Trending archive for Oct 25, 2026

deepseek-ai/DeepInfer

DeepSeek Opens Its Production Inference Engine — DeepInfer is the offline inference framework DeepSeek uses internally for large-scale serving, now open-sourced for DeepSeek-V4.1-Flash. It bundles optimized kernels (FlashMLA, DeepGEMM, DeepEP), MoE/CP parallelism, speculative decoding with confidence-scheduled verification, and batch-invariant/deterministic sampling—useful for teams deploying DeepSeek models at scale who need reproducible outputs.

DeepInfer: High-performance, large-scale inference framework for DeepSeek models

jyjblrd/single-propeller-drone

A drone that flies with just one spinning propeller — This project documents an under-actuated drone stabilized by a single propeller mounted atop a finned body, instead of the usual multi-rotor setup. It includes custom forks of ArduPilot and moteus for flight and motor control, plus Fusion 360/STEP files for the body and propeller designs, useful for hobbyists building unconventional drones.

Under-actuated drone with only one motor

mitmproxy/mitmproxy

The toolkit for intercepting, inspecting and scripting HTTPS traffic — mitmproxy offers three interfaces—console (mitmproxy), CLI (mitmdump, like tcpdump for HTTP), and web UI (mitmweb)—for intercepting SSL/TLS traffic across HTTP/1, HTTP/2, and WebSockets. It's built for penetration testers and developers who need to inspect, modify, or script HTTP traffic interactively.

An interactive TLS-capable intercepting HTTP proxy for penetration testers and software developers.

oosexgang/hakios-xnu-kernel

Booting Apple's actual XNU kernel and iOS 16.5 userspace on a Snapdragon Android phone — HackiOS runs Apple's open-source XNU kernel (from iOS 16.5) with the unmodified iOS userspace — launchd through SpringBoard — on a POCO X5 Pro 5G or QEMU, using a custom EFI loader, UEFI port, and a Linux Wi-Fi compatibility layer instead of Apple's closed drivers. No Apple files are bundled; IPSW content is fetched directly from Apple's CDN. A striking project for OS porting and reverse-engineering enthusiasts.

togethercomputer/HybridAgent

An LLM agent that switches between latent thinking and text actions — HybridAgent lets an LLM agent alternate between latent-space reasoning and text-based tool use within one generation pass, using a HybridLink module to map hidden states back to embeddings. It ships HybridSFT/HybridRL training stages, batch inference, and a tool to render latent thoughts as readable pseudo-text for interpretability.

php/php-src

The Source Behind PHP, Still Under Active Development — This is the official C source for the PHP interpreter, the engine behind countless websites and CMSs. The repo is useful mainly for contributors building PHP from source, testing new language features, or tracking engine internals rather than for typical application developers.

The PHP Interpreter

Zagitta/gpu-unlock

A Linux kernel module unlocks hidden BF16/FP16 speed on RTX 3000 GPUs — This kernel module combines the CMP unlocker's V67/PLM technique with UEFI-style pre-driver timing to double FP/BF16 FMA throughput on GA102/103/104/106 GeForce cards, without patching NVIDIA's driver each release. Benchmarks show BF16 GEMM gains of 25-80% depending on the card, useful for anyone running ML workloads on consumer RTX 3000 GPUs.

Your GeForce runs AI at half speed. This fixes it for 20/30/40 series cards. Up to 2× FP16 / BF16 / TF32 tensor throughput

receptron/mulmoclaude

A local-first AI assistant you grow with plain-file memories and no-code apps — MulmoClaude lets Claude build and maintain a personal wiki, schema-driven data, and small single-user apps (lists, trackers, flashcards) entirely as local files — no programming needed. Everything stays in your workspace under MIT, and you can reach the same assistant from your phone or a messaging app.

Nurture your own AI assistant on your own computer. Local-first and MIT: memories, data and apps stay as plain files in your workspace. Chat summons the right GUI — wiki, spreadsheet, chart, form, 3D. Build small apps for an audience of one, no programming required.

receptron/mulmoterminal

A Browser Grid for Juggling Several Coding Agents at Once — MulmoTerminal is a local, tmux-backed browser terminal that runs multiple coding-agent CLIs (Claude Code, Codex, Grok, Copilot CLI, Cursor CLI and more) side by side, colour-coding each cell as working, done, or waiting on you. Sessions survive reloads, isolate via git worktrees, and can push phone alerts when a turn finishes — built for people juggling several agents, not just one.

Run multiple Claude Code and Codex sessions in parallel — a browser terminal grid that shows which agent needs you. Local, tmux-backed, MIT.

CVEProject/cvelistV5

The Official CVE List Mirrored Daily in JSON 5 Format — This repo caches the complete official CVE List in CVE JSON 5 format, updated regularly with a deltaLog.json tracking recent record changes. It recently moved to Record Format 5.2.0, adding PackageURL (PURL) support for affected components, making it useful for anyone building vulnerability tooling or tracking CVE data changes over time.

CVE cache of the official CVE List in CVE JSON 5 format

electron/electron

The framework behind VS Code keeps powering desktop apps — Electron combines Node.js and Chromium so developers can ship cross-platform desktop apps using JavaScript, HTML and CSS. It powers VS Code and many other apps, with prebuilt binaries for macOS, Windows and Linux, including Apple Silicon and ARM support.

:electron: Build cross-platform desktop apps with JavaScript, HTML, and CSS

musoukun/skills

A grab-bag of drop-in skills for Claude Code and similar AI agents — This repo collects reusable skill folders for Claude Code-style AI agents, each with a SKILL.md you copy into your local skills directory. Included skills cover annotating screenshots for guides, DDD-style requirements modeling, generating narrated 3D explainer videos with Three.js, pattern/case classification, and turning architecture diagrams into animated MP4/GIFs.