Pre-Release of Polars 2.0

86 komape 2026-09-03 07:01:21 UTC pola.rs Llama Share 8

Polars 2.0 is released as a release candidate, featuring improved defaults and API changes for better performance and memory usage. The major version bump allows Polars to remove outdated design decisions and introduce stricter behavior for error handling and data type conversions.

Muse Spark 1.3

563 bvaldivielso 2026-09-02 20:01:21 UTC developer.meta.com Full Share 379

Three schoolgirls in Kinsale pulled up a pea plant covered in warts (2016)

30 DamonHD 2026-09-03 08:01:21 UTC scienceblog.com Llama Share 9

Three Irish teenagers, Émer Hickey and her friends Ciara Judge and Sophie Healy-Thow, discovered that soil bacteria can help improve cereal crop growth by up to 74%. They tested this idea over three years with 13,000 seeds and proved that it works.

Gemini 3.8 Flash and 3.8 Flash Cyber

1004 bratao 2026-09-02 16:01:21 UTC blog.google Llama Share 567

Gemini 3.8 introduces two variants, Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, offering significant improvements in software engineering, agentic tasks, and critical reasoning at the same speed and low cost as version 3.7 Flash.

The Browser's Main Thread Is Expensive

11 kciter 2026-09-03 08:01:21 UTC kciter.so Llama Share 2

Frontend optimization involves managing the main thread, which handles tasks such as JavaScript execution, drawing, and event handling. To optimize performance, developers can use techniques like splitting tasks into smaller chunks (splitting), batching similar tasks together (batching), prioritizing and deferring non-essential work to free up resources for critical tasks (prioritizing and deferring).

LLMs and Self-Referentiality

34 mavdol04 2026-09-03 06:01:21 UTC scottaaronson.blog Llama Share 21

Here is a 2-sentence summary of the discussion: Many experts disagree on whether current large language models (LLMs) demonstrate self-awareness or consciousness, with some arguing that their ability to talk about themselves and meta-reasoning capabilities are evidence of emergent properties. However, others point out that LLMs lack true self-reference and introspection, and that their ...

The Computer Museum of America reclamation project

39 rbanffy 2026-09-03 05:01:21 UTC computer-museum.org Llama Share 12

The Computer Museum of America's collection has been preserved at San Diego State University since 2004, containing a vast array of computer artifacts and archival materials. The collection is being indexed and promoted to serve as a resource for understanding the impact of technology on society over the past 70 years.

Holden's Lightning Flight

150 ColinWright 2026-09-02 23:01:21 UTC en.wikipedia.org Llama Share 31

Walter "Taffy" Holden, a 39-year-old RAF engineer, accidentally flew an English Electric Lightning without proper training or equipment on July 22, 1966. He successfully landed the aircraft after aborting two initial attempts and was later cleared of any wrongdoing by an inquiry.

Three sites made 215,128 “best software” pages for AI. Perplexity cites them

406 jakobgreenfeld 2026-09-02 14:01:21 UTC trellner.com Llama Share 198

Two web-grounded models were asked to provide recommendations in 380 software categories, with 7,534 citations retrieved from their URLs. The majority of these citations point at domains ranked lower than #100,000 or not in the top million.

Google avoids a breakup of its ad tech business

369 donohoe 2026-09-02 19:01:21 UTC nytimes.com Llama Share 276

A federal judge ruled that Google must change its advertising technology practices to benefit competitors, but did not need to break up the business. The ruling prevents significant changes to Google's $4.1 trillion power over the internet despite two previous court rulings declaring it a monopolist.

Can I opt out of my input or output data being used for training?

438 teekert 2026-09-02 13:01:21 UTC help.mistral.ai Llama Share 194

Users can opt out of Mistral's model training programs at any time, with Vibe users opting in by default and Enterprise customers opted out by default. Users can disable the toggle to prevent data sharing for both Vibe and API services.

Higher Multipoles of the Cow

66 MrOrelliOReilly 2026-09-03 02:01:21 UTC arxiv.org Llama Share 20

The spherical cow approximation is extended to a full multipole expansion, allowing for computation of bovine potentials and interactions beyond spherical symmetry. This enables the treatment of physical processes suppressed by spherical symmetry, such as gravitational wave emission.

Reverse Engineering Unknown File Formats with ImHex

196 carlos-menezes 2026-09-02 21:01:21 UTC werwolv.net Full Share 34

Fable 5.1 World Modeling

238 surreal_ 2026-09-02 20:01:21 UTC github.com/philolabs Llama Share 70

This project creates browser-native reconstructions of real places using autonomous agent swarms and open data, with 5.1 worlds available to explore in San Francisco's Union Square area. The code is open-source under the MIT license, derived from OpenStreetMap and USGS data.

Launch HN: RonanRX (YC S26) – Personalized Peptides and GLP-1s

58 lloydarmbrust 2026-09-02 23:01:21 UTC news.ycombinator.com Llama Share 65

Lloyd, co-founder of RonanRx, is building a vertically integrated pharmaceutical company with software for prescribing, telehealth, and manufacturing. He aims to connect the disconnected pieces of pharmaceutical manufacturing through his experience in mask production and software development.

