Personal Encyclopedias

159 jrmyphlmn 2026-03-26 07:01:41 UTC whoami.wiki Llama Share 40

The user created a personal encyclopedia called whoami.wiki using MediaWiki and a language model to organize and connect their family photos, videos, and data exports, revealing forgotten memories and strengthening relationships. The project is now open source, allowing others to create their own personal encyclopedias and run it on their machine with their own data.

Running Tesla Model 3's computer on my desk using parts from crashed cars

628 driesdep 2026-03-25 22:01:41 UTC bugs.xdavidhu.me Llama Share 195

The user successfully set up a Tesla Model 3 car computer on their desk by purchasing parts on eBay and using Tesla's electrical reference to connect the components. After several challenges, including a burned power controller chip, the user was able to get the system running and is now exploring the car's operating system and network interfaces.

Swift 6.3

33 ingve 2026-03-26 08:01:41 UTC swift.org Llama Share 8

Swift 6.3 introduces new features and improvements, including C interoperability, module selectors, and attributes for compiler optimizations. It also includes a preview of Swift Build, improvements to Swift Testing, and the first official release of the Swift SDK for Android.

Why Sora Failed: $15M/day inference cost vs. $2.1M lifetime revenue

6 vinodpandey7 2026-03-26 09:39:41 UTC revolutioninai.com Prefix Share 0

Everyone called Sora the future of video. They were wrong — and the numbers were screaming that for months before OpenAI finally pulled the plug on March 24, 2026. The prevailing narrative is that this is a sad story about a promising product killed too soon. That reading ignores the math entirely. Sora was burning through an estimated $15 million per day in inference costs at its peak, ...

What came after the 486?

26 jnord 2026-03-26 06:01:41 UTC dfarq.homeip.net Llama Share 17

Intel's CPUs didn't have brand names until the 1990s, using part numbers and clock speeds instead. The 486 was the last CPU with a number-based name, leading to the introduction of the Pentium in 1993.

The Last Contract: William T. Vollmann's Battle to Publish an Epic (2025)

5 benbreen 2026-03-26 09:01:41 UTC metropolitanreview.org Prefix Share 0

A few years ago, the novelist William T. Vollmann was diagnosed with colon cancer. The prognosis wasn’t great but he went ahead with the treatment. A length of intestine drawn out and snipped. It was awful but it worked. The cancer went into remission. William T. Vollmann spent “twelve or fifteen years” researching and writing a novel about the CIA called A Table for Fortune; as of this ...

ARC-AGI-3

393 lairv 2026-03-25 19:01:41 UTC arcprize.org Llama Share 261

ARC-AGI-3 is an interactive benchmark testing AI's ability to learn and adapt in novel environments. It measures intelligence across time, not just final answers, to bridge the gap between AI and human learning.

The Cassandra of 'The Machine'

7 Hooke 2026-03-26 09:01:41 UTC thenewatlantis.com Prefix Share 0

One day, Mrs. Pengelley came to London seeking the assistance of Hercule Poirot, Agatha Christie’s Belgian detective with the mustache, whose “little grey cells” assist him in solving mysteries. With a troubled look, she tells him that she fears she is being slowly poisoned. The doctor doesn’t see anything much the matter, she says. He attributes the stomach trouble to gastritis. She even ...

Obsolete Sounds

12 benbreen 2026-03-26 09:01:41 UTC citiesandmemory.com Llama Share 2

The Obsolete Sounds project reimagines lost sounds from the past and present, highlighting the need to preserve cultural heritage. It draws attention to disappearing soundscapes and encourages saving sounds before they're lost forever.

Earthquake scientists reveal how overplowing weakens soil at experimental farm

164 Brajeshwar 2026-03-25 21:01:41 UTC washington.edu Llama Share 69

Researchers used fiber optic cables to study the impact of tilling on soil moisture and found it disrupts soil's natural sponge-like quality. Tilling breaks soil channels, causing rain to pool and increase erosion and flood risk.

The truth that haunts the Ramones: 'They sold more T-shirts than records'

113 c420 2026-03-25 23:01:41 UTC english.elpais.com Llama Share 55

The Ramones released their self-titled debut album on April 23, 1976, which became one of the most influential albums in popular music history despite initial commercial failure. The album's logo, designed by Arturo Vega, became a highly successful promotional tool and is likely the best-selling T-shirt of all time.

Ashby (YC W19) Is Hiring Engineers Who Make Product Decisions

0 2026-03-26 07:01:41 UTC ashbyhq.com Full Share

Join a team that strives to do their best work every day.

Shell Tricks That Make Life Easier (and Save Your Sanity)

80 zdw 2026-03-26 01:01:41 UTC blog.hofstede.it Llama Share 28

The article discusses various tricks and shortcuts for Unix-like shells that can improve productivity and make the terminal more user-friendly. These tricks include line editing bindings, directory navigation, and command manipulation that can be used in most modern shells.

More precise elevation data for GraphHopper routing engine

44 karussell 2026-03-26 04:01:41 UTC graphhopper.com Llama Share 1

GraphHopper integrated new elevation data from mapterhorn for precise outdoor trip planning. This improved data reduces elevation errors near mountains and rivers, enhancing route accuracy and energy estimates for electric vehicles.

My DIY FPGA board can run Quake II

155 sznio 2026-03-25 22:01:41 UTC blog.mikhe.ch Llama Share 49

The user designed and built a custom FPGA board with a more advanced Efinix Ti60F256 and 1GB DDR3L memory, using a DDR3 Soft Controller Core and overcoming challenges with PCB layout and soldering BGA chips. The board features a custom SoC with a RISC-V processor core, DMA controller, and other components, and the user used SpinalHDL to generate Verilog code for the design.

