People: David Idea: Built a tool that lets you search a local collection of documents (example in this case is a collection from CTAHR) using natural language queries - no web searches needed, just Claude Code talking to a SQLite FTS5 index. Details: * The tool indexes local PDFs, DOCX, PPTX,
People: David Pickett Idea: Testing whether a 4GB NVIDIA Jetson Nano can act as an autonomous routing brain - classifying incoming queries with a local embedding model and deciding to answer locally or escalate to more powerful servers across four compute tiers. Details: * The Jetson Nano can run llama.cpp
People: David and Joe Idea: We built an automation that generates a weekly project status report by pulling activity from multiple data sources - no more chasing people down for updates every Friday. Details: * A scheduled Claude Cowork task runs every Friday at noon, starting from our master Excel sheet
Ran some queries and saw interesting output on the server and client component Next.js credentials. Thought it was funny on its constant self doubt and loop. The layout is updated. However, there's an issue - the layout is a server component but it's importing a
**People:** David **Idea:** Testing OpenClaw as a self-hosted agent runtime that connects LLM agents to Slack with multi-agent routing, tool permissions, and security controls baked in. **Details:** - OpenClaw runs in a docker container and gives you channel integrations, tool execution, policy controls, and session memory out of the box
Last week I wrote about two paths to vLLM on Apple Silicon - comparing vllm-metal and vllm-mlx as options for local inference. This week the picture changed. LM Studio shipped concurrent request support, and suddenly the simplest option became the most practical one. People David Idea We spent the week
People Donavan Idea Set up MCP servers for CRM and Google Drive to connect them directly to Claude Desktop - giving Claude real-time access to our deal pipeline, contacts, and documents without copy-pasting or context-switching. Details * MCP (Model Context Protocol) for connecting AI assistants to external tools and data sources.
People David Idea I tested both vllm-metal and vllm-mlx on the M3 Ultra Mac Studio to see which one gives better multi-user LLM throughput on a high-memory Apple Silicon machine. Details * Both projects showed up in late 2025 - vllm on Metal/MLX is very exciting because it should let
People David Idea Testing Nvidia's v1 LLM Router blueprint as a third approach to intelligent query routing - this time using policy-based task classification instead of semantic similarity or neural network intent matching. Details * Nvidia's v1 LLM Router blueprint (main branch) takes a three-step approach: apply
People: David Idea: Tested UIUC's LLMRouter framework as an alternative to LiteLLM's semantic routing - this one trains an actual neural network for intent classification and can run on hardware as small as a Raspberry Pi with 4GB of ram. Details: * Wanted to compare this against
People: David Idea: Testing semantic routing as a way to automatically send LLM requests to different models based on what the user is asking—part of a bigger vision for smart edge-to-hub-to-cloud routing on our Maui cluster. Details: * Our LiteLLM proxy already bundles multiple machines into one API endpoint, so
People: David Idea: Testing KIRA, Krafton's open-source project that lets you run a full Claude Code instance through Slack, on our local PMF hardware. Details: * Nicole mentioned wanting more Slack bot and automation options for the team * Found Krafton's KIRA project on GitHub which does exactly