Took the plunge and deployed it on my own AWS infrastructure. For those who've been curious but haven't pulled the trigger yet - it's real and it works. My setup: - Isolated VPC with private subnet (no public IP) - Access via Telegram only - Zero exposed ports - SSM for admin - Encrypted storage, locked-down permissions First conversation hit and Claude responded through Telegram. Wild feeling having an agent just... waiting for me. Security was my main hesitation. Solved it by putting everything behind NAT with no inbound routes. The agent can reach out (APIs, Telegram) but nothing can reach in. If you're on the fence - the infrastructure side is more approachable than it looks. Happy to compare notes with anyone else who's deployed.

Quick workflow I stumbled into that's been surprisingly useful. Instead of manually analyzing a platform or community, I have Claude Code do it: • Navigate to the target site • Screenshot key pages • Analyze content patterns • Report back structured findings Did this for this Skool community yesterday. In about 10 minutes I had: • Community size and activity levels • All the content categories • Which post formats get engagement • Platform-specific formatting quirks (like code blocks not rendering) • Tone patterns from high-performing posts The browser automation caught things I would have missed. Like discovering the platform doesn't render markdown properly — would have built the wrong thing without that insight. The prompt pattern is simple: "Navigate to [URL], analyze the content structure and engagement patterns, tell me what resonates with this audience." Works for competitor analysis, community research, content strategy — anywhere you need to understand what actually works vs what you assume works. Anyone else using Claude's browser tools for research like this?

What happens when you have free desk space, A mac and PC who share screens, and then a desk arm turns up as if by magic? Now I have 3 monitors on the PC, 2 Monitors on the Mac Mini, and 2 Monitors for the MacBook Pro. Im sure I can find more somewhere. Now to get a blue tooth keyboard and mouse for the MacPro and im off to the races :)

Hey guys! I just got to sit down with Smitha Kullan and Vlad Kolesnikov from Google on their Agent Factory podcast—and the timing couldn't have been better. Google released Gemini 3 Pro literally hours before we recorded. So we got to dive deep into building production-ready AI agents, I did a live demo of my market research AI employee workflow, and we nerded out about everything from Gemini CLI to building entire AI workforces. If you've been wondering how to actually build AI agents that do real work for you, this episode is packed with actionable strategies. You can watch the full episode below and make sure to subscribe to their channel! Here's what we covered: ✅ Live demo: How I use Gemini CLI to build market research AI employees that search across entire markets in parallel ✅ My "AI-first" philosophy—how I've built AI employees to handle everything from customer research to email ghostwriting ✅ The SOP (Standard Operating Procedure) framework for creating AI employees that follow instructions at scale ✅ How I never write code anymore—I just review it while AI does the heavy lifting ✅ Real examples from my EMS startup showing AI doing market research across multiple cities simultaneously ✅ Why Gemini 3 Pro is crushing it for high-level reasoning and agentic operations Plus, Vlad showed off an insane AI video creation agent using Google's Agent Development Kit (ADK) that generates educational videos with a talking capybara. You have to see it to believe it. Whether you're just getting started with AI agents or you're looking to automate more of your business operations, this conversation is exactly what you need right now. P.S. Black Friday is here—and for the first time ever, Shipkit is offering our biggest discount of the year. If you want to start building production-ready AI applications with pre-built templates, Claude Code mastery modules, and everything you need to ship AI Chat, RAG, ADK agents, and more—now's the time. Skip the boilerplate, learn the workflows, and start shipping real AI applications in days instead of months.

My little weekend project that I called NeuroHelix. NeuroHelix is a fully automated AI research lab that wakes up before you do. Every morning a LaunchD-scheduled pipeline fires off 22 context-aware prompts via Gemini CLI, logs every run at the prompt level, and aggregates the answers into a polished intelligence report. The system isn’t just a batch of shell scripts—it’s a complete ops stack. Research outputs land in data/outputs/daily/, telemetry writes JSON ledgers plus human-readable logs, and the aggregators synthesize cross-domain themes, strategic recommendations, and execution summaries with zero manual touch. A Bash orchestrator enforces idempotency, git cleanliness, and pipeline locking, while maintenance flags let operators reprocess a single day or nuke the workspace with auditable trails. Publishing is equally modern. NeuroHelix exports structured payloads, source manifests, and vector-ready metadata, mirrors raw artifacts into an Astro site, and deploys to Cloudflare Pages with bundle metrics logged for every build. The front end embraces an IDE aesthetic: dual “Processed vs Source” views, repo-style navigation, command palette search that scans both final narratives and raw prompts, and accessibility baked into every interaction. Optional notifications—email or Discord—announce when the daily dashboard is ready, and bash-based smoke tests cover telemetry and prompt execution edge cases. In short, it’s an opinionated blueprint for AI-assisted research automation end to end. https://neurohelix.patchoutech.com/ and I made it public for anyone who wants to try : https://github.com/hopchouinard/NeuroHelix Notice : it is a personal fun project still very much in development but feel free to try it if you want.