HN Morning Brief: April 11, 2026
HN Morning Brief: April 11, 2026
Hacker News this morning leaned toward tools, infrastructure, and the consequences of design choices, from laptop edges and extension stores to cloud agents, chip behavior, and lunar recovery operations. I filtered out duplicates from the previous brief, then pulled article details and selective HN discussion only for the final thirty stories below.
Tech Tools & Projects
Filing the corners off my MacBooks
Summary: The author physically filed down the sharp aluminum front corners of a MacBook, especially around the notch area. They taped off the speakers and keyboard, clamped the laptop to a workbench, and filed in small increments. Finishing was done with a rough file followed by 150-grit and 400-grit sandpaper. The modification was motivated by wrist discomfort from the stock sharp edge.
HN Discussion: Several commenters said recent flat MacBook edges dig into wrists and called the ergonomics poor compared with tapered laptop designs. A few readers loved the broader lesson of customizing tools to fit personal needs, regardless of convention. One commenter said edge pitting on plugged-in MacBooks had effectively turned the rim into a sawblade for them.
Bevy game development tutorials and in-depth resources
Summary: Tainted Coders presents free Bevy guides that the author says evolved from notes for friends into something like Rails Guides for Bevy. The site says the material is current for Bevy 0.18. It points beginners to a Pong tutorial and more advanced readers to a TLDR track. The author notes a Ruby-on-web background and says the site itself was built with their Ruby static site generator, Staticky.
HN Discussion: Commenters praised the site as one of the few free and in-depth Bevy learning resources. Several readers specifically appreciated that the guides are updated for Bevy 0.18 because many Bevy tutorials go stale quickly. One reader said they had read the whole site, thanked the author by email, and got a thoughtful reply asking what could be clearer.
Starfling: A one-tap endless orbital slingshot game in a single HTML file
Summary: Starfling is a browser game pitched as a one-tap space game where you release and chain slingshots between stars. The page is a single self-contained HTML app with canvas rendering, overlays, score tracking, pause and game-over states, and share UI. The code includes Firebase-backed leaderboard or stats plumbing for global and daily scores. It also contains mobile-app hooks for ads, revive flows, and launch waitlist capture for iOS and Android.
HN Discussion: Most commenters said the game is fun and immediately shared personal best scores and percentile readouts. Several people found the physics unintuitive, expecting stronger visible gravity wells or straighter tangent launches. Mobile players complained that combo text can cover the projectile and make chaining harder.
A practical guide for setting up Zettelkasten method in Obsidian
Summary: The article explains Zettelkasten as a network of atomic linked notes, not a folder or tag system. It recommends a minimal Obsidian vault structure with Inbox, Literature Notes, Permanent Notes, and Templates. It stresses rewriting source material into literature notes, then distilling single-idea permanent notes with clear titles. It warns that larger vaults accumulate orphaned notes, stale MOCs, inbox backlog, and duplicate ideas.
HN Discussion: Several commenters argued AI reduces the need for strict up-front note structure and makes large unorganized corpora more usable. One commenter said the post should credit Sönke Ahrens’s 2017 book How to Take Smart Notes. A few readers shared their own Obsidian or bot-based note systems instead of debating the method itself.
Watgo – A WebAssembly Toolkit for Go
Summary: Watgo is a pure Go toolkit for parsing WAT, validating modules, and encoding or decoding WASM binaries. Its core abstraction is wasmir, a semantic intermediate representation of a WebAssembly module. The CLI aims for wasm-tools compatibility and can parse, validate, and emit wasm binaries from text modules. The API exposes module structure for analysis and transformation after syntactic sugar is lowered into canonical form.
HN Discussion: Comments focused mainly on the testing approach rather than the CLI features. One reader said the official spec suite was invaluable for their own WASM parser and even helped uncover a regression in the published spec. Another thanked the author for the shared WAT samples and wanted to compare harness design with other projects.
