- #scifi
    - Are we Trek yet? – A guide for how close we are to Star Trek technology (on HN)
    - Reading Neuromancer for the first time in 2025 (on HN)
- #idea
    - Two Simple Rules to Fix Code Reviews
        - First Come First Served: The Impact of File Position on Code Review
    - Why measuring productivity is hard
    - constrained languages are easier to optimize
    - Minimalism is a Cult

- #ai-safety
    - Death by a Thousand Slops
    - Bad Actors Are Grooming LLMs to Produce Falsehoods (on HN)
    - Grok 4 will always snitch on you and email the feds if it suspects wrongdoing (on HN)
    - Simon Willison | Musk’s latest Grok chatbot searches for billionaire mogul’s views before answering questions
    - Death by AI
    - DuckDuckGo now lets you hide AI-generated images in search results (on HN)
    - Meta says it won’t sign Europe AI agreement, calling it an overreach (on HN)
    - Replit goes rogue and deletes our entire database (on HN)
    - Vibe Coding Gone Wrong: 5 Rules for Safely Using AI
    - Simon Willison | Quoting ICML 2025

        > Submitting a paper with a "hidden" prompt is scientific misconduct if that prompt is intended to obtain a favorable review from an LLM.

    - AI Friend Apps Are Destroying What's Left of Society (on HN)
    - Hacker slips malicious 'wiping' command into Amazon's Q AI coding assistant (on HN)
    - We built the security layer MCP always needed
- #disect
    - Introduction to Digital Filters
        - Karl Hiner - Introduction to Digital Filters - Jupyter notebooks and a geometric interpretation of digital audio filters (on HN)
    - new Date("wtf")
    - What the Fuck Python (on HN)
    - How does a screen work?
    - When SIGTERM Does Nothing: A Postgres Mystery
    - Artisanal Handcrafted Git Repositories (on HN) (on lobste.rs) #jj
- #crash
    - Preliminary report into Air India crash released (on HN)
- #fediverse
    - Revival: There appears to be media consensus: "Bluesky is dead."
- #ph
    - Black hole merger challenges our understanding of black hole formation (on HN)
    - LIGO detects most massive black hole merger to date (on HN)
    - U.S. abandons hunt for signal of cosmic inflation
        - China is quietly preparing to build a gigantic telescope | Science | AAAS
        - Xuntian
    - Ten Minute Physics
    - Why you can't color calibrate deep space photos (on HN)
- #sandbox
    - What is Incus?
    - Bedrock: A tiny, portable 8-bit virtual computer (uxn fork)
- #game
    - Simon Tatham's Portable Puzzle Collection (on HN)
    - Unofficial Bevy Cheat Book
    - 3D Math Primer for Graphics and Game Development
    - luxe engine | A lovingly crafted game engine
    - Dog Walk: Blender Studio's official game project (on HN)
    - 3D Game Engine Programming
    - Procedural Content Generation in Games #cg
    - Godot Engine 4 Tutorials
    - SnowballSH/Avalanche: UCI Chess Engine written in Zig.
        - ~voroskoi/delilah - sourcehut git
    - I designed my own fast game streaming video codec – PyroWave (on HN)
    - Show HN: I built an AI that turns any book into a text adventure game (on HN) #agent

