V1 is live. I built an AI assistant that calls me on the phone. 78 seconds, no fluff, just the demo. YouTube: https://youtu.be/qeRwLXY1TCc I'll be in the comments today replying to everyone. If you're seeing this here first (founding members get the inside track), drop a thought when you watch. V2 is the full build walkthrough so you can run your own Beckett. That one will live here too with the code in a separate post.

This is the room where I build Beckett in public and steal good ideas from whatever you're building. Here's what's actually here: Build logs. Every week I post what I shipped, what broke, what I almost shipped and pulled. Specifics, not theory. The actual code, the architecture choices, the moments I picked one tool over another. It's all here. Teardowns. When something interesting lands in the AI agent space (new model, new framework, new technique), I run the experiment and post what I'd actually do with it. No "top 10" lists. Member builds. Post what you're working on, what's stuck, what worked. Other members weigh in. I weigh in. We figure it out together. Done-for-you. If you want me to build it instead of you, that conversation also happens here. Three things to do right now: 1. Read the first build log (pinned above) 2. Drop a comment so I know who's in the room 3. Tell me what you're building or stuck on. Specifics. I read every post. The thesis: nobody has agents running in production at scale yet. The people who figure it out fastest will own the next five years. Welcome to the lab. Tuni

Beckett is my personal AI engineer. It runs on a $5/mo VPS, answers my Telegram messages (text and voice), picks up the phone when I call (or calls me when something matters), and remembers every conversation we've ever had. It also ships code. Last month I asked it on a phone call to change my landing page copy. It pushed the commit before I hung up. The stack - Brain: Claude (Opus 4.7, with Sonnet fallback). Runs as the Claude Code CLI on the VPS. - Voice: Vapi for phone calls (custom assistant, ElevenLabs voice). Telegram bot for text and voice memos. - Memory: Supabase Postgres + pgvector. Hybrid search (dense embeddings + BM25 + RRF + cross-encoder reranker). Three chunking strategies routed by data type: Q/A pairs for conversations, agentic chunking for sessions, single-entry for facts. - Today: 476 chunks across 51 sessions, growing every day. - All-in cost: ~$2-5/month. Why this stack Most personal AI projects fail at memory. They use one embedding strategy across everything and pretend that's good enough. I tried that first and watched Beckett confidently "remember" things that never happened. The fix wasn't a better embedding model. It was three different chunking strategies routed by what type of data we're storing, and a reranker on top of hybrid search to catch the cases where dense embeddings get fooled by surface similarity. Most recent upgrade: adding a cross-encoder reranker on top of hybrid search. Cost: +200ms per query. Gain: +3.6% MRR, +7.3% NDCG@10 on a labeled benchmark built from real conversations. Worth it. What's next The current build is an idea vetting board: three agents (Research, IP, Critic) that run in parallel when I float a new project idea, and feed Beckett a synthesis before I commit time to anything. The whole point is making me less likely to start things that won't ship. After that: voice mode improvements and the multi-agent orchestration that lets Beckett decide what's worth doing on its own. Your turn