Make Humans Analog Again (bhave.sh) AI

The opinion piece argues that AI agents can make people more “analog” by boosting hands-on creation, movement, and communication rather than replacing human work. It describes examples of using agents for coding, diagramming, and implementing ideas, and argues that better engineering practices (refactoring, documentation, testing) help agents work faster. The author also frames software development skills like delegation and orchestration as new forms of management and emphasizes that AI’s capabilities have limits that humans must bridge.

The Reason VO₂ Max Declines with Age (gethealthspan.com)

The article explains that VO₂ max declines with age not only because the heart delivers less oxygen, but increasingly because muscles lose the ability to extract and use it efficiently. It quantifies that maximal cardiac output drops less than VO₂ max, shifting a larger share of the limitation to peripheral factors like mitochondrial changes, reduced capillaries, and sarcopenia. It argues these peripheral systems can still adapt—endurance training supports capillary growth, while HIIT/SIT can drive mitochondrial improvements efficiently—so long-term consistency in aerobic training matters.

Usenet Archives (usenetarchives.com)

UsenetArchives.com describes itself as a large free archive preserving hundreds of millions of Usenet posts from 1980 to the present, with search and browsing by author, subject, content, and group. The site highlights Usenet’s role in early internet discussion and notes notable contributions made there, while explaining that it is maintained by Jozef Jarosciak and supported via Patreon to cover infrastructure costs.

Winners of the 2026 Kokuyo Design Awards (spoon-tamago.com)

Japan’s Kokuyo Design Awards 2026 named Hiroki Kannari’s “Before Note” as Grand Prix, a customizable “pre-notebook” concept positioned between mass production and personalization. Other merit winners include Takashi Higashide’s “Gram” pen series focused on subtle weight changes, and Yuji Tsukamoto’s “Notebooks Identified by Edges,” which uses colored edges for quick organization while minimizing ink use. The article also lists several finalists spanning innovative notebook layouts, writing aids, and packaging ideas.

Rendering arbitrary-scale emojis using the Slug algorithm (leduyquang753.name.vn)

The post explains how to adapt Eric Lengyel’s Slug GPU algorithm to render vector color fonts—like emojis—at arbitrary scales. It describes a HarfBuzz GPU pipeline for single-color glyphs (using fragment-shader coverage of curve outlines), then extends the approach to COLRv0 flat layered colors and COLRv1’s render-tree features such as clipping, gradients, affine transforms, and blending. The author outlines how glyph outlines and (for gradients) gradient definitions and transforms can be encoded into GPU buffers so colors can be computed per pixel during rendering.

Case study: recovery of a corrupted 12 TB multi-device pool (github.com)

A GitHub issue documents a case study of recovering a severely corrupted 12 TB btrfs multi-device pool after a hard power cycle left core trees inconsistent. The author reports that `btrfs check --repair` entered an infinite 46,000+ commit loop with no progress, and final recovery required 14 custom C tools using the internal `btrfs-progs` API, with only ~7.2 MB of data loss from 4.59 TB. The post also proposes nine potential upstream changes and publishes a reference tool set for similar incidents.

The 1987 game "The Last Ninja" was 40 kilobytes (twitter.com)

A post highlights that the 1987 game “The Last Ninja” fits within just 40 kilobytes of storage, underscoring how much could be done with very limited space on early computer systems.

An open-source 240-antenna array to bounce signals off the Moon (moonrf.com)

MoonRF is an open-source communications project building phased-array hardware for amateur use, including a 240-antenna (100 cm) array designed for Earth–Moon–Earth “moon-bounce” experiments. The site outlines upcoming kits—starting with the QuadRF 4-antenna SDR tile—and describes how larger arrays can be scaled for higher beam-forming gain and EIRP, alongside other terrestrial RF imaging and satellite-link use cases. Operation is tied to Amateur Radio Technician+ licensing and includes export-control limits for radar-related functionality.

Media scraper Gallery-dl is moving to Codeberg after receiving a DMCA notice (github.com)

The maintainer of the media-scraping tool gallery-dl says they received a DMCA takedown notice from FAKKU, LLC covering multiple extractor files and asking for either their removal or rewriting repository history within a week. Rather than rewrite history, they said they considered other options such as removing specific extractors or filing legal responses, and ultimately migrated and mirrored the repository to Codeberg and GitLab. The discussion also includes debate over the scope and specificity of the alleged infringements and potential next steps under DMCA processes.

More Americans Are Breaking into the Upper Middle Class (wsj.com)

The article argues that an increasing share of Americans are moving into the upper-middle class, suggesting changes in income growth, employment, and household finances over time despite continuing economic pressures.

We replaced Node.js with Bun for 5x throughput (trigger.dev)

Trigger.dev reports that it replaced Node.js with Bun for its Firestarter warm-start connection broker and saw about a 5x throughput increase. The team first removed a CPU-heavy SQLite-based metadata lookup, then switched to Bun’s native HTTP server APIs and trimmed hot-path overhead (notably Zod validation, header conversion, and debug logging). After deploying, they also found and fixed a Bun-specific memory leak tied to pending Promise<Response> not being resolved on client disconnects, which they corrected by resolving the request with an HTTP 499 status.

Scientists mapped all the nerves of the clitoris for the first time (livescience.com)

Researchers used high-resolution synchrotron X-ray imaging to build a 3D micron-scale map of the clitoris’s nerves in two donated pelvises, tracing the main dorsal nerve and showing it branches into the glans and other areas. The non–peer-reviewed results, published as a preprint, could help surgeons avoid nerve damage and improve reconstructive procedures after female genital mutilation or other vulvar surgeries.

NMS Ceefax: Remember teletext? This is the same (nmsceefax.co.uk)

NMS Ceefax is an interactive re-creation of UK teletext, letting users enter three-digit page numbers to load themed pages like “Sport Headlines 302,” using on-screen controls or keyboard inputs.

