Faiss is a C++ library with Python and numpy wrappers for efficient similarity search and clustering of dense vectors, developed primarily at Meta's Fundamental AI Research group. It handles billions of embeddings within a fixed memory budget by offering two distinct scaling strategies: compact quantization codes that compress representations to fit in RAM (trading precision for scale), and graph-based indexes like HNSW and NSG that layer structure over raw vectors for speed.
- The GPU path is a drop-in replacement; swapping `IndexFlatL2` for `GpuIndexFlatL2` handles memory copies automatically
- The only hard dependency is a BLAS implementation; CUDA, ROCm, and the Python interface are all optional
- It explicitly includes tooling for parameter tuning rather than hiding the trade-offs, and maintains a dedicated troubleshooting page
- The GPU path is a drop-in replacement; swapping `IndexFlatL2` for `GpuIndexFlatL2` handles memory copies automatically
- The only hard dependency is a BLAS implementation; CUDA, ROCm, and the Python interface are all optional
- It explicitly includes tooling for parameter tuning rather than hiding the trade-offs, and maintains a dedicated troubleshooting page
A repository of bug-fixed and patched software projects for DEC PDP-11 systems, targeting real hardware, SimH, and PiDP-11 emulators running RT-11 and RSX-11M-PLUS. The two main projects are FIG-FORTH v1.4, which corrects garbled terminal output and a corrupted screen file, and FOCAL-11 V11-01C, which fixes an infinite recursion crash in the FEXP function.
- FIG-FORTH originally marked dictionary names with a high bit that modern terminals rendered as a literal "?" character; the patch strips it before display without affecting the dictionary
- The FORTH.DAT file was distributed with a VARIABLE RMS record format, embedding 2-byte length headers every 512 bytes that appeared as stray characters in screens
- FOCAL-11's XFEXP computed e^x by recursively halving the exponent until it fell below -8, but zero's exponent never decreases, so the recursion ran the stack into memory; all other math functions already had a depth guard
- FOCAL is a BASIC-like language from the late 1960s; Lunar Lander was originally written in it
- FIG-FORTH originally marked dictionary names with a high bit that modern terminals rendered as a literal "?" character; the patch strips it before display without affecting the dictionary
- The FORTH.DAT file was distributed with a VARIABLE RMS record format, embedding 2-byte length headers every 512 bytes that appeared as stray characters in screens
- FOCAL-11's XFEXP computed e^x by recursively halving the exponent until it fell below -8, but zero's exponent never decreases, so the recursion ran the stack into memory; all other math functions already had a depth guard
- FOCAL is a BASIC-like language from the late 1960s; Lunar Lander was originally written in it
Rulin Shao writes about Context Language Models (CLMs), which manage their own context by treating it as a file the model can edit with Bash, replacing harness-defined compaction and retrieval rules. Zero-shot applications to existing models outperform state-of-the-art context management strategies across diverse benchmarks.
- Multiple agent contexts can coexist as files, extending the design to agent swarms and subagents.
- Suffix Cache Reuse was co-designed to cut server-side compute by 35% against standard SGLang at matched performance.
- A model-editable context creates a new channel through which injected or self-written instructions can persist across turns.
- Multiple agent contexts can coexist as files, extending the design to agent swarms and subagents.
- Suffix Cache Reuse was co-designed to cut server-side compute by 35% against standard SGLang at matched performance.
- A model-editable context creates a new channel through which injected or self-written instructions can persist across turns.
A collection of 20 copy-paste prompts designed to test local coding LLMs by asking them to generate complete, playable browser games in a single HTML file. Each challenge enforces the same constraints: no external assets, libraries, or network requests, offline playable, and a priority on polish and game feel over explanations. The prompts span genres from arcade racing and survival horror to stealth and colony management, and the document suggests specific prompts as stronger benchmark candidates for evaluating complex coding ability.
- The prompts intentionally stress-test gameplay state, input handling, rendering, collision, enemy AI, progression systems, UI, persistence, and visual polish all in one shot.
- Prompts like OUTBREAK, DUNGEON ZERO, and CYBER SURVIVOR are flagged as particularly good for benchmarking because they demand more than basic rendering and movement.
- All graphics must be drawn procedurally using Canvas and CSS; Web Audio is used for synthesized sound effects.
- The single-file constraint makes it easy to compare models because there is no build system, dependency installation, or asset pipeline hiding behind the output.
- The prompts intentionally stress-test gameplay state, input handling, rendering, collision, enemy AI, progression systems, UI, persistence, and visual polish all in one shot.
- Prompts like OUTBREAK, DUNGEON ZERO, and CYBER SURVIVOR are flagged as particularly good for benchmarking because they demand more than basic rendering and movement.
- All graphics must be drawn procedurally using Canvas and CSS; Web Audio is used for synthesized sound effects.
- The single-file constraint makes it easy to compare models because there is no build system, dependency installation, or asset pipeline hiding behind the output.
DAC is an open-source dashboard-as-code tool built by Bruin Data that allows users to define, validate, and serve interactive dashboards entirely from declarative YAML and dynamic TSX components. The platform features a built-in semantic layer for defining reusable metrics and dimensions, supports all major SQL databases, and is explicitly designed to be utilized by AI agents to construct standardized, reviewable data visualizations.
Jenny List writes about how an undocumented feature in certain ESP32 chips allows the raw I/Q stream to be tapped before the WiFi modem, effectively turning them into software-defined radio receivers. Discovered independently by Espargos and a Reddit user, the mode supports both 2.4 and 5 GHz bands but is limited to receive-only operation.