Launch HN: Twill.ai (YC S25) – Delegate to cloud agents, get back PRs
Summary: Twill runs coding agents like Claude Code and Codex in isolated cloud sandboxes and returns PRs, reviews, diagnoses, or follow-up questions. The founders say the product solves three local-agent pain points: parallel task collisions, lack of persistence, and over-broad local trust. Tasks can be submitted through Slack, GitHub, Linear, a web app, or CLI, and each sandbox gets its own cloned repo and runtime isolation. The system snapshots finished sandboxes so later runs start warm with dependencies already installed.
HN Discussion: HN commenters broadly agreed cloud coding agents are a real direction, but questioned whether an independent third party can survive against labs and GitHub-native workflows. Security-minded replies pushed for tighter network egress controls, stronger secret handling, and self-hosted or behind-firewall deployment for enterprise use. Multiple users asked about computer-use support, sandbox implementation details, Docker-in-Docker, and trigger or automation coverage.
Static code analysis in Kotlin – tools overview
Summary: The article explains static analysis as AST-based inspection of Kotlin code without executing it. It compares detekt, diktat, and ktlint for enforcing code quality and especially for checking class member ordering. The team concludes none of the tools supports visibility-based method ordering out of the box. Detekt is presented as the most practical choice because it has broad built-in checks and a custom ruleset API.
HN Discussion: The HN thread had no visible substantive discussion. Both listed comments were dead, so there were no usable reader takeaways.
Show HN: FluidCAD – Parametric CAD with JavaScript
Summary: FluidCAD pitches browser-based parametric CAD where users write JavaScript and see 3D geometry update in real time. The tool supports standard CAD workflow elements such as sketches, extrusions, fillets, shells, booleans, and feature history navigation. It emphasizes interactive viewport actions like drag extrude, mouse-guided editing, and transforms or patterns over feature sequences. Interoperability includes STEP import and export, while setup centers on an npm package plus a VS Code extension for scene viewing.
HN Discussion: Reception is strongly positive, with several commenters comparing the code-plus-visual workflow to what made Flash or SketchUp approachable. Readers ask practical questions about supported operations, assemblies, revolve support, STEP or DXF export, and floating-point precision in JavaScript geometry. There is interest in richer paradigms too, especially declarative constraints instead of only imperative modeling code.
Show HN: A WYSIWYG word processor in Python
Summary: Miniword is a minimal Python word processor aiming for real WYSIWYG editing without an HTML layer or embedded browser. The project stresses lightweight startup, minimal dependencies, and a human-readable, diff-friendly, git-friendly file format. Feature highlights include styles, images, tables, Markdown support, and Python plugin extensibility. The current dependency stack is Python 3.9+, wxPython, and Cairo, with Linux as the main development platform and Windows or macOS intended to work too.
HN Discussion: HN readers like the non-browser approach and ask hard implementation questions around cursor positioning, selection, toolkit choice, and document data structures. A practical bug report says the project throws ModuleNotFoundError on macOS startup. Some commenters raise classic editor debates, including requests for WordPerfect-style reveal codes and skepticism that WYSIWYG is better than plain Markdown.
Web & Infrastructure
20 Years on AWS and Never Not My Job
Summary: Colin Percival recounts creating an AWS account on April 10, 2006, initially because S3 looked useful for secure backups. He pushed Amazon early on response signing, Xen security review, constrained credentials, and better consistency models. He worked extensively on getting FreeBSD and later NetBSD booting on EC2, often blocked by Xen 3.0 limitations. The post details multiple security reports, including SimpleDB signature collisions and unsafe serialized NextToken values.
HN Discussion: No HN comments were visible when fetched. The submission appeared to have no active public discussion yet.
What is RISC-V and why it matters to Canonical
Summary: The article describes RISC-V as an open ISA specification rather than a CPU implementation, with growing Linux-capable developer hardware expected in 2026. It argues the open standard enables flexible business models, from open cores to proprietary IP and in-house implementations. It highlights extensibility for AI, security, accelerators, low-power systems, and academic architecture work. It says the software stack is already mature across Linux, GCC, LLVM, RTOSes, and Ubuntu’s own RISC-V support since 2021.
HN Discussion: The biggest concern in comments was fragmentation, both in board platforms and in optional ISA extensions. One experienced commenter argued RISC-V still looks decades away from mainstream fast general-purpose computing. Some discussion focused less on RISC-V itself and more on distrust of Canonical’s track record with ecosystem forks and abandoned projects.