Show HN: OwnTime – a chess clock for your day's priorities

73 fmos 2026-09-03 04:01:21 UTC owntime.app Llama Share 37

OwnTime is an app that helps users prioritize tasks with daily time budgets, syncing between iPhone and Apple Watch, and providing real alarms when budgets are spent. The data is stored locally on devices, can be exported as a SQLite file for analysis, and remains outside the app's privacy policy once shared.

Biggest dark matter detector spots a single weird particle

291 randycupertino 2026-09-02 14:01:21 UTC science.org Full Share 100

Wendell Berry has died

185 Curiositry 2026-09-02 19:01:21 UTC nytimes.com Full Share 97

Please enable JS and disable any ad blocker

Aging brains blend memories together instead of just forgetting them

274 mdp2021 2026-09-02 13:01:21 UTC studyfinds.com Llama Share 117

Older adults' brains may blend memories together, causing mix-up errors, rather than simply losing memory accuracy. This blending effect is linked to stronger similarity in brain patterns between learning and recall phases.

Qantas Airbus A380 engine failure in 2010 (2023)

128 gumby 2026-09-02 19:01:21 UTC admiralcloudberg.medium.com Llama Share 80

A Qantas Airbus A380 experienced a catastrophic engine failure minutes after takeoff from Singapore, but the crew successfully landed the plane without incident. The failure was caused by a faulty oil pipe in one of the engines, which led to an internal fire and subsequent damage to multiple systems on board.

Engineering of the fastest WebAssembly interpreters

89 herobird 2026-09-02 22:01:21 UTC wasmi-labs.github.io Llama Share 8

Wasmi 2.0, a fast and feature-rich WebAssembly interpreter, has been released after eight months of focused work. It runs ~2.2x faster than Wasmi 1.0 in geometric mean across the wasmi-benchmarks suite on an Apple M2 Pro, with new knobs reducing binary artifact size and stable fuel metering added features.

Exit the Cave

256 akkartik 2026-09-02 15:01:21 UTC turtlespace.blog Llama Share 92

The author reflects on the "Cave" of private training and how it can lead to a false sense of progress, while missing out on real-world competition and growth. They share their personal experience as a wrestler, where they learned that true victory comes from embracing challenges and risks in public, rather than hiding in solitude.

Async Rust vs RTOS showdown (2022)

88 kooi 2026-09-02 19:01:21 UTC tweedegolf.nl Llama Share 43

This blog post compares Embassy/Rust and FreeRTOS/C on an STM32F446 microcontroller, evaluating their performance in terms of interrupt latency, program size, RAM usage, and ease of programming. The comparison focuses on a simple application that tracks a user button's state.

Altair Basic Interpreter Source Code (1975) [pdf]

59 Eridanus2 2026-09-02 21:01:21 UTC images.gatesnotes.com Full Share 33

You don't have permission to access "http://images.gatesnotes.com/12514eb8-7b51-008e-41a9-512542cf683b/34d561c8-cf5c-4e69-af47-3782ea11482e/Original-Microsoft-Source-Code.pdf" on this server.

A Selection of Los Alamos Rolodex Business Cards

166 1970-01-01 2026-09-02 15:01:21 UTC clui.org Full Share 44

A Selection of Los Alamos Rolodex Business Cards

We could save petabytes of cache storage with Zstandard and Pingora

110 torutofu 2026-09-02 18:01:21 UTC blog.cloudflare.com Llama Share 46

Cloudflare has developed a system called Cache Transcoding, which uses Zstandard compression to reduce storage and bandwidth costs by encoding eligible assets before writing them to disk. This approach preserves content integrity while offering significant savings in storage and bandwidth usage.

Poisson Disk Sampling

161 vismit2000 2026-09-02 14:01:21 UTC stripeacross.com Llama Share 19

A mathematician discusses Bridson's algorithm for generating Poisson disk distributions, which efficiently places points in a grid with a minimum distance between them. The algorithm has been improved upon by adding parental optimization and changing the distribution of distances to generate more random configurations.

Nango (YC W23) is hiring across eng, product and GTM (SF and remote)

0 2026-09-02 21:01:21 UTC nango.dev Full Share

From DevExp to high scale to API fragmentation, we work on hard problems.

Reasons robotics is hard

91 ddp26 2026-09-02 23:01:21 UTC secondthoughts.ai Llama Share 50

The development of broadly capable robots is racing along, but they still face significant technical challenges. Despite impressive demo videos, robots lack the flexibility, strength, and reliability needed for general-purpose work.

WebLLM: high-performance in-browser LLM inference engine

114 saikatsg 2026-09-02 15:01:21 UTC github.com/mlc-ai Llama Share 20

WebLLM is a high-performance in-browser LLM inference engine that leverages WebGPU for hardware acceleration, enabling powerful LLM operations directly within web browsers. It supports OpenAI API and various models, including custom models in MLC format.