Show HN: Getting GLM 5.2 running on my slow computer https://ift.tt/wUr2mso

Show HN: Getting GLM 5.2 running on my slow computer The capabilities and security I was getting from this LLM are similar to those I've gotten from models like Claude or GPT, and this really surprised me. But then I thought, "I wonder how it would work on a normal computer like mine," and above all, "I wonder if it would work without going into OOM on a computer like mine." So I started working with the help of agents to test this possibility. I started converting the model to int4, understanding MTP usage, and if possible implementing DSA for long context. How it responds in int4 and whether the quality is maintained or not. Until I got to the point, on my computer with 32GB of RAM, I was able to communicate with GLM 5.2 with times that, of course, aren't high in cold start, but even then, we're talking about 0.1 tok/s, but that wasn't important to me. The important thing was the journey to reach this goal and, above all, changing the perspective on the project. I wanted it to work at all costs, even slowly. So I created Colibrì, which was born from a very simple idea, to be honest, but tested in every way, where a 744B Mixture-of-Experts model activates only ~40B parameters per token—and only ~11 GB of those change from token to token (the routed experts). So: The dense part (attention, shared experts, embeddings—~17B params) stays resident in RAM at int4 (~9.9 GB); The 21,504 routed experts (75 MoE layers × 256 experts + the MTP head, ~19 MB each at int4) live on disk (~370 GB) and are streamed on demand, with a per-layer LRU cache, an optional pinned hot-store, and the OS page cache as a free L2. The engine is a single C file (c/glm.c, ~1,300 lines) plus small headers. No BLAS, no Python at runtime, no GPU.No GPU or serious hardware because I don't have that hardware so I can't test it on hardware that is more powerful than my computer.Colibrì is a one-person project, written and tested entirely on a 12-core laptop with 25 GB of RAM — the numbers above are the ceiling of what I can measure at home. Any feedback is welcome! Repo: https://ift.tt/bQ8Xi2v https://ift.tt/bQ8Xi2v July 9, 2026 at 01:35PM

Show HN: Codex Explorer, a local session manager for Codex CLI https://ift.tt/G6Vp92a

Show HN: Codex Explorer, a local session manager for Codex CLI https://ift.tt/zpRKMbi July 9, 2026 at 11:53PM

Show HN: Onboard-CLI, a LLM powered and AST-based tool to visualize codebase https://ift.tt/CNG9Kda

Show HN: Onboard-CLI, a LLM powered and AST-based tool to visualize codebase https://ift.tt/UEXyNsS July 9, 2026 at 01:39AM

Show HN: Skill-extractor turns coding agent transcripts into reusable skills https://ift.tt/JtLaQHI

Show HN: Skill-extractor turns coding agent transcripts into reusable skills https://ift.tt/HK7fx4M July 9, 2026 at 01:33AM

Show HN: REST - Living Without Burnout. A manifesto about sustainable discipline https://ift.tt/lW7iT84

Show HN: REST - Living Without Burnout. A manifesto about sustainable discipline Lately, I’ve been thinking a lot about what gives me energy and what slowly takes it away. Those thoughts eventually turned into a small manifesto I called REST. https://themanifesto.rest/ July 9, 2026 at 01:12AM

Show HN: Hnwork.app – UI for Who is hiring posts https://ift.tt/lALUxfZ

Show HN: Hnwork.app – UI for Who is hiring posts Hey HN, I built a UI on top of the "Who is hiring" posts. Take a look at https://hnwork.app ! One of the downsides of unstructured text posts is the readability due to it being free-form and having little to no format. While there are other tools that have been built over the years to make perusing Who is hiring posts easier, I took a try on making my own (I actually tried to build this at a YC hackathon a few years back, but got around to completing it recently). Features: - Text search and search filters - Original post text with call outs to important information - Removes posts that aren’t on topic (complaints, seeking work, vague or missing contact info) - Analytics - API In addition, job posters can create accounts to submit postings through the app. While I don’t expect posting to move over to this app, it’s what I envisioned what a Who is hiring thread would like as an app: - Structured postings with required fields (e.g., salary range required) - Job posters get notifications about comments on their posts - Job posters get verified through their email before posting (e.g., someone posting a Sony job has a Sony email address) - Companies with multiple job posters can coordinate postings and view past postings - Admins can audit and approve companies and posts Job seekers can also create an account to post comments or get access to a simple API but otherwise browsing doesn’t require any kind of signup/signin. I’m open to feedback: let me know if you’d like me to ingest more data from past months, something is missing or broken, or there’s a new feature you’d like to see. Thanks! https://hnwork.app/ July 9, 2026 at 01:30AM