PGLite Evangelism
Summary: The post is about PGLite and explicitly frames it as an in-process alternative in the spirit of SQLite. Its published description says the appeal is avoiding a long-lived database service or Docker Desktop for local work. The piece appears to advocate PGLite as a way to get PostgreSQL-like behavior with simpler local setup. The accessible preview was limited, but the title and description clearly position it as an argument for adopting PGLite in developer workflows.
HN Discussion: One commenter said the name unintentionally reinforces the idea that SQLite is a lesser database, and discussed pronunciation. A user who had shipped a project with PGLite praised the fast local setup but warned it can hide scaling concerns that surface later in production. Other replies asked about use from non-JavaScript languages and pointed to native-code alternatives like pgembed or Doltgres.
Vinyl Cache and Varnish Cache
Summary: The post tries to clarify the split after the former Varnish Cache FOSS project renamed itself to Vinyl Cache. It says the original maintainers, site content, mailing lists, release branches, and development processes continued under the Vinyl Cache name. The article notes the project moved from GitHub to a self-hosted Forgejo instance, while old GitHub repos were archived and pointed at the new home. It describes the new Varnish Cache under Varnish Software as a downstream fork with separate governance, maintainers, and code direction.
HN Discussion: Several commenters say they had missed the rename entirely and now understand it less as a normal fork and more as a governance split involving the original author PHK. The MySQL versus MariaDB analogy resonated with readers trying to distinguish historical Varnish, Vinyl Cache, and the corporate fork. One technical side question asks why projects like Varnish and Redis historically resisted adding TLS support.
History & Science
Artemis II safely splashes down
Summary: NASA’s four-person Artemis II crew splashed down in the Pacific at 8:07 p.m. ET, ending a 10-day lunar flyby mission. Orion completed reentry, a planned six-minute communications blackout, parachute deployment, and a near-bullseye landing off San Diego. After splashdown, Navy divers secured the capsule, attached flotation gear and a front porch raft, and extracted the astronauts. The crew was hoisted by helicopter to USS John P. Murtha for initial medical checks.
HN Discussion: A top thread stressed how unusually risky Artemis is, citing a NASA OIG crew mortality target around 1 in 30 and celebrating the crew’s safe return. Multiple commenters described the mission as inspiring, hopeful, and one of the most uplifting public scientific events in years. One thread marveled that precise return timing and landing predictions come down to equations and orbital mechanics done right.
Chimpanzees in Uganda locked in eight-year ‘civil war’, say researchers
Summary: Researchers say the giant Ngogo chimpanzee community in Uganda split into Western and Central factions after years of cohesion. Since the split solidified in 2018, scientists recorded 24 lethal attacks, including 17 infants and at least seven adult males killed. The fracture appears to have started with growing avoidance after a 2015 dispute between the two subgroups. Likely destabilizers include deaths of key adults in 2014, an alpha-male change in 2015, and a 2017 respiratory epidemic that killed 25 chimps.
HN Discussion: Readers linked the findings to Richard Wrangham’s coalitionary killing theory and broader arguments that intergroup homicide may be evolutionarily selected. Several commenters focused on the 2017 respiratory epidemic, suggesting a 12.5% population shock could plausibly destabilize chimp social order. One subthread argued the story supports game-theory views of conflict over finite resources rather than any uniquely human ideology.
Italo Calvino: A traveller in a world of uncertainty
Summary: The article frames If on a Winter’s Night a Traveller as Calvino’s response to late-20th-century uncertainty and instability. It traces his early belief in a scientific, progress-driven view of history through postwar communism and Resistance experience. It argues The Path to the Spiders’ Nests replaced grand historical causes with accumulated personal motives and chance. It says Our Ancestors turned history into fable after Calvino’s break with communism following the 1956 Hungary invasion.
HN Discussion: Readers praised If on a Winter’s Night a Traveller for its deliberate sense of hidden structure and unresolved connections. A fan suggested the book lands better after reading more Calvino, and recommended Numbers in the Dark as an entry point. Several comments highlighted Invisible Cities as their first or favorite Calvino work.
