Show HN: Yori – Isolating AI Logic into "Semantic Containers" (Docker for Code) https://bit.ly/4kzdMHP

Show HN: Yori – Isolating AI Logic into "Semantic Containers" (Docker for Code) Hi HN, I've been a developer for some time now, and like many of you, I've been frustrated by the "All-or-Nothing" problem with AI coding tools. You ask an AI to fix a bug or implement a function, and it rewrites the whole file. It changes your imports, renames your variables, or deletes comments it deems unnecessary. It’s like giving a junior developer (like me) root access to your production server just to change a config file. So, 29 days ago, I started building Yori to solve the trust problem. The Concept: Semantic Containers Yori introduces a syntax that acts like a firewall for AI. You define a $${ ... }$$ block inside a text file. Outside the block (The Host): Your manual code, architecture, and structure. The AI cannot touch this. Inside the block (The Container): You write natural language intent. The AI can only generate code here. Example: myutils.md ```cpp EXPORT: "myfile.cpp" // My manual architecture - AI cannot change this #include "utils.h" void process_data() { // Container: The AI is sandboxed here, but inherits the rest of the file as context $${ Sort the incoming data vector using quicksort. Filter out negative numbers. Print the result. }$$ } EXPORT: END ``` How it works: Yori is a C++ wrapper that parses these files. Whatever is within the EXPORT block and outside the containers ($${ }$$) will be copied as it is. When you run `yori myutils.md -make -series`, it sends the prompts to a local (Ollama) or cloud LLM, generates the syntax, fills the blocks, and compiles the result using your native toolchain (GCC/Clang/Python). If compilation fails, it feeds the error back to the LLM in a retry loop (self-healing). Why I think this matters: 1. Safety: You stop giving AI "root access" to your files. 2. Intent as Source: The prompt stays in the file. If you want to port your logic from C++ to Rust, you keep the prompts and just change the compile target. 3. Incremental Builds (to be added soon): Named containers allow for caching. If the prompt hasn't changed, you don't pay for an API call. It’s open source (MIT), C++17, and works locally. I’d love feedback on the "Semantic Container" concept. Is this the abstraction layer we've been missing for AI coding? Let me hear your ideas. Also, if you can't run yori.exe tell what went wrong and we see how to fix it. I opened a github issue on this. I am also working in making documentation for the project (github wiki). So expect that soon. GitHub: https://bit.ly/4qysa4w Thanks! February 13, 2026 at 05:17AM

Show HN: MicroGPT in 243 Lines – Demystifying the LLM Black Box https://bit.ly/4aMZhN8

Show HN: MicroGPT in 243 Lines – Demystifying the LLM Black Box The release of microgpt by Andrej Karpathy is a foundational moment for AI transparency. In exactly 243 lines of pure, dependency-free Python, Karpathy has implemented the complete GPT algorithm from scratch. As a PhD scholar investigating AI and Blockchain, I see this as the ultimate tool for moving beyond the "black box" narrative of Large Language Models (LLMs). The Architecture of Simplicity Unlike modern frameworks that hide complexity behind optimized CUDA kernels, microgpt exposes the raw mathematical machinery. The code implements: The Autograd Engine: A custom Value class that handles the recursive chain rule for backpropagation without any external libraries. GPT-2 Primitives: Atomic implementations of RMSNorm, Multi-head Attention, and MLP blocks, following the GPT-2 lineage with modernizations like ReLU. The Adam Optimizer: A pure Python version of the Adam optimizer, proving that the "magic" of training is just well-orchestrated calculus. The Shift to the Edge: Privacy, Latency, and Power For my doctoral research at Woxsen University, this codebase serves as a blueprint for the future of Edge AI. As we move away from centralized, massive server farms, the ability to run "atomic" LLMs directly on hardware is becoming a strategic necessity. Karpathy's implementation provides empirical clarity on how we can incorporate on-device MicroGPTs to solve three critical industry challenges: Better Latency: By eliminating the round-trip to the cloud, on-device models enable real-time inference. Understanding these 243 lines allows researchers to optimize the "atomic" core specifically for edge hardware constraints. Data Protection & Privacy: In a world where data is the new currency, processing information locally on the user's device ensures that sensitive inputs never leave the personal ecosystem, fundamentally aligning with modern data sovereignty standards. Mastering the Primitives: For Technical Product Managers, this project proves that "intelligence" doesn't require a dependency-heavy stack. We can now envision lightweight, specialized agents that are fast, private, and highly efficient. Karpathy’s work reminds us that to build the next generation of private, edge-native AI products, we must first master the fundamentals that fit on a single screen of code. The future is moving toward decentralized, on-device intelligence built on these very primitives. Link: https://bit.ly/3ODXJfM February 13, 2026 at 03:38AM

