He Taught Himself to Code in a Colombian Gold Mine — Now He Builds AI Employees

He was laid off in COVID, down to his last penny, and taught himself to code on the night shift of a barbed-wire Colombian gold mine. Now he builds AI employees that run entire businesses. The lesson underneath it: while everyone chases the shiny agents, he mastered the fundamentals first — and that's the whole game.🎧 Close more by striving less. Rewire your nervous system with the free DreamState audios → https://www.dreamstatelivin.com/future-authoring (or comment STATE on Instagram and I'll send them).In this episode, Dan Gauerke breaks down the real path from zero to AI operator: how he went from make.com automations to building LeadLock (his voice-AI platform) and Big Dog (the AI employee that runs his go-to-market). We get into RAG, context engineering, the 3 stages of AI, why 80–90% of the work is planning, sub-agents and orchestrators, cron jobs and daemons, and where voice AI is headed in the next 12 months.You'll walk away knowing exactly where you actually are on the AI curve — and the next move to make instead of getting overwhelmed and quitting.⏱️ CHAPTERS00:00 From the gold mine to the frontier of AI00:30 20 on, 10 off: life inside an armored Colombian mine01:48 No TV, no distractions: teaching himself code on the night shift02:20 ChatGPT lands (Nov 2022) and everything shifts02:50 Contract ends — forced into entrepreneurship05:19 The path: make.com → n8n → agents → Claude Code06:34 First client Max Toro & building the whole infrastructure07:32 RAG explained: 400 videos into an AI that answers in your voice09:32 The voice-AI gap → building LeadLock10:20 Why emotion-aware speech-to-speech is the future12:07 What LeadLock actually does for agencies13:02 Voice agents for appointment setting: keep the scope small15:26 AI employees & the trust problem: automate the boring admin first18:02 What actually makes an agent an agent (tools, sub-agents, orchestrator)21:06 Context engineering: why most of the work is planning27:16 Beating overwhelm: think in systems, first principles29:34 The 3 stages of AI — and where you actually are32:29 Stage 3: commanding AI employees from your phone37:12 Meet Big Dog: building a go-to-market AI employee40:25 Rebuilding Big Dog inside the agentic OS (the 15-stage plan)44:42 How agents run for hours: /loop, /goal & daemons48:09 Where AI goes in the next 12 months50:03 Are they getting us addicted? Fable, pricing & the honest close🔗 GUEST — DAN GAUERKEInstagram: https://instagram.com/danielgauerke🛠️ TOOLS MENTIONEDClaude Code · Apify · ElevenLabs · Higgsfield · NotebookLM · S👇 CONNECT WITH DREAMSTATESubscribe for more conversations on AI, state control & closing without the grind.Book a call: https://www.dreamstatelivin.comInstagram: https://instagram.com/ethantryxPodcast: https://open.spotify.com/show/0qqtjisw9huL8HkxkXIRpy#AIagents #ClaudeCode #AIautomation #DreamState #highperforman---Spotify variant (shorter)Dan Gauerke taught himself to code on the night shift of a barbed-wire Colombian gold mine after being laid off in COVID. Now he builds AI employeesthat run entire businesses. In this episode he breaks down theI operator — his voice-AI platform LeadLock, Big Dog (hisgo-to-market AI employee), RAG, context engineering, the 3 stages of AI, and where voice AI is headed next. The through-line: master the fundamentalsbefore you chase the shiny agents.Close more by striving less — get the free DreamState audios: in.com/future-authoring (or comment STATE on IG).Guest: Dan Gauerke — https://instagram.com/danielgauerkeBook a call: https://www.dreamstatelivin.com