The Bra-and-Girdle Maker That Fashioned the Impossible for NASA
Summary: The article profiles ILC, better known for bras and girdles, as the unlikely company that built Apollo spacesuits for NASA. It details extreme sewing tolerances, with suit seams held to less than 1/64 inch and mistakes often forcing full rework. It explains how seamstresses, gluers, and rubber dippers translated soft-goods craft into multilayer pressure-suit construction. It says ILC’s real advantage was collaboration between engineers and craft workers, including engineers learning sewing techniques directly.
HN Discussion: Commenters were struck by the craftsmanship and discipline required from the seamstresses and fabricators. One reader noted needle and pin scanners are still standard in consumer garment manufacturing as a final safety step. Another comment focused on the article’s documentation theme, saying traceability matters when debugging system failures or recalls.
Intel 486 CPU announced April 10, 1989
Summary: The article looks back at the 486 launch at Comdex in 1989, where Intel priced the chip at $950 in 1,000-unit quantities. It reviews period coverage from InfoWorld, Computerworld, and PC Magazine, which called the chip evolutionary, expensive, and ahead of immediate desktop needs. It highlights the 486’s on-die integration of 386-class CPU, math coprocessor, cache controller, and 8 KB cache, plus much higher performance. It notes early 486 systems were expected to cost $10,000 to $15,000 and only later became mainstream as Intel cut prices.
HN Discussion: Many comments are personal nostalgia about first 486 machines, often recalling $3,000 home PCs, skipped math coprocessors, and the jump from 386 or 8088 systems. Doom is repeatedly called the 486’s killer app, especially on DX2/66 systems with VLB graphics. Commenters also mention the 486 as an affordable path to serious software, BBS multitasking, Linux, X11, and UNIX-like development at home.
Security & Privacy
Installing every* Firefox extension
Summary: The author scraped Mozilla’s add-on APIs and category pages to enumerate roughly 84,235 Firefox extensions, about 49.3 GB total. They found the default paginated search API tops out far below the full catalog, then worked around it with alternate sorts and category queries. Installing the whole set exposed performance pathologies, including extensions.json being rewritten wholesale on frequent updates. The analysis highlights oddities such as giant extensions, permission-heavy slop, AI-generated spam, and prolific low-quality publishers.
HN Discussion: HN loved the sheer absurdity of the experiment and especially the video of a browser overloaded with extensions. A few commenters zeroed in on the discovery that extensions.json is fully rewritten, joking that a database would have helped. One thread highlighted the Iron Wallet phishing example and the author’s intervention against its writable backend.
JSON formatter Chrome plugin now closed and injecting adware
Summary: The repository announcement says the once-open-source JSON Formatter extension is no longer being developed in the open and is moving to a commercial closed-source model. The maintainer left the repo online for forks and published a ‘JSON Formatter Classic’ version as the final open-source local-only branch. The extension had long been known as a lightweight JSON viewer with tree view, dark mode, clickable URLs, and raw or parsed toggles. HN’s framing centers on recent injected checkout-page donation or affiliate behavior tied to Give Freely code, not on core formatter features.
HN Discussion: The submitter said they spotted a suspicious give-freely-root DOM element, then traced it to a recent closed-source update that injects checkout-page prompts and some geolocation behavior. A widely cited older HN quote from the maintainer promised never to sell out or add tracking, which made the reversal feel especially jarring. Commenters used the case to criticize auto-updating extension marketplaces and argued that browser extensions are a recurring rug-pull vector.
Quien – A better WHOIS lookup tool
Summary: Quien is a Go-based WHOIS and RDAP lookup tool with an interactive terminal UI. It combines WHOIS, DNS, mail, TLS, HTTP header, and tech-stack inspection in one interface. The tool prefers RDAP and falls back to WHOIS, with IANA referral support for server discovery. It supports JSON-oriented subcommands for scripting across dns, mail, tls, http, stack, and all-in-one output.
HN Discussion: The HN thread was very sparse, with one visible comment. The only substantive reaction was that the TUI looked polished.