Show HN: WebExplorer – a tool for preview file in browser https://bit.ly/3ZzZk8N

Show HN: WebExplorer – a tool for preview file in browser https://bit.ly/3ODVglw February 13, 2026 at 03:10AM

Show HN: Double blind entropy using Drand for verifiably fair randomness https://bit.ly/461dPpO

Show HN: Double blind entropy using Drand for verifiably fair randomness The only way to get a trust-less random value is to have it distributed and time-locked three ways, player, server and a future-entropy. In the demo above, the moment you commit (Roll-Dice) a commit with the hash of a player secret is sent to the server and the server accepts that and sends back the hash of its secret back and the "future" drand round number at which the randomness will resolve. The future used in the demo is 10 secs When the reveal happens (after drand's particular round) all the secrets are revealed and the random number is generated using "player-seed:server-seed:drand-signature". All the verification is in Math, so truly trust-less, so: 1. Player-Seed should matches the player-hash committed 2. Server-Seed should matches the server-hash committed 3. Drand-Signature can is publicly not available at the time of commit and is available at the time of reveal. (Time-Locked) 4. Random number generated is deterministic after the event and unknown and unpredictably before the event. 5. No party can influence the final outcome, specially no "last-look" advantange for anyone. I think this should be used in all games, online lottery/gambling and other systems which want to be fair by design not by trust. https://bit.ly/4rP3SEv February 12, 2026 at 03:10AM

Show HN: CodeRLM – Tree-sitter-backed code indexing for LLM agents https://bit.ly/3Mt9THH

Show HN: CodeRLM – Tree-sitter-backed code indexing for LLM agents I've been building a tool that changes how LLM coding agents explore codebases, and I wanted to share it along with some early observations. Typically claude code globs directories, greps for patterns, and reads files with minimal guidance. It works in kind of the same way you'd learn to navigate a city by walking every street. You'll eventually build a mental map, but claude never does - at least not any that persists across different contexts. The Recursive Language Models paper from Zhang, Kraska, and Khattab at MIT CSAIL introduced a cleaner framing. Instead of cramming everything into context, the model gets a searchable environment. The model can then query just for what it needs and can drill deeper where needed. coderlm is my implementation of that idea for codebases. A Rust server indexes a project with tree-sitter, builds a symbol table with cross-references, and exposes an API. The agent queries for structure, symbols, implementations, callers, and grep results — getting back exactly the code it needs instead of scanning for it. The agent workflow looks like: 1. `init` — register the project, get the top-level structure 2. `structure` — drill into specific directories 3. `search` — find symbols by name across the codebase 4. `impl` — retrieve the exact source of a function or class 5. `callers` — find everything that calls a given symbol 6. `grep` — fall back to text search when you need it This replaces the glob/grep/read cycle with index-backed lookups. The server currently supports Rust, Python, TypeScript, JavaScript, and Go for symbol parsing, though all file types show up in the tree and are searchable via grep. It ships as a Claude Code plugin with hooks that guide the agent to use indexed lookups instead of native file tools, plus a Python CLI wrapper with zero dependencies. For anecdotal results, I ran the same prompt against a codebase to "explore and identify opportunities to clarify the existing structure". Using coderlm, claude was able to generate a plan in about 3 minutes. The coderlm enabled instance found a genuine bug (duplicated code with identical names), orphaned code for cleanup, mismatched naming conventions crossing module boundaries, and overlapping vocabulary. These are all semantic issues which clearly benefit from the tree-sitter centric approach. Using the native tools, claude was able to identify various file clutter in the root of the project, out of date references, and a migration timestamp collision. These findings are more consistent with methodical walks of the filesystem and took about 8 minutes to produce. The indexed approach did better at catching semantic issues than native tools and had a key benefit in being faster to resolve. I've spent some effort to streamline the installation process, but it isn't turnkey yet. You'll need the rust toolchain to build the server which runs as a separate process. Installing the plugin from a claude marketplace is possible, but the skill isn't being added to your .claude yet so there are some manual steps to just getting to a point where claude could use it. Claude continues to demonstrate significant resistance to using CodeRLM in exploration tasks. Typically to use you will need to explicitly direct claude to use it. --- Repo: github.com/JaredStewart/coderlm Paper: Recursive Language Models https://bit.ly/4rG4RXf — Zhang, Kraska, Khattab (MIT CSAIL, 2025) Inspired by: https://bit.ly/3MrxwAn https://bit.ly/4rOWKIf February 11, 2026 at 02:10PM

