptocurrencies that rely on Elliptic Curve Discrete Logarithm Problem (ECDLP) cryptography. They estimate that a quantum computer with fewer than half a million physical qubits could break ECDLP-based cryptography ...

Prefer do notation over Applicative operators when assembling records (2024)

https://haskellforall.com/2024/05/prefer-do-notation-over-applicative/opengraph-image?133960dfd41055ac
Use do notation to assemble records for better ergonomics and approachability. It's order-insensitive and produces clearer error messages.

GitHub has DMCA'd nearly all forks of the official Claude-code repo

https://opengraph.githubassets.com/6c0e62b6c081bfb8e88ce6771b103496e4925c63629873a583ac47f73b814eeb/anthropics/claude-code
Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands. - Forks · anthropics/claude-code

Breaking Enigma with Index of Coincidence on a Commodore 64

https://imapenguin.com/images/2026/04/enigma-ic-found.webp
The Index of Coincidence (IC) attack is a method to crack Enigma codes without a crib, using the statistical properties of the German language to identify the correct rotor settings. The IC attack requires decrypting the entire message for every candidate setting, which is slower than a crib-based attack but can be done on a Commodore 64 in 82 hours, producing 18,165 candidates.

Claude Code full source code leaked on NPM

https://opengraph.githubassets.com/92000bd2198251a1b05324a790c7ac1451a5ed0c868e546da433e8a6ebd876ba/chatgptprojects/claude-code
This repository extracts and preserves the original TypeScript source code of Anthropic's Claude Code CLI tool from its npm package. It unpacks the source map to make the code easier to read and reference.

Why Inventing Color TV Was So Difficult [video]

Super Micro Computer Investors Look for Exits

https://catenaa.com/wp-content/uploads/2025/11/Baidu-Unveils-Two-New-AI-Chips-To-Replace-US-Chips-1024x683.webp
Super Micro Computer faces investor exodus due to recent indictment of its co-founder and past financial issues. Despite soaring sales, the company's stock price has dropped 63% since July 2025.

Objections to systemd age-attestation changes go overboard

A pull request to add a field to store a user's birth date in systemd's JSON user records has sparked a hostile response from some community members, with some receiving death threats and doxxing. The change aims to facilitate compliance with age-attestation and -verification laws, but critics argue it's a step towards surveillance and should be targeted at the people pushing for these laws ...

Bourbon waste could provide next-gen supercapacitor components

https://spectrum.ieee.org/media-library/an-adult-man-in-safety-goggles-and-gloves-holding-up-a-beaker-filled-with-opaque-bourbon-waste.jpg?id=65400140&width=1200&height=900
Researchers in Kentucky have found a way to convert bourbon waste into high-performance energy-storage materials. This could help distilleries turn a waste stream into a source of profit and meet the growing demand for energy storage.

Closed Source AI = Neofeudalism

Good people in AI labs are driven by a desire to contribute to science, not power, but institutions push towards concentration of power and control. A free technical order prioritizes safety, openness, and local control to prevent a single entity from curating the future of AI.

Forth VM and compiler written in C++ and Scryer Prolog

https://opengraph.githubassets.com/325162b95cbfed114af5c4c3939af7aa9c0bcb642a1336b65cea4440849ac9ab/no382001/forth-vm
The sets language is a statically-typed s-expression compiler with a 16-bit stack VM in C++20 and a Prolog compiler. It has a flat byte-addressable memory model with fixed-size stacks and a heap region for named variables.

In Expanding de Sitter Space, Quantum Mechanics Gets More Elusive

https://www.quantamagazine.org/wp-content/uploads/2026/03/Shalma-Wegsman-alt-profile.webp
Physicists are struggling to understand the quantum world in an expanding universe, particularly in de Sitter space where space expands exponentially. They are trying to learn from black holes to make sense of quantum mechanics in de Sitter space.

Ask HN: Academic study on AI's impact on software development – want to join?

Researchers at New York University and City, University of London are conducting a study on AI's impact on software development. They are seeking US-based developers to share their experiences and perspectives on using AI tools in their day-to-day work via a 45-60 minute Zoom interview.