90% of Claude-linked output going to GitHub repos w <2 stars

285 louiereederson 2026-03-25 19:01:41 UTC claudescode.dev Llama Share 173

Multiple contributors made various improvements to a project, including fixing build warnings, enhancing user interface, and adding new features such as custom caption bars and chat progress reporting. These improvements also included security hardening, rate limiting, and input validation, as well as updates to documentation and database connections.

The EU still wants to scan your private messages and photos

1159 MrBruh 2026-03-25 21:01:41 UTC fightchatcontrol.eu Full Share 307

🚨 The Conservatives (EPP) are attempting to force a new vote on Thursday (26th), seeking to reverse Parliament's NO on indiscriminate scanning. This is a direct attack on democracy and blatant disregard for your right to privacy. No means no. Take action now!

Two studies in compiler optimisations

82 hmpc 2026-03-26 01:01:41 UTC hmpcabral.com Llama Share 7

The user explores the implementation of LLVM optimisation passes using simple C++ examples to demonstrate how small source changes can trigger different paths in the compiler's internal processing with unexpected consequences. The user identifies various optimisation passes, including InstCombine, CodeGenPrepare, and DAGCombiner, and explains how they can be used to optimise code for ...

Supreme Court Sides with Cox in Copyright Fight over Pirated Music

348 oj2828 2026-03-25 16:01:41 UTC nytimes.com Full Share 267

Please enable JS and disable any ad blocker

Show HN: Robust LLM Extractor for Websites in TypeScript

42 andrew_zhong 2026-03-26 04:01:41 UTC github.com/lightfeed Llama Share 32

Lightfeed Extractor is a TypeScript library for robust web data extraction using LLMs and Playwright, allowing for browser automation, AI navigation, and structured data extraction. It supports various providers like OpenAI, Google Gemini, and Anthropic, and offers features like URL cleaning, image extraction, and safe sanitization of LLM outputs.

Show HN: Optio – Orchestrate AI coding agents in K8s to go from ticket to PR

53 jawiggins 2026-03-25 23:01:41 UTC github.com/jonwiggins Llama Share 31

Optio automates coding tasks by provisioning an isolated environment, running an AI agent, and merging pull requests. It handles tasks from GitHub Issues, Linear, or its web UI, and resumes the agent on failures or review feedback.

False claims in a widely-cited paper

276 qsi 2026-03-26 01:01:41 UTC statmodeling.stat.columbia.edu Llama Share 108

A business school professor is seeking advice on correcting a misreported study by Eccles, Ioannou, and Serafeim (2014) in Management Science, which has been widely cited but found to have methodological issues. The authors have acknowledged the mistake but refused to submit a correction, leaving the professor to consider submitting a comment or contacting Research Integrity Offices.

Government agencies buy commercial data about Americans in bulk

52 nuke-web3 2026-03-26 07:01:41 UTC npr.org Llama Share 20

Congress is set to reauthorize Section 702 of the Foreign Intelligence Surveillance Act, which may include closing the data broker loophole allowing government agencies to buy Americans' location data without warrants. Privacy advocates say this is the best chance to protect Americans' data from mass surveillance.

Thoughts on slowing the fuck down

856 jdkoeck 2026-03-25 15:01:41 UTC mariozechner.at Llama Share 387

The author warns that relying on coding agents can lead to brittle, complex, and unmaintainable codebases due to their inability to learn from errors and lack of global understanding. To avoid this, humans should slow down, use agents for non-critical tasks, and maintain agency over their codebases to ensure quality and maintainability.

Maxell MXCP-P100 – wireless cassette player

21 ChrisArchitect 2026-03-26 04:01:41 UTC maxell-usa.com Llama Share 9

This player combines 80's nostalgia with modern wireless tech, delivering high-fidelity sound from cassette collections. It offers up to 11 hours of Bluetooth playtime and fast USB-C charging.

Quantization from the Ground Up

252 samwho 2026-03-25 17:01:41 UTC ngrok.com Llama Share 47

Sam Rose discusses how to make large language models (LLMs) smaller and faster by reducing the number of parameters and using quantization, a process that reduces the precision of floating-point numbers. By applying quantization techniques, such as symmetric and asymmetric quantization, LLMs can be made 4x smaller and 2x faster with only a 5-10% loss in accuracy.

Apple randomly closes bug reports unless you "verify" the bug remains unfixed

404 zdw 2026-03-25 20:01:41 UTC lapcatsoftware.com Llama Share 227

You file bug reports with Apple Feedback Assistant due to a mix of addiction and the occasional fix of reported issues. However, Apple's lack of response and apparent disregard for users' time and effort is a major complaint.

Show HN: A plain-text cognitive architecture for Claude Code

98 marciopuga 2026-03-26 00:01:41 UTC lab.puga.com.br Llama Share 31

Cog is a persistent AI memory system that uses plain text files and Unix tools to store and maintain knowledge. It allows for transparent decision-making and rule-editing, with changes tracked in a git log.

Niche Museums

3 bookofjoe 2026-03-26 08:01:41 UTC niche-museums.com Prefix Share 0

John M. Mossman (1846-1912) was a bank vault engineer who operated out of New York City in the late 19th and early 20th centuries. His customers included the Stock Exchange and Bank of America, and he often found himself replacing older locks with new, upgraded versions. He kept hold of the locks that he replaced, which formed the basis for his 370-item lock collection. As a result, the ...

Jury finds Meta liable in case over child sexual exploitation on its platforms

390 billfor 2026-03-25 10:01:41 UTC cnn.com Llama Share 482

A New Mexico jury found Meta liable for failing to protect children from predators on its platforms, ordering the company to pay $375 million in damages. Meta plans to appeal the decision, which marks the first time the company has been held accountable in a jury trial for issues related to kids' safety.