Show HN: Agent framework that generates its own topology and evolves at runtime https://bit.ly/4ky6Omu

Show HN: Agent framework that generates its own topology and evolves at runtime Hi HN, I’m Vincent from Aden. We spent 4 years building ERP automation for construction (PO/invoice reconciliation). We had real enterprise customers but hit a technical wall: Chatbots aren't for real work. Accountants don't want to chat; they want the ledger reconciled while they sleep. They want services, not tools. Existing agent frameworks (LangChain, AutoGPT) failed in production - brittle, looping, and unable to handle messy data. General Computer Use (GCU) frameworks were even worse. My reflections: 1. The "Toy App" Ceiling & GCU Trap Most frameworks assume synchronous sessions. If the tab closes, state is lost. You can't fit 2 weeks of asynchronous business state into an ephemeral chat session. The GCU hype (agents "looking" at screens) is skeuomorphic. It’s slow (screenshots), expensive (tokens), and fragile (UI changes = crash). It mimics human constraints rather than leveraging machine speed. Real automation should be headless. 2. Inversion of Control: OODA > DAGs Traditional DAGs are deterministic; if a step fails, the program crashes. In the AI era, the Goal is the law, not the Code. We use an OODA loop to manage stochastic behavior: - Observe: Exceptions are observations (FileNotFound = new state), not crashes. - Orient: Adjust strategy based on Memory and - Traits. - Decide: Generate new code at runtime. - Act: Execute. The topology shouldn't be hardcoded; it should emerge from the task's entropy. 3. Reliability: The "Synthetic" SLA You can't guarantee one inference ($k=1$) is correct, but you can guarantee a System of Inference ($k=n$) converges on correctness. Reliability is now a function of compute budget. By wrapping an 80% accurate model in a "Best-of-3" verification loop, we mathematically force the error rate down—trading Latency/Tokens for Certainty. 4. Biology & Psychology in Code "Hard Logic" can't solve "Soft Problems." We map cognition to architectural primitives: Homeostasis: Solving "Perseveration" (infinite loops) via a "Stress" metric. If an action fails 3x, "neuroplasticity" drops, forcing a strategy shift. Traits: Personality as a constraint. "High Conscientiousness" increases verification; "High Risk" executes DROP TABLE without asking. For the industry, we need engineers interested in the intersection of biology, psychology, and distributed systems to help us move beyond brittle scripts. It'd be great to have you roasting my codes and sharing feedback. Repo: https://bit.ly/4rjN60f https://bit.ly/4612RRa February 11, 2026 at 08:39PM

Show HN: Yan – Glitch Art Photo/Live Editor https://bit.ly/4rOHLhr

Show HN: Yan – Glitch Art Photo/Live Editor Everything evolves in digitality, and deconstructs in logic. Tired of filters that make everyone look like a glazed donut? Same. Yan is not another beauty app. It's a digital chaos engine that treats your pixels like they owe it money. We don't enhance photos — we interrogate them at the binary level until they confess their true nature. [What We Actually Do] • Luma Stretch: Grab your image by its light and shadow, then yeet it into oblivion. Speed lines included. • Pixel Sort: Let gravity do art. Pixels fall, colors cascade, Instagram influencers cry. • RGB Shift: That drunk 3D glasses effect, but on purpose. Your eyes will thank us. Or sue us. • Block Jitter: Ctrl+Z had a nightmare. This is what it dreamed. [Why Yan?] Because "vintage filter #47" is not a personality. Because glitch is not a bug — it's a feature. Because sometimes the most beautiful thing you can do to a photo is break it. Warning: Side effects may include artistic awakening, filter addiction withdrawal, and an uncontrollable urge to deconstruct everything. Your camera roll will never be boring again. https://bit.ly/46lFOkn February 11, 2026 at 05:19AM

