HN Top 10 — June 03, 2026
Today’s Top 10 on Hacker News
1. Show HN: Edsger – A handwritten Clojure REPL for the reMarkable 2
⭐ 64 💬 14 👤 nathell 🔗 Discuss on HN
Edsger is a fully on-device Clojure REPL for the reMarkable 2 e-ink tablet. The system works by polling the Xochitl note-taking app’s filesystem for handwritten input, running OCR on the captured text, and sending it to a Go-based Clojure implementation (let-go) compiled for ARM. Results are rendered as images and injected back into the notebook via a custom XOVI plugin that hooks into Xochitl’s QML rendering pipeline. The project showcases the surprisingly open nature of the reMarkable platform and demonstrates creative use of handwriting recognition for interactive programming on an e-ink device.
2. 1-Click GitHub Token Stealing via a VSCode Bug
⭐ 441 💬 65 👤 ammar2 🔗 Discuss on HN
Just by clicking a link, it’s possible for an attacker to steal a GitHub token that can read and write to your repos, including private ones. !Image 2: CPython file opened in github.dev, a VSCode web interface This browser instance of VSCode is pretty powerful, you can view all the files in the repo (even if it’s a private one), you can send out pull requests and even…
3. Show HN: I reverse-engineered the world maps of Test Drive III (1990 DOS game)
⭐ 114 💬 31 👤 s-macke 🔗 Discuss on HN
A five-year reverse engineering project to extract and reconstruct the 3D world maps from Test Drive III: The Passion, a 1990 DOS racing game by Accolade. The author grew up exploring the game’s world rather than racing, and spent years on and off reverse-engineering the tile-based 3D format. Each tile and object is stored as small meshes with 16-bit signed vertex coordinates and polygon records. The project now includes interactive map visualizations built with modern web tools, bringing childhood nostalgia to life through decades of persistence and recent AI assistance.
4. Hacking your PC using your speaker without ever touching it
⭐ 5 💬 0 👤 xx_ns 🔗 Discuss on HN
What initially started as simply wanting to write a Linux tool for communicating with my speaker ended up with me discovering vulnerabilities which allow any attacker within a ~15M range of any Katana V2X to turn it into a covert spying tool and Rubber Ducky - all without ever having to pair with or physically touch the device. CTprotocol background As I explained in my…
5. Use your Nvidia GPU’s VRAM as swap space on Linux
⭐ 331 💬 88 👤 tanelpoder 🔗 Discuss on HN
nbd-vram is a tool that lets you use an NVIDIA GPU’s unused VRAM as high-priority swap space on Linux, designed for hybrid-graphics laptops with soldered RAM and no upgrade path. A small daemon allocates VRAM via the CUDA driver API and serves it as a block device using the Network Block Device (NBD) protocol. Tested on AMD + RTX 3070 Laptop setups: 7 GB of VRAM allocated for swap, combined with zram and SSD swap for ~46 GB total addressable memory — tripled from stock. Overflow order: RAM fills, VRAM absorbs via PCIe, zram compresses the rest, SSD only as last resort.
6. AI outperforms law professors in Stanford Law study
⭐ 265 💬 206 👤 berlianta 🔗 Discuss on HN
A groundbreaking study led by Stanford Law School Professor Julian Nyarko reveals that law professors overwhelmingly prefer AI-generated answers to student questions over responses written by their fellow instructors—a finding that could reshape how legal education is delivered. The study, titled “Law Professors Prefer AI Over Peer Answers,” was conducted with 16 law professors across U.S. law…
7. MAI-Code-1-Flash
⭐ 477 💬 213 👤 EvanZhouDev 🔗 Discuss on HN
MAI-Code-1-Flash is a new Microsoft coding model optimized for speed and efficiency in everyday developer workflows, rolling out to GitHub Copilot users in VS Code. Built end-to-end by Microsoft with appropriately licensed data, it features agentic coding trained directly with production GitHub Copilot harnesses, adaptive thinking that stays concise on simple tasks and reasons deeper on complex ones, and strong instruction-following. It outperforms Claude Haiku 4.5 across SWE-Bench benchmarks with up to 60% fewer tokens — delivering higher pass rates (including a +16-point lead on SWE-Bench Pro) at lower cost.
8. The Unreasonable Redundancy of Nature’s Protein Folds
⭐ 100 💬 28 👤 ray__ 🔗 Discuss on HN
Arda Goreci · May 20, 2026 Over the last few years, deep neural networks have made generative language modeling dramatically more powerful, giving us large language models. A similar leap happened for continuous modalities like images and videos. Recently, similar techniques have been applied to the generative modeling of biomolecules with great success.
9. It is an amazing time for programmers
⭐ 73 💬 39 👤 jlundberg 🔗 Discuss on HN
Published: 29 May, 2026 This is something I have been thinking a lot about but not yet written down in an article or blog post. It truly is a unique time to be alive as a person writing software. Many of the legends in the computer science space are still alive.
10. CT scans of BYD car parts
⭐ 394 💬 237 👤 viasfo 🔗 Discuss on HN
Lumafield CT-scanned four components from BYD’s EV lineup — the world’s largest electric vehicle manufacturer by volume: a lithium iron phosphate battery cell, a window switch panel from the BYD Tang, a Type 2 portable AC charger, and a key fob. The LFP battery cell reveals dual jellyroll electrode stacks with tight process control. The window switch panel shows a sparse PCB managing mirrors, windows, locks, and child safety on a LIN network. The portable charger reveals the handshake pins and single-phase power conductors in the Type 2 connector. A fascinating deep-dive into Chinese EV manufacturing quality from the inside out.