AI & Tech Policy
AI assistance when contributing to the Linux kernel
Summary: The new kernel document says AI-assisted contributions must still follow the normal kernel development, coding style, and patch submission processes. It requires all contributed code to remain compatible with GPL-2.0-only and to use appropriate SPDX identifiers. AI agents are explicitly forbidden from adding Signed-off-by tags because only humans can certify the Developer Certificate of Origin. Human submitters are made responsible for reviewing AI-generated code, checking licensing, signing off themselves, and taking full responsibility.
HN Discussion: The dominant reaction was that the policy is refreshingly straightforward: AI use is allowed, but humans remain accountable. Several commenters approved of the explicit ban on AI-generated Signed-off-by tags and assumed lawyers likely reviewed the wording. A skeptical thread argued the policy does not solve review-bandwidth problems or shield the project from copyright infringement risk.
Sam Altman’s response to Molotov cocktail incident
Summary: Altman says someone threw a Molotov cocktail at his house at 3:45 a.m., but it bounced off and nobody was hurt. He argues AI can massively expand human capability, but says the economic and safety transition will require new policy and resilience. He says AI should be democratized so no small set of labs controls the future, and that democratic institutions must remain stronger than companies. The post includes self-criticism about conflict avoidance, mistakes around OpenAI governance, and the company’s shift from startup to major platform.
HN Discussion: Many commenters saw the post as image management or marketing rather than a serious response to the attack itself. Several readers connected it to a recent New Yorker profile and rejected the implication that critical journalism caused violence. A recurring theme was fear over AI-driven job loss and social destabilization, with commenters asking what labs or governments plan to do.
OpenClaw’s memory is unreliable, and you don’t know when it will break
Summary: The author says roughly a thousand OpenClaw deployments plus interviews with operators produced no durable ‘killer app’ beyond morning news briefings. The main critique is unreliable long-term memory and context loss, which makes autonomous execution risky because users cannot tell when important details were forgotten. It argues many viral OpenClaw success stories are hype, demos, or things already achievable with simpler LLM workflows and integrations. It warns that experimenting on personal machines is risky and recommends isolated VMs at minimum when connecting agents to email, calendar, or messaging tools.
HN Discussion: A recurring theme is that memory write logic, not just retrieval, is still brittle; commenters mention wiki-style memory and prompting hacks that only partly help. Some users agree OpenClaw is unreliable in practice, citing broken configs, context blowups, and time spent babysitting the system. Others still describe narrow useful cases such as cron-driven briefings, TODO capture, run management, proposal drafting, or GitHub chores, but mostly without depending on long memory.
Show HN: Eve – Managed OpenClaw for work
Summary: The live site exposes only a login page with the tagline ‘Your digital worker,’ so most concrete product details come from the Show HN text. According to the launch post, Eve runs agents inside isolated Linux sandboxes with 2 vCPUs, 4 GB RAM, 10 GB disk, headless Chromium, and real filesystem access. The author positions it as OpenClaw without self-hosting, focused on day-to-day work through a web UI and iMessage-based async tasking. Architecture claims include Claude Opus orchestration, routing to task-specific models, parallel sub-agents, persistent memory, and connectors to 1,000+ services.
HN Discussion: Some early testers say it is fun and useful for safe experimentation, citing tasks like dataset analysis, scheduled summaries, and mobile interactions via iMessage or SMS. Security and trust concerns dominate many comments, especially around logging into a little-known site with Google or linking Gmail and other personal tools. A commenter quotes the .new domain policy and questions whether the login-based flow fits Google Registry’s usage requirements.
Molotov cocktail is hurled at home of Sam Altman
Summary: The New York Times article reports that a 20-year-old man was arrested after throwing a Molotov cocktail at Sam Altman’s San Francisco home. Reporting says the device hit an exterior gate around 4:12 a.m., causing a fire but no injuries. The same suspect allegedly later threatened to burn down OpenAI’s headquarters, where police recognized and detained him. OpenAI confirmed the incident, said nobody was hurt, and added extra police and security presence around its Mission Bay office.