Show HN: Model Training Memory Simulator https://bit.ly/4tuFsBI

Show HN: Model Training Memory Simulator https://bit.ly/46lDeLd February 8, 2026 at 10:39AM

Show HN: I vibecoded 177 tools for my own use (CalcBin) https://bit.ly/4tMlOS3

Show HN: I vibecoded 177 tools for my own use (CalcBin) Hey HN! I've been building random tools whenever I needed them over the past few months, and now I have 177 of them. Started because I was tired of sketchy converter sites with 10 ads, so I just... made my own. Some highlights for the dev crowd: Developer tools: - UUID Generator (v1/v4/v7, bulk generation): https://bit.ly/4qNOvLH - JWT Generator & Decoder: https://bit.ly/4aHg9ot - JSON Formatter/Validator: https://bit.ly/4aJfF1a - Cron Expression Generator (with natural language): https://bit.ly/3MeEZCZ - Base64 Encoder/Decoder: https://bit.ly/4ra9OI2 - Regex Tester: https://bit.ly/4reyteL - SVG Optimizer (SVGO-powered, client-side): https://bit.ly/4ttUvLP Fun ones: - Random Name Picker (spin wheel animation): https://bit.ly/45YUUvM - QR Code Generator: https://bit.ly/45UvIXq Everything runs client-side (Next.js + React), no ads, no tracking, works offline. Built it for myself but figured others might find it useful. Browse all tools: https://bit.ly/4aHy9z4 Tech: Next.js 14 App Router, TypeScript, Tailwind, Turborepo monorepo. All open to feedback! https://bit.ly/4tvp9o4 February 11, 2026 at 03:46AM

Show HN: ArtisanForge: Learn Laravel through a gamified RPG adventure https://bit.ly/4apgIls

Show HN: ArtisanForge: Learn Laravel through a gamified RPG adventure Hey HN, I built ArtisanForge, a free platform to learn PHP and Laravel through a medieval-fantasy RPG. Instead of traditional tutorials, you progress through kingdoms, solve coding exercises in a browser editor, earn XP, join guilds, and fight boss battles. Tech stack: Laravel 12, Livewire 3, Tailwind CSS, Alpine.js. Code execution runs sandboxed via php-wasm in the browser. What's in there: - 12 courses across 11 kingdoms (PHP basics to deployment) - 100+ interactive exercises with real-time code validation using AST analysis - AI companion (Pip the Owlox) that uses Socratic method – never gives direct answers - Full gamification: XP, levels, streaks, achievements, guilds, leaderboard - Multilingual (EN/FR/NL) The idea came from seeing too many beginners drop off traditional courses. Wrapping concepts in quests and progression mechanics keeps motivation high without dumbing down the content. Everything is free, no paywall, no premium tier. Feedback welcome – especially from Laravel devs and educators. https://bit.ly/3O6FZtr February 8, 2026 at 08:15AM

Show HN: I built a cloud hosting for OpenClaw https://bit.ly/4r4v5CX

Show HN: I built a cloud hosting for OpenClaw Yet another OpenClaw wrapper. But I really enjoyed the techy part of this project. Especially server provisionings in the background. https://bit.ly/4a6nhe2 February 9, 2026 at 11:39PM

Show HN: Reef – Bash compatibility layer for Fish shell, written in Rust https://bit.ly/4aGPJDq

Show HN: Reef – Bash compatibility layer for Fish shell, written in Rust Fish is the fastest, friendliest interactive shell, but it can't run bash syntax, which has kept it niche for 20 years. Reef fixes this with a three-tier approach: fish function wrappers for common keywords (export, unset, source), a Rust-powered AST translator using conch-parser for structural syntax (for/do/done, if/then/fi, $()), and a bash passthrough with env capture for everything else. 251/251 bash constructs pass in the test suite. The slowest path (full bash passthrough) takes ~3ms. The binary is 1.18MB. The goal: install fish, install reef, never think about bash compatibility again. Your muscle memory, Stack Overflow commands, and tool configs all just work. https://bit.ly/3O6BYFp February 10, 2026 at 12:44AM

Show HN: Ported the 1999 game Bugdom to the browser and added a bunch of mods https://bit.ly/4klu5b4

