What happens to human agency when machines start planning, deciding, and acting for us?In this episode of Archaeology of Now, we explore the idea of preference cocoons: environments where algorithms continuously adapt to what we like, slowly removing friction, surprise, and challenge. Will AI not only take our job but also the ability to think? and the future role of humans in an agent-driven world. How AI systems shape our choices, how productivity has become a defining lens for human value, and what a post-productivity future might look like.This is the Archaeology of Now — a podcast by Ruslan and Reza, two AI enthusiasts working in tech and interested in how technology is reshaping the world we live in. We’re not chasing news. We dig into big, uncertain questions to understand how reality is changing — and invite you to think it through with us.0:00 Welcome to Tempelhofer Feld (Berlin)0:56 Episode agenda: purpose, autonomy, and human agency in an AI-built world2:08 “Preference cocoons”: from echo chambers to AI-shaped comfort loops4:41 Self-worth, intelligence, and “value”: what’s left for humans8:22 Creativity vs rules: systems, regulation9:54 Agents as assistants: prediction, preference-shaping, and the mirror effect10:55 AGI + economy: productivity gains don’t automatically end work13:35 If planning/execution is automated, what’s the human role?21:35 Humans as computers: what disappears, what evolves23:00 Thinking vs computing: Penrose, non-rational paradigms, poetry vs rationality26:59 Homo faber → “Homo non-faba”: the post-productivity human32:28 Free will as a social contract: Sapolsky, Searle, predictability of behavior37:05 Simulation/encoded thoughts39:06 The autonomy slider in dev: vibe coding, when to let autopilot run41:10 Orchestrating agents: coder/evaluator/orchestrator, PRD → tests → metrics47:29 Limits of autonomy today: minutes of reliability, beautiful nonsense49:11 Future of interfaces: generative UI, “phones as pixel receivers,” dreams of machine52:45 LLMs as mirrors: semantic space, training on past eras, language drift57:26 Memory as rewriting: recall changes memories; associations → facts

Can you shut off an intelligent robot — or would that be murder?

On a particularly cold day near the Berlin Wall, we met to explore this question.

This is the first episode of Archaeology of Now — a podcast by Ruslan and Reza, two AI enthusiasts working in tech and interested in how technology is reshaping the world we live in.

We’re not chasing news. We dig into big, uncertain questions to understand how reality is changing — and invite you to think it through with us.

References & context mentioned in this episode:

• Claude Funeral (WIRED): A real-world ritual staged around an AI system, blurring the line between performance, belief, and machine agency.
• Mortal AI Performance: An artistic exploration of artificial mortality and human emotional projection onto machines.
• Podcast with Andrej Karpathy: A wide-ranging discussion on modern AI systems, their capabilities, limitations, and what people project onto them.
• Shanahan: Artificial Intelligence and the Moral Status of Machines: A philosophical analysis of whether advanced AI systems could ever warrant moral consideration.