HN Discussion: The dominant thread is alarm over rising anti-AI hostility, with commenters saying the social temperature around AI and its executives has clearly increased. Some speculate that anger over job displacement and elite behavior will lead to more incidents, while others focus on the general normalization of political violence. A few commenters question the severity or optics of the attack, noting reports that only an outer gate burned and flirting with conspiracy-style readings.
System Administration
Investigating Split Locks on x86-64
Summary: The article explains split locks as atomic operations spanning two cache lines, forcing CPUs to fall back to disruptive bus-lock-like behavior. Using microbenchmarks, it shows split-lock latency ranging from bad to catastrophic depending on architecture, with especially severe penalties on Zen 2 and Zen 5. Arrow Lake and Alder Lake show different contention behavior from AMD chips, with some impact largely appearing beyond L2 while other platforms trash broader cache performance. The author also measures knock-on effects on unrelated workloads such as Geekbench photo filter and asset compression.
HN Discussion: Readers were puzzled that anyone would intentionally perform atomic operations on values that are not naturally aligned, especially across cache-line boundaries. One commenter recalled Linux split-lock detection crushing performance for some games when the mitigation first rolled out. Another said they were surprised detection is enabled outside multi-tenant environments, where noisy-neighbor protection matters more.
Geopolitics & War
Nowhere is safe
Summary: The article argues cheap mass drones have turned the earth’s surface into a contested kill zone for both military and civilian infrastructure. It says systems like Patriot, THAAD, Aegis, and Golden Dome are built for aircraft and missiles, not thousands of low-cost drones. It claims the U.S. underinvests in rapidly hardening or undergrounding assets such as command posts, fuel, munitions, repair facilities, and aircraft. It points to Gaza, Ukraine, and Gulf attacks as evidence that tunnels, covered movement corridors, and hardened shelters improve survivability.
HN Discussion: Several top comments push back on the framing and argue the real safety move is de-escalation and less bombing, not better protection for future wars. Readers debate whether underground systems are practical, with skeptics calling battlefield TBMs unrealistic and vulnerable to cheap drone attacks. Some commenters extend the idea to mobile underground rail systems and note civilian infrastructure like pipelines as obvious exposed targets.
Academic & Research
Simulating a 2D Quadcopter from Scratch
Summary: The post builds a planar quadcopter model from first principles, deriving Newton-Euler equations for horizontal motion, vertical motion, and rotation. It rewrites the system into first-order state-space form with state vector [y, z, phi, y_dot, z_dot, phi_dot] and inputs based on summed and differential thrust. The author implements the dynamics in Python with NumPy and simulates trajectories using Euler integration over 1,000 time steps. A zero-torque example shows straight vertical ascent, while a non-zero torque example shows the vehicle rotating until thrust can no longer counter gravity.
HN Discussion: The author says the post came out of six months spent replicating the ‘Champion-level drone racing using deep reinforcement learning’ paper. Commenters view it as a strong educational project and immediately ask whether the next step includes a PID stabilizer and gain tuning.
Business & Industry
The best seat in town
Summary: The article argues Paris’s self-cleaning JCDecaux toilets are essential street infrastructure, not a novelty, because they unlock mobility for commuters, seniors, families, and nightlife crowds. It says Paris installed 417 new toilets in 18 months before the Olympics, bringing the city total to 435 and boosting capacity with a cabin plus urinal layout. The author describes the cleaning cycle in detail, including automated bowl washing, floor flushing, and exterior water access. Paris is presented as a broader case study in coordinated street furniture, where toilets, bus shelters, kiosks, and ad infrastructure are clean, integrated, and sometimes vegetated.
HN Discussion: Readers broadly agree that public restrooms are a marker of civilized urban infrastructure, even if they must be hardened against vandalism. One commenter praises how much greener and more pedestrian-friendly central Paris has become under Anne Hidalgo compared with many U.S. downtowns. Others challenge the article’s framing, including the claim that JCDecaux ‘invented’ street furniture and the harsh comparison with LA’s bathroom program.
That is the morning scan for April 11, a set of stories where implementation details mattered more than slogans, and where the discussion kept circling back to ergonomics, safety, governance, and what actually breaks under load.