Explorable Experiments

Visualize neurons, weights, and activations in real time.

Demystifying AI parameters through interactive weight manipulation.

How learning actually happens — gradients, loss, and iteration.

Gradient descent, loss landscapes, and what training actually does.

How neural networks build internal representations of the world.

From binary to floating point — what numbers look like to a machine.

Deconstructing GPT — what it means (and doesn't) to "learn".

Perception, reasoning, and action loops in agent systems.

Build and watch AI agents execute tool-driven workflows.

From chatbots to autonomous systems — the design patterns in between.

A structured look at how agentic systems are designed and composed.

The architecture of agent harnesses — how orchestration actually works.

How 'Beads' gives AI agents long-term memory beyond the context window.

From amnesia to expertise: how agents learn and retain knowledge.

Where agent autonomy breaks down — and why the ceiling keeps moving.

Beating AI context rot with iterative self-summarisation.

The recursive language model and what lies beyond the token limit.

Rolling vs folding through volume, void space, and time.

Why folding fitted sheets is a battle against geometry.

How tape holds under tension, peel, and shear forces.

The slip force and how card magicians control your choice.

Rhythm, weight transfer, and timing as controllable parameters.

Domino physics, timing, and chain reactions.

The physics and pattern behind glitter, shimmer, and light scatter.

An interactive battle — how your body identifies and destroys threats.

An audit of human sitting — why posture is a dynamic problem.

The mechanics of pixel art — colour, coverage, and craft.

The physics of lock cracking and why flaws enable picking.

How locks work, from pins to shear lines.

What it means to think adversarially about AI systems.

The mechanics of AI red-teaming — probing, provoking, patching.

Interactive double-slit experiment: particles, waves, and observation.

A linear algebra playground: vectors, transformations, and racing.

Where quantum computing actually stands — and why it's so hard.

Halting problems, Gödel, and what no computer can ever solve.

Is consciousness beyond the reach of any algorithm?

Commits, branches, and merges as a journey through time.

An anatomy of Git — why history is a graph, not a line.

Content-addressed storage and why you can't lose history.

Why failure dominates, incentives matter, and hype beats utility.

From data to wisdom — a sceptic's guide to AI expectations.

What coding with AI actually does to the engineer over time.

Thread-based engineering and what it means for software teams.

Why long agentic threads degrade — and what to do about it.

Navigating the internet — epistemics, signals, and noise.

BOLA, open databases, and bot swarms: why Moltbook died in 72 hours.

A scrollytelling post-mortem of "Vision-First" engineering.

Deploy your vision, spam the API, watch the Signal/Noise collapse.

When a system appears to recurse — but doesn't.