Show HN: Fork – Let users build features on top of existing applications https://ift.tt/Qbe3KUN

Show HN: Fork – Let users build features on top of existing applications Our initial release is a Chrome extension that builds on top of Gmail and Google Calendar. We keep limited trace activity of coding sessions for 30 days to troubleshoot and improve our offering. No data (events, emails) from Google is ever logged or retained. 50 second demo: https://www.youtube.com/watch?v=JQ292bncO_c Chrome extension (free + no login): https://ift.tt/GBlcWA4... We have a lot to learn and build. Would love any and all feedback! Paul & Dalton https://withfork.co/ July 7, 2026 at 11:47PM

Show HN: Record, replay, and improve AI agents in production https://ift.tt/DgRMn3C

Show HN: Record, replay, and improve AI agents in production At the AI Engineering World's Fair a big part of the conversation was to nail the self improvement loop. Our take on this is to record state of the agent execution with a durable runtime, then allow users to replay from state checkpoints and run 'what-if' experiments. It's OSS and free to use. Would love some feedback from the community. https://ift.tt/dKjtxMD July 6, 2026 at 10:56PM

Show HN: Handoff – a verified context bridge between Claude Code sessions https://ift.tt/5JoXiBI

Show HN: Handoff – a verified context bridge between Claude Code sessions https://ift.tt/ozrv3eP July 5, 2026 at 10:48PM

Show HN: I built an encrypted BLE dongle for pasting stuff to air-gapped devices https://ift.tt/V8miX0g

Show HN: I built an encrypted BLE dongle for pasting stuff to air-gapped devices Definitely one of those "20 minute adventure gone wrong" projects where all I wanted initially was a quick wireless rubber ducky for bitlocker keys and the like and then I kept adding stuff like AES-256..... Currently working on adding WebAuthn/FIDO support because the hardware is already there and scope creep is a lifestyle at this point. Would love feedback, especially on the security side. Repo and PCB files are fully open source. https://ift.tt/tB3ZcD6 July 5, 2026 at 02:43AM

Show HN: Gemma 3 inference in pure C++ with Metal acceleration https://ift.tt/OLeSf1w

Show HN: Gemma 3 inference in pure C++ with Metal acceleration https://ift.tt/kHWniYq July 4, 2026 at 09:24PM

Show HN: Opbox – CRDT based sync for text files on disk https://ift.tt/TEKAXHI

Show HN: Opbox – CRDT based sync for text files on disk Hi! I’m one of the founders of s2.dev, and recently have been hacking on opbox, which is an open-source daemon that turns directories of text files (code, markdown, etc) into collaborative, multi-player workspaces. This started as a bit of an intellectual curiosity, to see if it was possible to do real-time sync at the filesystem level (i.e., in an editor-agnostic way). The idea is pretty simple: - Opbox workspaces are roughly analogous to git repositories (and can be used alongside existing git repos, to share live changes between commits) - When the opbox daemon is running in a workspace (ob start), it listens for local filesystem events within its directory (writes, deletes, new files), and translates them into operations (the titular “op”) on shadow CRDT documents (Yrs) corresponding to each text file (as well as one doc for the namespace as a whole, which handles paths) - These shadow CRDT docs are maintained in a workspace-local sqlite db (Turso) - The ops, which represent diffs on a corresponding CRDT document, are then appended to a durable stream (S2) that acts as a shared journal for all sync participants - Opbox also reads from that journal, receiving ops from other participants, which are then used to update the local documents, first in the db, then by materializing them into actual files on disk This has worked surprisingly well for sharing things like Obsidian graphs in real-time. It’s most helpful in cases where you want the ability to edit local files from arbitrary editors, but still collaborate live. The experience is best from editors where you can configure an aggressive autosave policy, and where edits to an open file are reflected in the editor in a timely way. To gain confidence in the correctness of the core opbox flows (particularly all the nuances around bidirectional sync) I invested in wiring up deterministic simulation testing using the turmoil library, which has been incredibly helpful (see the opbox-sim crate in the repo). https://www.opbox.dev/ July 4, 2026 at 12:26AM

Show HN: Auto-continue Claude Fable 5 the second your 5-hour limit lifts https://ift.tt/miNvqK6

Show HN: Auto-continue Claude Fable 5 the second your 5-hour limit lifts https://ift.tt/PXD371k July 4, 2026 at 01:05AM

Show HN: Dockside – I turned unused space around the macOS Dock into a workspace https://ift.tt/eUv4Vl8

Show HN: Dockside – I turned unused space around the macOS Dock into a workspace https://ift.tt/nXKj0c4 July 3, 2026 at 11:35PM