- Only works on ESP32 variants with earlier Tensilica hardware; Bluetooth-only and non-radio P4 chips won't work
- Transmit capability appears possible but remains undocumented, with the discoverer refusing to share due to "rule 3"
- The 80 Msps 32-bit/sample data rate creates bandwidth challenges for external transfer via SDIO or UART interfaces
- Only works on ESP32 variants with earlier Tensilica hardware; Bluetooth-only and non-radio P4 chips won't work
- Transmit capability appears possible but remains undocumented, with the discoverer refusing to share due to "rule 3"
- The 80 Msps 32-bit/sample data rate creates bandwidth challenges for external transfer via SDIO or UART interfaces
Clef is a 27B multimodal model developed by Cloudflare that transforms state data and schemas of typed questions into structured decisions. By accepting text, JSON, images, or video as input, it returns a probability for every allowed option of each question in a single forward pass, entirely eliminating free-form text generation and output parsing. The model is post-trained from Qwen/Qwen3.8-27B and features a joint schema head that routes evidence to each question and scores all options simultaneously. It is fully compatible with Jev and SystemOne APIs, making it well-suited for deterministic, structured decision-making tasks.
- Supports `choice`, `score`, and `noul` (true/false) question types
- A smaller, faster variant called Clef-Flash is also available
- Achieves strong benchmark results, particularly on BFCL (98.5%), ToolRet (69.2), and API-Bank (91.9%)
- Significantly faster than comparable models, with a median latency of 209.3 ms
- Released under the Apache-2.0 license
- Supports `choice`, `score`, and `noul` (true/false) question types
- A smaller, faster variant called Clef-Flash is also available
- Achieves strong benchmark results, particularly on BFCL (98.5%), ToolRet (69.2), and API-Bank (91.9%)
- Significantly faster than comparable models, with a median latency of 209.3 ms
- Released under the Apache-2.0 license
Shapeshift is a Next.js + React 19 component that turns a single text input into context-aware UI cards. As you type, the box morphs into event cards, checklists, timers, color pickers, bill splitters, converters, polls, and more based on detected intent. Intent classification is powered by TypeSafe Jev, which answers 14 typed questions in a single parallel call, while all value extraction (dates, amounts, units, math) is handled by deterministic code. The project works fully offline by default using a built-in keyword classifier and can optionally use the online Jev model with an API key.
- One Jev call answers 14 typed questions in parallel (speculative fan-out), covering card type plus signals like "is it a video call?" and "is it urgent?"
- A state machine with hysteresis prevents flickering: a card only changes when a challenger wins twice in a row or is very confident
- 19 card types supported, including trip planner, time zone converter, dice roller, countdown, and goal tracker
- Built with TypeScript (strict), Tailwind CSS v4, shadcn/ui, Motion, chrono-node, and zod
- Live demo at shapeshiftui.vercel.app
- One Jev call answers 14 typed questions in parallel (speculative fan-out), covering card type plus signals like "is it a video call?" and "is it urgent?"
- A state machine with hysteresis prevents flickering: a card only changes when a challenger wins twice in a row or is very confident
- 19 card types supported, including trip planner, time zone converter, dice roller, countdown, and goal tracker
- Built with TypeScript (strict), Tailwind CSS v4, shadcn/ui, Motion, chrono-node, and zod
- Live demo at shapeshiftui.vercel.app
Giorgio Mendoza writes that Makerfabs has released the MaTouch ESP32-S3 7.5-inch Quad-Color E-Ink board, pairing an 800×480 four-color electrophoretic display with an ESP32-S3R8 controller, a four-microphone array, a six-axis IMU, and microSD storage for low-power displays and voice-controlled interfaces.
- Priced at $59.80, with an optional OV3660 3-megapixel camera adding $4.80
- The E Ink panel uses a JD79665 controller with a full-screen refresh time of about 20 seconds
- Software support includes ESP-IDF 5.4.2 and Arduino, with published examples for camera capture and voice-activated image browsing
- The board supports 2.4GHz Wi-Fi and Bluetooth 5 LE, with 8MB PSRAM and 16MB SPI flash
- Priced at $59.80, with an optional OV3660 3-megapixel camera adding $4.80
- The E Ink panel uses a JD79665 controller with a full-screen refresh time of about 20 seconds
- Software support includes ESP-IDF 5.4.2 and Arduino, with published examples for camera capture and voice-activated image browsing
- The board supports 2.4GHz Wi-Fi and Bluetooth 5 LE, with 8MB PSRAM and 16MB SPI flash
Thomas Claburn writes that the Pi coding agent hit its 1.0 milestone with a notable reversal on Model Context Protocol (MCP) support, which its creator Mario Zechner previously dismissed as unnecessary. Earendil, the company that acquired Pi in April 2026, attributed the about-face to improvements in MCP and the developers' own changes that made integrating other capabilities easier. The release also introduced Pi Durable, a separate component for orchestrating long-running agentic applications that was split out to preserve Pi's minimalist design.
- Earendil was formed by Armin Ronacher and Colin Daymond Hanna.
- The MCP changes also enable easier use of the Jev decision model within Pi.
- Pi is described as the Flask of agent harnesses (minimal and extensible), contrasted with Django.
- New features include Codemode (a sandbox for tool calls), virtual model support, deferred tool loading, and cache warming for Anthropic models.
- Earendil was formed by Armin Ronacher and Colin Daymond Hanna.
- The MCP changes also enable easier use of the Jev decision model within Pi.
- Pi is described as the Flask of agent harnesses (minimal and extensible), contrasted with Django.
- New features include Codemode (a sandbox for tool calls), virtual model support, deferred tool loading, and cache warming for Anthropic models.
