Meta Muse Glimmer – open weights 30B local coding model

487 riordan 2026-08-10 10:58:07 UTC research.meta.ai Llama Share 258

Meta Superintelligence Labs introduces Muse Glimmer, a 30-billion-parameter model optimized for local agent workflows, enabling use anywhere, anytime. Muse Glimmer delivers strong performance on key agentic use cases and benchmarks, with optimized integrations and developer documentation available for building and running agents.

Squeak/Smalltalk 6.1 Release Notes

64 fniephaus 2026-08-10 12:58:07 UTC squeak.org Llama Share 19

Squeak 6.1 release notes highlight major changes, including a new tree browser, improved kernel infrastructure, and enhanced toolset and user interface.

50k Boat Names

60 jonathanmkeegan 2026-08-10 13:58:07 UTC beautifulpublicdata.com Llama Share 30

A dataset of over 50,000 boat names from NOAA's Marine Cadastre Vessel Traffic data reveals a wide range of creative and often humorous names, offering glimpses into the personalities and professions of boat owners. The names include references to TV shows, movies, songs, and mythological figures, as well as puns, jokes, and clever wordplay.

Docker Sandboxes – Disposable, isolated sandboxes for AI agents

397 etoxin 2026-08-10 06:58:07 UTC docker.com Llama Share 255

Docker Sandboxes provide isolated environments for AI agents like Claude Code and Gemini CLI to run freely and safely without affecting the host system. They offer a balance between speed and safety through permissive modes and microVM isolation.

Tail-call optimization in C is relatively recent

65 prakashqwerty 2026-08-10 11:58:07 UTC lwn.net Llama Share 32

The C calling convention has historically not supported tail calls, but GCC now supports them with a separate calling convention. This allows for improved performance in interpreters like Gforth and a toy Forth interpreter.

Parametron: 50s Japanese computer that uses neither transistors nor vacuum tubes

69 xeonmc 2026-08-10 10:58:07 UTC ethw.org Full Share 24

OpenAI's letter to Governor Abbott on responsible AI infrastructure in Texas

8 hackerBanana 2026-08-10 14:41:07 UTC openai.com Full Share 0

Over 181,000 AI meeting recordings left wide open in note taking app

194 colesantiago 2026-08-10 12:58:07 UTC bobdahacker.com Llama Share 65

A security researcher found a huge vulnerability in tl;dv, an AI meeting recording platform, that exposes millions of users' sensitive conversations and meeting metadata. Despite reporting the issue six months ago, the CTO has not responded, leaving users' data exposed and vulnerable to unauthorized access.

Mark Zuckerberg attacks 'closed' AI rivals as Meta returns to open models

21 root-parent 2026-08-10 14:09:07 UTC ft.com Full Share 11

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

Mistral Patent for "Code implemented tool calls"

72 theanonymousone 2026-08-10 13:58:07 UTC patentsgazette.uspto.gov Llama Share 73

A server receives a user request for tool calls and generates a code block using a large language model. The server resumes execution, substituting the first result for the pending tool call in the code block.

What Happened to HackerOne?

315 hipparchus 2026-08-10 02:58:07 UTC blog.teknogeek.io Llama Share 165

HackerOne has undergone significant changes since its founding, shifting from a community-driven platform to a corporate entity focused on sales and AI-powered solutions, raising concerns among hackers about data usage and platform stagnation. The company has faced criticism for its handling of AI technology, transparency, and community engagement, with some hackers questioning the use of ...

Defending my own brain against enshittification

53 mrmarket 2026-08-10 13:58:07 UTC mrmarket.lol Llama Share 36

The author found a new job that provides a welcome change from previous high-risk startups, with a more structured and normcore work environment. They've developed a morning routine and small daily rituals, such as making tea and eating Belvita cookies, to find joy and calm in their workdays.

Run Android ARM64 VR APKs on Apple Vision Pro

137 LorenDB 2026-08-10 03:58:07 UTC github.com/shinyquagsire23 Llama Share 40

Klepton-ld translates Android libraries to macOS compatible formats for Java-thin applications. It patches x18 usage and allows runtime patching of libraries on macOS.

Tail-Call Interpreters in Rust – Jimmy Ostler

63 amatheus 2026-08-10 07:58:07 UTC lordgoati.us Full Share 25

An Interesting Fourier Transform – 1/F Noise

86 q7m 2026-08-10 05:58:07 UTC dsprelated.com Llama Share 17

Power law functions in science and engineering have surprising properties, including the Fourier transform being a power law. This may be key to solving the 80-year mystery of 1/f noise, which has a power spectrum close to 1/f but an amplitude spectrum of 1/f1/2.