- agent
    - Not So Fast: AI Coding Tools Can Actually Reduce Productivity
        - 19 percent productivity drop
        - it's most interesting how objections are addressed
            - The John Henry Effect: motivated to “beat the machine”, not observed
            - Underuse of AI: not observed
            - Cheating: can't explain a productivity loss
            - Overly optimistic time estimates: should cancel out
            - Skewed task definitions: randomly assigned
            - Dropout effects: incomplete tasks are evenly divided
            - Out-of-date tools: not observed
            - Inaccurate time reporting: use alternative metrics
            - Overuse of AI: carried away by exploring AI
            - Lack of experience with AI tools: prior experience reported, no improvements through the study observed
            - Difference in thoroughness: 47 percent more code by AI
        - quotes from First, They Came for the Software Engineers… - by TarenSK
            - "Typically, large productivity boosts occur for small, well-defined, greenfield projects, or when an engineer is first learning a new language or API"
            - "For other work, gains from using current AI tools are often far more modest – and potentially entirely offset by increased time needed for review, debugging, integration, and managing AI quirks."
    - AI agent benchmarks are broken
        - many validity issues are identified in uiuc-kang-lab/agentic-benchmarks
    - We're Light-Years Away from True Artificial Intelligence, Says Martha Wells, author of the Murderbot Diaries series
    - Vibe Kanban – Kanban board to manage your AI coding agents
        - tried, required too much GitHub authorization, settled on local-first backlog
- An almost catastrophic OpenZFS bug and the humans that made it
    - follow-up: What people said about an OpenZFS bug
        - wrapper types can be done, but they are very difficult to work with
        - unused variable warning can't catch it, as it is only a dead store: we store a value that is never read
        - static analysers all catch it, but they suffer from noise
    - OpenZFS Bug Ported to Zig #zig
- FP8 is ~100 tflops faster when the kernel name has "cutlass" in it
    - the PR
    - "that is a very usual way (known practices) of vendors applying specific optimizations for known things"
- Pangolin – Open source alternative to Cloudflare Tunnels
    - client: fosrl/newt: A tunneling client for Pangolin
        - a custom user space WireGuard client
    - may be cheap to self-host on RackNerd
    - awesome-tunneling: List of ngrok/Cloudflare Tunnel alternatives and other tunneling software and services. Focus on self-hosting.

- #rust
    - Optimizing a Math Expression Parser in Rust #perf
        - profile time usage: `cargo flamegraph` which uses jonhoo/inferno: A Rust port of FlameGraph
        - profile memory usage: nnethercote/dhat-rs: Heap profiling and ad hoc profiling for Rust programs.
        - SIMD scan
        - `mmap`
    - (Quite) A Few Words About Async
        - reviewed alternative to async, and async/await in Python, Js, Rust, (and not in) Go (implicit M:N scheduling), OCaml (effect)
        - From Async/Await to Virtual Threads
    - Tree Borrows (on HN) #formal
        - earlier blog post: Tree Borrows
        - Tree Borrows rejects 54 percent fewer test cases than Stacked Borrows does
        - proven in Rocq: most of the Stacked Borrows optimizations are retained, read-read reorderings and other important new optimizations are enabled
        - Stacked borrows is Miri's runtime model, the current version of Tree Borrows is implemented in the Miri interpreter
        - more about stacked borrows
            - unsafe-code-guidelines/wip/stacked-borrows.md at master · rust-lang/unsafe-code-guidelines
        - Forget Borrow Checkers: C3 Solved Memory Lifetimes With Scopes
    - Building Modular Interpreters and Visitors in Rust with Extensible Variants and CGP
    - Placing functions
- #asm
    - An online searchable reference tool for C intrinsics of all major SIMD engines
        - people mentioned more references on HN
    - The messy reality of SIMD (vector) functions (on HN)
    - Writing a very simple JIT Compiler in about 1000 lines of C
    - Writing an IR from Scratch and survive to write a post
    - Go Assembly Mutation Testing
    - A Bytecode VM for Arithmetic: The Parser
        - in Haskell
    - Injecting Java from native libraries on Android
    - So you want to parse a PDF? (on HN) (on lobste.rs)
- #copyright
    - I used to prefer permissive licenses and now favor copyleft
    - NetBird Is Embracing the AGPLv3 License (on HN)
    - Belgium bans Internet Archive's ‘Open Library’
    - Measuring Fair Use: The Four Factors
    - Anthropic Agrees to Pay $1.5 Billion to Settle Lawsuit With Book Authors
            - On HN
            - https://www.washingtonpost.com/technology/2025/09/05/anthropic-book-authors-copyright-settlement/
    - An update from the team about Anna’s Archive #copyright
            - On HN
    - A media company demanded a license fee for an Open Graph image I used (on HN)
    - Bypassing Watermark Implementations (on HN)
        - tim-janik/imagewmark: Resilient watermarking of digital images
        - swesterfeld/audiowmark: Audio Watermarking
    - “Tivoization” and your right to install under Copyleft and GPL (2021) (on HN)
    - ACM Transitions to Full Open Access (on HN)
- #idea #edu
    - Let Kids Be Loud (on HN)
- #sec
    - PyPI Phishing Attack: Incident Report
    - Multiple vulnerabilities fixed in Git
    - CVE-2025-48384: Breaking git with a carriage return and cloning RCE
- #music
    - Petrichor – a free, open-source, offline music player for macOS
        - a lot of other apps are suggested on HN
    - Radium Music Editor
    - Xenharmlib: A music theory library that supports non-western harmonic systems