How to Get Better at Guitar (jakeworth.com)

The article argues that improving at guitar comes less from reading tabs and more from listening closely and transcribing what you hear note by note. It recommends starting with easy songs, stopping the track at each note to write it down, then comparing your transcription to online tabs or videos and revising as needed. The writer says this builds both the ability to hear notes accurately and muscle memory, and suggests practicing the newly learned songs through a playlist.

We analyzed 5,480 hospital cost reports. Supply spending varies 3-7x (andrewrexroad.substack.com)

An analysis of 5,480 US hospitals’ CMS cost reports finds large, case-mix-adjusted variation in supply spending, with efficient hospitals using far less than peers (up to ~3–8x differences by bed size) and no clear link to quality or outcomes. The author estimates about $28–29B in “addressable” supply waste (rising in FY2024 to $177.6B total supply spending), attributing much of it to procurement incentives, operating-room preference bloat and implant price dispersion, and disposal of items based on labeled expiration dates. The piece also estimates several billion dollars per year in donated medical supplies flowing through nonprofits, arguing hospitals may be buying and discarding more than they use rather than efficiently purchasing supplies.

My university uses prompt injection to catch cheaters (varun.ch) AI

A first-year computer science course reportedly embeds hidden prompt text in nearly invisible font to detect students who paste cheating instructions into LLMs, and the author suggests a floormate discovered the technique after testing it.

LLMs can't justify their answers–this CLI forces them to (wheat.grainulation.com) AI

The article describes “wheat,” a CLI/framework that helps teams using Claude Code turn technical questions into structured decision briefs. It gathers evidence through research, prototype, and adversarial challenge steps, records findings as typed claims with evidence grades, and uses a multi-pass compiler to catch contradictions and block output until issues are resolved. The output is a shareable, self-contained recommendation with an audit trail, illustrated with an example GraphQL migration decision.

Riddle solved: Why was Roman concrete so durable? (news.mit.edu)

MIT and collaborators report that the durability of Roman concrete likely came from “hot mixing” that used reactive quicklime and produced lime-rich mineral inclusions (“lime clasts”). These clasts create a brittle, reactive calcium source that helps the material spontaneously self-heal as cracks form by reacting with water to recrystallize calcium carbonate and further strengthen the surrounding composite. Experiments with hot-mixed samples showed cracks sealed within weeks, unlike concrete made without quicklime.

Sheets Spreadsheets in Your Terminal (github.com)

Sheets is a terminal-based spreadsheet tool (written in Go) that provides a text UI for viewing, editing, and navigating CSV data. It supports reading from files or stdin, jumping to cells or ranges, and using keyboard-driven commands for search, selection, editing, and formula insertion. Installation is via Go tooling or prebuilt binaries, with the project released under the MIT license.

New Copilot for Windows 11 includes a full Microsoft Edge package, uses more RAM (windowslatest.com) AI

A new Copilot update for Windows 11 replaces the native app with a web-based “hybrid” version that ships with its own bundled Microsoft Edge/Chromium components. The installer is distributed via the Microsoft Store, but it downloads an installer rather than the full app directly. In tests, the updated Copilot uses significantly more memory—up to around 500MB in the background and about 1GB during use.

AI agents promise to 'run the business,' but who is liable if things go wrong? (theregister.com) AI

The Register examines how liability remains unclear when AI agents “run the business” and errors cascade through automated decisions like HR, finance, and supply chain processes. UK regulators stress that accountable responsibility still sits with the using firm and its responsible individuals, even if the technology is provided by a vendor. Lawyers and analysts say contracts may shift blame through warranties, testing, monitoring, and explainability—yet non-deterministic agent behavior makes it hard to promise (or assign) predictable outcomes, with negotiations focusing on safeguards and the limits of what vendors will accept.

Copilot is 'for entertainment purposes only', per Microsoft's terms of use (techcrunch.com) AI

Microsoft’s terms of use for Copilot say it’s intended for entertainment only and that users shouldn’t rely on its outputs for important advice, as it can make mistakes. The company said it plans to update older wording, which had been criticized online. The article notes that similar disclaimers are used by other AI providers such as OpenAI and xAI.

Show HN: Gemma Gem – AI model embedded in a browser – no API keys, no cloud (github.com) AI

Gemma Gem is a Chrome extension that runs Google’s Gemma 4 model entirely on-device in the browser using WebGPU. It avoids API keys or cloud calls and can use a simple agent loop to read page content, click and fill forms, run page JavaScript, and answer questions about the site you’re viewing.

US War Machine Is Built on Decades of Lies. The Assault on Iran Is No Exception (truthout.org)

The article argues that the Trump administration’s claim that an alleged bombing in Iran was “done by Iran” reflects a longer U.S. pattern of war-related misinformation. It traces how successive administrations—across Hiroshima and Nagasaki, Vietnam, Korea, Iraq, and Afghanistan—used myths, omissions, and dehumanizing narratives to justify wars and minimize accountability for civilian deaths. The piece links these practices to structural choices like limiting international scrutiny and resisting civilian-harm safeguards.

Iran's IRGC Publishes Satellite Imagery of OpenAI's $30B Stargate Datacenter (newclawtimes.com) AI

Iran’s IRGC released satellite imagery and a video targeting OpenAI’s planned $30B Stargate AI datacenter in Abu Dhabi, threatening “complete and utter annihilation.” The article frames this as an escalation from earlier, broader IRGC warnings toward specific identification of the facility, citing prior regional attacks affecting Oracle and AWS-related infrastructure. It argues the main risk for AI “agent builders” is disruption to the compute layer behind OpenAI APIs, increasing the importance of multi-provider resiliency.