How Blackwing Pencils are Made [video]

57 NaOH 2026-08-10 07:58:07 UTC youtube.com Embed Share 25

Show HN: Voice driven murder mystery, Interview AI suspects with your voice

156 MrRowTheBoat 2026-08-10 03:58:07 UTC whodunnitai.com Full Share 63

Step into the interrogation room. Interview AI suspects with your own voice, catch their lies, and accuse the killer to their face. Solve the murder at Blackwood Manor — if you can.

Findphone: Locate a nearby Bluetooth device by signal strength

24 helsinkiandrew 2026-08-10 09:58:07 UTC github.com/ben-z Llama Share 12

The tool findphone helps locate nearby Apple devices within Bluetooth range by signal strength, with modes for surveying all devices and tracking one device by name. It requires macOS 13 or later and Bluetooth access, and can be used to sweep a room by ear with sound cues.

Taxi drivers rarely die of Alzheimer's

346 jader201 2026-08-09 15:58:07 UTC theconversation.com Llama Share 246

Two studies found that taxi and ambulance drivers have a lower risk of dying from Alzheimer's disease due to their continuous real-time navigation work. This association is linked to the hippocampus, a brain region that governs memory and spatial navigation, which is damaged early in Alzheimer's disease.

How I use LLMs to learn complex topics

746 laurentiurad 2026-08-09 19:58:07 UTC laurentiugabriel.github.io Llama Share 480

User uses generative AI for various tasks but finds LLM explanations too simplistic. They created a game, ChipTycoon, to learn chip building by asking AI to build knowledge, review accuracy, and create a low-poly animation.

The Future Is for Everyone – The Path to a Positive AI Future

15 insumanth 2026-08-10 10:58:07 UTC meta.com Full Share 4

Open QEC harness – greedy vs. GE, uniform vs. clustered k=4

4 Pachanoi 2026-08-10 12:58:07 UTC github.com/mrblakessinger-rgb Prefix Share 0

Digital evaluation / telemetry only — not hardware, not a production decoder, not an asymptotic threshold paper. Layer C framing: tensor-network model of bulk-boundary encoding with greedy recovery (HaPPY); optional matched-sample GF(2) GE ablation. Non-dominance: GE is not a recovery superset of greedy on this instance. Gap widens at depth-2 under the linear face-parameter model. k=4 stays ...

How We Pushed CDC into Postgres

121 craigkerstiens 2026-08-10 01:58:07 UTC snowflake.com Llama Share 23

Snowflake's data mirroring feature provides highly resilient data replication from Postgres to Snowflake with low cost and low lag. It uses push-based change data capture and transactional boundaries to simplify replication and unify operational and analytical workloads.

Ask HN: What are you working on? (August 2026)

285 david927 2026-08-09 17:58:07 UTC news.ycombinator.com Llama Share 986

The user has been working on various projects including an offline meeting document generator, a software house, a PlayStation bare-metal programming tutorial repo, a free and open source Chrome extension, and several other tools and platforms such as a music player, a calorie tracker, and a screen recording tool. They are also working on analytics for AI chat and search, a language with AI, ...

ATProto for Distributed Systems Engineers

118 LelouBil 2026-08-10 00:58:07 UTC atproto.com Llama Share 25

The AT Protocol is a decentralized backend system that allows applications to share state and interconnect, using a stream processing architecture with a shared data model and event logs to ensure scalability and reliability. The system is designed to be open and convenient, using standard practices for high-scale backends and peer-to-peer techniques to create a global network of applications ...

An alias-based formulation of the borrow checker (2018)

23 parksb 2026-08-10 06:58:07 UTC smallcultfollowing.com Llama Share 1

A new approach to Rust's borrow checker has been proposed, which views regions as sets of loans rather than program points, and uses Datalog rules to analyze the relationships between regions and loans. The analysis computes subset relationships between regions and propagates them across control-flow edges, and also tracks which regions depend on which loans to detect borrow check errors.

Cool URIs Don't Change (1998)

272 Klaster_1 2026-08-09 14:58:07 UTC w3.org Full Share 63

Picophysics: Single file physics for games on platforms like N64, PSX, DC

78 klaussilveira 2026-08-09 23:58:07 UTC gitlab.com Full Share 23

GitLab.com

Everything you do is being recorded

375 ike_usawa 2026-08-09 14:58:07 UTC theatlantic.com Llama Share 326

A new device called Spectre I is being developed to prevent others from recording you, using a technology that resembles speech to jam AI-enabled recorders. However, experts warn that this may not be enough to defeat advanced AI algorithms, which could always find new ways to recover speech.

Because It's Not Fun Enough: why languages fail

78 jottinger 2026-08-10 11:58:07 UTC bytecode.news Llama Share 90

Programming languages rise and fall based on how well they serve the three aspects of a programmer's work: vocation, art, and job, with difficulty being a key factor in their success or failure. Languages that make any of these aspects unnecessarily difficult will not survive, but languages that balance them well, like C++ and Rust, can be immortal.