8epochs is a workshop that demonstrates experiments using AI. Each project here is a self-contained probe into a single idea, built to be useful or to amuse, and perhaps to provoke thinking and trigger ideas.
A career profile for the decade after AI, built on Claude Opus and three validated psychological frameworks. It maps the specific intersection of who you are, what drives you, and what interests you, showing where your human edge becomes worth more in the era of AI.
Take the assessment →A never-ending soap opera written by a loop, not by prompts. You submit what happens next; an Author agent drafts it and a Director agent argues it into canon (the version of events that sticks), approving, revising, or rejecting, with no human in between.
Watch the loop →A holiday-destination finder that rebuilds itself. Set what matters (flight time, month, sunshine) and AI finds where you should be. Missing a criterion? Ask for it, and an AI product team designs, tests and releases it while you watch. No human in the loop.
Find your destination →A divergent-thinking engine. State a problem and get a board of ideas across seven AI-designed lanes, from painfully obvious to deliberately terrible, because flipping a bad idea exposes the assumption underneath. Star the keepers, export the shortlist.
Generate ideas →More probes as they're built. Each one ships when it's honest, not when it's finished.
Most of what gets built with frontier models is a demo or a pitch. 8epochs is neither. It's a place to take one idea at a time seriously enough to build it, put it in front of real people, and see what holds up.
Some projects are useful tools. Some are closer to art. What they share is that the AI isn't decoration; it's load-bearing. Take it out and the thing stops working.
Everything here is independent and built by one person, Julian Reed. Not a company, not a product roadmap, and not affiliated with Anthropic or any model provider.