Show HN: Ported the 1999 game Bugdom to the browser and added a bunch of mods I think the very first video game I ever played was Bugdom by Pangea Software, which came with the original iMac. There was also a shooter called Nanosaur, but my 7-year-old heart belonged to the more peaceable Bugdom, which featured a roly-poly named Rollie McFly needing to rescue ladybugs from evil fire ants and bees. Upon seeing the port to modern systems ( https://bit.ly/4tuDWzI ), I figured it should be able to run entirely in-browser nowadays, and also that AI coding tools "should" be able to do this entire project for me. I ended up spending perhaps 20 hours on it with Claude Code, but we got there. Once ported, I added a half-dozen mods that would have pleased my childhood self (like low-gravity mode and flying slugs & caterpillars mode), and a few that please my current self (like Dance Party mode). EDIT: Here are some mod/level combinations I recommend * https://bit.ly/3Msk7Iq... * https://bit.ly/4rJLFbr... * https://bit.ly/4a49Aw6... https://bit.ly/4rvsdPe February 9, 2026 at 04:07AM

Show HN: IsHumanCadence – Bot detection via keystroke dynamics (no CAPTCHAs) https://bit.ly/3Zp5E2U

Show HN: IsHumanCadence – Bot detection via keystroke dynamics (no CAPTCHAs) https://bit.ly/4agbc4I February 9, 2026 at 01:40AM

Show HN: A custom font that displays Cistercian numerals using ligatures https://bit.ly/4to51Er

Show HN: A custom font that displays Cistercian numerals using ligatures https://bit.ly/4twswLN February 8, 2026 at 11:39PM

Show HN: LocalGPT – A local-first AI assistant in Rust with persistent memory https://bit.ly/4a31tzU

Show HN: LocalGPT – A local-first AI assistant in Rust with persistent memory I built LocalGPT over 4 nights as a Rust reimagining of the OpenClaw assistant pattern (markdown-based persistent memory, autonomous heartbeat tasks, skills system). It compiles to a single ~27MB binary — no Node.js, Docker, or Python required. Key features: - Persistent memory via markdown files (MEMORY, HEARTBEAT, SOUL markdown files) — compatible with OpenClaw's format - Full-text search (SQLite FTS5) + semantic search (local embeddings, no API key needed) - Autonomous heartbeat runner that checks tasks on a configurable interval - CLI + web interface + desktop GUI - Multi-provider: Anthropic, OpenAI, Ollama etc - Apache 2.0 Install: `cargo install localgpt` I use it daily as a knowledge accumulator, research assistant, and autonomous task runner for my side projects. The memory compounds — every session makes the next one better. GitHub: https://bit.ly/3O2Lc5C Website: https://bit.ly/4a31uDY Would love feedback on the architecture or feature ideas. https://bit.ly/3O2Lc5C February 8, 2026 at 02:26AM

Show HN: More beautiful and usable Hacker News https://bit.ly/4r4GfaX

Show HN: More beautiful and usable Hacker News gives you keyboard navigation.. let me know what you think. https://twitter.com/shivamhwp/status/2020125417995436090 February 8, 2026 at 02:33AM

Show HN: Witnessd – Prove human authorship via hardware-bound jitter seals https://bit.ly/3ZsfPnk

Show HN: Witnessd – Prove human authorship via hardware-bound jitter seals https://bit.ly/4akLMmz February 7, 2026 at 11:40PM

Show HN: FastLog: 1.4 GB/s text file analyzer with AVX2 SIMD https://bit.ly/3OqOAXQ

Show HN: FastLog: 1.4 GB/s text file analyzer with AVX2 SIMD https://bit.ly/4qpevwB February 7, 2026 at 02:32AM

Show HN: A Kubernetes Operator to Validate Jupyter Notebooks in MLOps https://bit.ly/4rAX8tG

Show HN: A Kubernetes Operator to Validate Jupyter Notebooks in MLOps I built an open-source Kubernetes operator to automate the validation of Jupyter Notebooks in MLOps workflows. It's called the Jupyter Notebook Validator Operator and it's designed to catch issues with notebooks before they hit production. It runs notebooks in isolated pods and can validate them against deployed ML models on platforms like KServe, OpenShift AI, and vLLM. It also does regression testing by comparing notebook outputs against a "golden" version. The goal is to make notebooks more reliable and reproducible in production environments. It's built with Go and the Operator SDK. We're looking for contributors. There are opportunities to work on features like smarter error reporting, observability dashboards, and adding support for more platforms. GitHub: https://bit.ly/3ObO3Jf... https://bit.ly/46dId0r February 7, 2026 at 01:10AM