DCJ11Hack+ – DEC PDP/11 based homebrew computer

https://codeberg.org/TechPaula/DCJ11HackPlus/media/branch/main/images/DCJ11_CPU.png?raw=true
You're sharing a DCJ11 based board with a backplane for adding/removing functionality. It includes a main CPU board, RAM, ROM, and text display board with schematics and PCB files available.

On the trail of ancient art, deep in the Sahara

The following information can help our support team to resolve this issue.

JSSE: A JavaScript Engine Built by an Agent

https://p.ocmatos.com/img/2026/03/jsse-agent-dreaming.png
The author built a JavaScript engine called JSSE from scratch in six weeks using a single agent, Claude Code, without writing any Rust code, and it passed 100% of test262 non-staging tests. The project demonstrated the power of agentic coding and the importance of a good plan, and it laid the groundwork for future improvements in software production.

Show HN: PhAIL – Real-robot benchmark for AI models

https://phail.ai/phail-static/phail-icon.png
Five leading models. One commercial task. Production metrics.

Windows++: C++ Application Framework for Windows by Paul DiLascia

http://pauldilascia.com/images/wpp.jpg
Windows++ is a lightweight C++ framework for Windows that simplifies programming and reduces code size. It supports multiple compilers and runs on various Windows versions, with a small 84K DLL.

Mad Bugs: Vim vs. Emacs vs. Claude

https://substackcdn.com/image/fetch/$s_!bU3_!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F704d8e1c-2d23-40d2-ac1c-e0e9111169e1_570x460.png
A user reported an RCE 0-day in GNU Emacs, but maintainers declined to fix it, attributing it to git. A new bug-hunting initiative, MAD Bugs, is launching to publish AI-discovered bugs and exploits.

Project Mario: the inside story of DeepMind

https://colossus.com/wp-content/uploads/2026/03/ColossusMagazine_AIgovernance_eshakespeare_WEB-horizontal_FINALv1-scaled.jpg
Demis Hassabis and Mustafa Suleyman sought to create a governance structure for AI at DeepMind, but faced resistance from Google. They eventually secured a $1 billion investment from Reid Hoffman to pursue a spin-out, but negotiations with Google stalled.

Show HN: Pardus Browser- a browser for AI agents without Chromium

https://opengraph.githubassets.com/968bbaa0b5d8db76fc20eb99c2ba65d3a67ee5f53b6c893b8d733dc54a716a35/JasonHonKL/PardusBrowser
Pardus-browser is a headless browser for AI agents that fetches URLs, parses HTML, and outputs a clean semantic tree in milliseconds. It supports various formats, including Markdown, tree, and JSON, with options for custom headers, JavaScript execution, and verbose logging.

Show HN: Cerno – CAPTCHA that targets LLM reasoning, not human biology

https://cerno.sh/og.png
01 Proof of work SHA-256 hash prefix, 14–24 bits. Adaptive difficulty based on client signals. 03 Motor-control analysis 12 behavioral features (7 public + 5 secret, server-only) extracted from raw pointer events. Scored against per-maze baselines. 05 Signature binding ECDSA P-256 ephemeral keypair. Public key bound at challenge issuance, verified on submission.

The story of Britain's oldest sweet, the Pontefract Cake (2019)

https://ichef.bbci.co.uk/images/ic/480xn/p07gf54q.jpg.webp
The Oldest Sweet Shop in England in Pateley Bridge sells Pontefract Cakes, a treacle-coloured sweet with a 1,000-year history. The sweet's origins date back to the 11th Century when monks brought liquorice to Yorkshire.

Show HN: Raincast – Describe an app, get a desktop app (open source)

https://raw.githubusercontent.com/tihiera/raincast/main/src-tauri/icons/128x128%402x.png
Raincast is a native desktop app generator that builds real, shippable applications from natural language descriptions. It uses AI to create fully functional apps with UI, backend commands, file system access, and system integration in various layouts.

Show HN: Hyprmoncfg – Terminal-based monitor config manager for Hyprland

https://paolino.me/images/hyprmoncfg-demo.gif
hyprmoncfg is a TUI tool for configuring Hyprland monitors with spatial layout and workspace planning features. It's a lightweight, portable, and SSH-friendly alternative to other tools with a focus on reliability and ease of use.