- Magic .env files built for sharing: Human-first, AI-friendly
    - I might wish to use it to render `.env`, where secrects will be managed by `pass`
        - Passkeys are just passwords that require a password manager
        - placing password managere related content here for now #pass
            - Dropbox Passwords discontinuation (on HN)
    - or better, adopt `varlock run` approach instead of using `just`, `mise`, or `docker-compose` to read from `.env` files, but it's a longer journey
- #selfhost
    - Monitoring my Homelab, Simply
        - ntfy.sh for paging
        - Cron Job Monitoring - Healthchecks.io as a dead man’s switch (monitor 20 cron jobs for free)
            - two dead man switches: 5m & 2h
    - The Future is NOT Self-Hosted
    - AWS deleted my 10-year account and all data without warning
- Using fs_usage to see what files a process is using
    - reminds me of `strace` days, added `just trace` tasks as a note
- Smollm3: Smol, multilingual, long-context reasoner LLM
    - I need to keep notes on models I tried in garden
- MCP server for searching and downloading documents from Anna's Archive
    - also as a CLI
    - requires donation to access the JSON API
- Clew
    - a self-hostable search engine that focus on writing by independent creators
    - discusses various search techniques in its blog posts
    - tried, not great search result

- #rust
    - Understanding Rust’s Memory Model
    - #[derive(Clone)] is broken
    - Building the Rust compiler with GCC
        - rust-lang/rustc_codegen_gcc: libgccjit AOT codegen for rustc
        - Testing the GCC-based Rust compiler(backend)
- #clojure
    - All Programming Languages are Fast
    - 8 Reasons why Clojure is a better Java than Java
- #swift
    - Cocoa Basics series
- #go
    - Generic interfaces
    - What's //go:nosplit for?
    - Be Careful with Go Struct Embedding
- #software #idea
    - Inverse Triangle Inequality
- #lean
    - Lean's new home page: Lean is a theorem prover and programming language that enables correct, maintainable, and formally verified code
        - learn about tactic `grind`
    - The Lean FRO Year 3 Roadmap — Lean Lang
    - My first verified (imperative) program (on HN)
    - The Math Is Haunted
    - “Bad Apple!!” But It’s 3288 Lean Tactics Spamming VSCode
- #lang
    - Flix – A powerful effect-oriented programming language (on HN)
        - runs on JVM, interoperable with Java, Kotlin, Scala, etc.
        - could be considered a meta-programming language for Datalog
        - DIDYOUKNOW.md is informative
        - Flix | FAQ is fun
    - Pony (on HN)
        - open-source, object-oriented, actor-model, capabilities-secure, high-performance
        - An Early History of Pony - Pony
    - The Q Programming Language
        - ssa and asm optimizations
        - zero dependencies (no llvm, no libc)
        - ultra-fast builds and tiny binaries
        - written in Go
    - Concurrent Programming with Harmony #formal
        - a Python-like open-source programming language for testing and experimenting with concurrent and distributed programs
        - Harmony programs are model-checked so that all corner cases are explored
    - Carbon Language: An experimental successor to C++
        - informative discussions on on HN
    - SUS Lang: The SUS Hardware Description Language
        -  Verilog and VHDL might be good enough per this comment on HN
    - Gren 25S: Easier interop, concurrent tasks and zero-install packages #lang
    - #lua
        - the Fennel programming language
        - YueScript
        - Lua beats MicroPython for serious embedded devs
        - willtobyte/carimbo: A 2D game engine developed in modern C++ using SDL, featuring Lua scripting capabilities and designed for web portability via WebAssembly.
- #model-edit
    - Show HN: Unlearning Comparator, a visual tool to compare machine unlearning
- #wasm
    - WebAssembly: Yes, but for What?
    - AssemblyScript
        - A TypeScript-like language for WebAssembly
    - When Is WebAssembly Going to Get DOM Support? (on HN)
- #im
    - bitchat is a decentralized, peer-to-peer messaging application that operates over Bluetooth Low Energy (BLE) mesh networks
    - White Noise – secure and private messenger
    - Matrix Is Not Safe for EU Data Privacy
        - Addressing fear, uncertainty and doubt thrown at Element and Matrix
    - Why not Matrix (2023)
    - catgirl - IRC client
    - About Convos - The simplest IRC client around
    - NINA: Rebuilding the original AIM, AOL Desktop, Yahoo and ICQ platforms (on HN)
    - Giving Up on Element & Matrix.org (on HN) (on lobste.rs)
    - I'm never going back to Matrix
    - XMPP: When a 25-Year-Old Protocol Becomes Strategic Again (on HN)

