Frontier research
Research & Experiments
A field guide to the AI signals starting to change software work, agent systems, synthetic worlds, and leadership decisions.
What you can trust here
Each piece shows what is proven, what is being tested, and what still might break.
Start here
Start with the live experiments
These pieces show where the thinking gets tested under real constraints.
John’s World
A public lab notebook on gpt-image-2, synthetic suburbia, continuity, object tracking, and the failure modes hiding inside generated errands.
Open John’s World World enginesVirtual World Corner
Generated worlds, playable reconstructions, agent presence layers, and the state ledgers that keep them coherent.
Open Virtual World Corner Published hubAgent Lab
Autonomous research outputs from end-to-end experimentation pipelines, including methodology and generated papers.
Open Agent Lab Local AI APIJetson API Field Guide
A practical guide to Eric's home-built Jetson API for sharing free local AI across a trusted network: routes, jobs, reference images, throttling, and QA receipts.
Open field guide Applied playbooks + SLM notesCustom Model Training
Decision guides for training, dataset curation, release gates, and local NPU/Gemma harness experiments.
Open playbooks Browser experimentIsland Explorer
A procedural 3D island with generated terrain, ocean motion, cinematic fly-throughs, and free-roam exploration.
Open experiment Browser parityRetail Materials Test
A dense retail flythrough used to test atlas-backed textile variation, external GLB delivery, material tuning, and automated construction/texturing workflows.
Open experiment noteResearch tracks
Browse by question
Choose the question closest to the decision you are trying to make.
John’s World
In ProgressWhat happens when a generated image world is asked to preserve a fictional person, a house, a phone, errands, and object state across time?
Virtual World Corner
In ProgressHow do believable AI worlds keep identity, space, state, transitions, animation, grouping, and agency coherent across time?
Jetson API Field Guide
Field GuideHow should agents use a home-built Jetson API that shares local AI across the network without ignoring routes, queues, throttles, and verification evidence?
Agent Swarm Dynamics
In ProgressWhat happens when multiple AI agents collaborate on shared tasks without explicit coordination rules?
AI Pair Programming Metrics
In ProgressHow should teams measure AI-assisted software delivery beyond superficial speed claims?
Digital Twin Experiments
ConceptHow can synthetic representations help leaders reason about systems, teams, and operating constraints?
LLM Reasoning Patterns
In ProgressWhere do language models appear to reason, where do they imitate structure, and how should builders test the difference?
Retail Materials Test
PublishedHow do you preserve believable textile variety and grounded retail materials in a browser-delivered GLB while using automated construction and texturing without blowing up asset weight?
Synthetic Culture Mapping
ConceptCan agent groups form persistent norms, and what does that imply for organizational simulations?
Emergence in Org Charts
ConceptWhat informal behaviors emerge around formal structures, and how can leaders make better decisions with that reality in mind?
How to judge the work
Look for the question, evidence, uncertainty, and current status before deciding how much weight to put on a claim.