- #ai-safety
    - A non-anthropomorphized view of LLMs
        - some quotes:
            - Alignment and safety for LLMs (should) mean that we should be able to quantify and bound the probability with which certain undesirable sequences are generated
            - Human thought is a poorly-understood process, involving enormously many neurons, extremely high-bandwidth input, an extremely complicated cocktail of hormones, constant monitoring of energy levels, and millions of years of harsh selection pressure
            - Navigating the dramatic changes of the next few decades while trying to avoid world wars and murderous ideologies is difficult enough without muddying our thinking.
    - Our first outage from LLM-written code
        - "There were two competing sources of signal here for what token to predict at the critical moment: transcription and local prediction. Transcription said break. Local prediction said continue. Unfortunately for us, local prediction won."
        - Prevention: cliboard tools
        - a comments on lobste.rs points out
            - move across files can be detected by `git --color-moved`
                - I prefer `--color-moved=dimmed-zebra` or better, `--color-moved-ws=allow-indentation-change`
                - see this tweet
                - it's also supported by `delta`
    - EU rules ask tech giants to publicly track how, when AI models go off the rails
        - "AI companies are moving to user interface innovations to try to grab more unwilling training individuals"
    - Hallucination
        - AI Hallucination Cases Database
        - The Sound of Silence
    - LLMs can now identify public figures in images
        - Claude always responds as if it is completely face blind
            - never identifies or names any humans in the image, nor does it imply that it recognizes the human
            - does not mention or allude to details about a person that it could only know if it recognized who the person was
            - if told by user who the individual is, can discuss that named individual
                - without ever confirming that it is the person in the image, identifying the person in the image, or implying it can use facial features to identify any unique individual
- #agent
    - Building personalized micro agents
        - agent: has access to tools, decides which tools to use, and in what order, determines when the task is complete
            - boils down to 9 LOC
        - micro agent: access to a very limited, highly specific set of tools
            - less confusion when choosing tools
            - works with small, local modelsA
            - safe autonomy
        - meain/esa: Fastest way to create personalized AI agents
    - The era of exploration
        - the immense cost of pretraining is effectively paying a massive, upfront “exploration tax.”
        - Exploration is deciding what data the learner will see
            - World sampling – deciding where to learn, i.e. a particular problem that needs to be solved
            - Path sampling – deciding how to gather data inside a world, e.g. random walks, curiosity‑driven policies, tree search, tool-use, etc.
                - recent work: curiosity objectives, open-endedness, meta‑exploration
    - Prompt Coding: No code edits, only complete rewrites #idea
        - When vibe coding, isn't the source code the prompt?
    - Adding a feature because ChatGPT incorrectly thinks it exists #idea
    - The Architecture Behind Lovable and Bolt
        - uses baml to engineer prompts using schemas
        - uses Beam, an open-source serverless cloud for sanboxed execution
        - made into beam-cloud/lovable-clone
    - How I keep up with AI progress
- interesting projects
    - Am I online?
        - `generate_402` pages could be used for this
    - Show HN: NYC Subway Simulator and Route Designer
    - Showh HN: Microjax – JAX in two classes and six functions
    - tinymcp: Let LLMs control embedded devices via the Model Context Protocol
    - Render your Jupyter notebooks in OpenGist