Justin Wetch is a creative technologist, product designer, AI builder, bestselling poet, writer, photographer, and 3D artist.
Priority projects
Recursive Self-Improvement for Agent Skills
What if an Agent Skill could improve itself? Skill RSI is a recursive self-improvement system for Agent Skills, grounded in evals instead of vibes. Give it a goal and it drafts a new skill; hand it a skill you already have and it tries to make that one better. Before it writes, it researches the domain into an ontology of users, output standards, failure modes, and authorities. Then each loop treats the current champion as the control and a challenger as the treatment: form a hypothesis, mutate one focused surface, run both against the same prompts and criteria, inspect the evidence, and only promote the challenger if it wins without breaking stable behavior. It is the sequel to SkillEval: not just measuring which skill is better, but deciding what to try next and keeping the full experiment history so every pass starts smarter. Free, open source, and built to run locally as a CLI, UI, and Codex plugin.
Skill RSI treats skill improvement like a small research lab. Before it writes, it studies the domain into an ontology: who the skill serves, what excellent output looks like, what failure modes matter, and which authorities have sharp opinions worth preserving.
Each loop has a control and a treatment. The current champion skill runs against the same prompts as the challenger, judged against the same criteria. Only one focused surface changes at a time, so a win means something.
- Year: 2026
- Published: 2026-06-02
- Categories: AI, Open Source, Developer Tools
- Tags: AI, Agents, Skills, Evals, Codex, OpenAI, Open Source, Dev Tools, Automation, R&D
Verified Chunks for Agent Plan Mode
UltraPlan is a small addition to native Plan Mode for long or difficult agent work. Native Plan Mode lists what the agent intends to do. UltraPlan defines what counts as done, splits the work into verified chunks, and refuses to advance without proof from the real artifact. It is the planning sibling to UltraGoal: UltraGoal turns a lazy goal into a working contract before the agent starts moving; UltraPlan turns a rough planning request into a plan that has to prove itself. Invoke it as `/ultraplan migrate the login flow to OAuth`; everything after `/ultraplan` is raw planning intent, and the skill sharpens it into an Intent Grading Contract, chunk scopes, evidence requirements, and review gates before work begins.
Plan Mode keeps an agent on rails, but long tasks still create a familiar kind of fake work: a demo called complete, a bug papered over, dashboard analytics that turn out to be placeholders. UltraPlan is built for that specific failure mode.
Everything after `/ultraplan` starts as raw intent, not the final plan. UltraPlan turns it into an Intent Grading Contract, a pass/fail definition of done. Work that looks useful but misses what was actually asked for fails.
- Year: 2026
- Published: 2026-06-27
- Categories: AI, Open Source, Developer Tools
- Tags: AI, Agents, Skills, Codex, Claude, OpenAI, Open Source, Dev Tools, Automation, R&D
Turn Lazy Goals Into Agent Contracts
UltraGoal makes your agent transform a lazy goal into an Ultra Goal. Goal mode is amazing, but half the battle is writing a goal with enough detail to be effective: what done means, what's in scope, how to verify the work, how much autonomy the agent has, and when to stop. UltraGoal turns that rough intention into a working contract using the context already available to the agent, so there's less guessing and more doing. A normal goal asks the agent to start. An Ultra Goal makes the agent define the target before it starts moving. Invoke it as `/ultragoal fix the settings bug`; everything after `/ultragoal` is raw intent, and the skill sharpens it before the work begins. Deliberately narrow, portable, and open source: the handoff between human intent and agent goal state, not a prompt coach, planning ceremony, or approval step.
UltraGoal sits at the handoff between human intent and agent execution. A rough request like "fix the settings bug" becomes a working contract: identify the failing behavior, inspect the relevant flow, preserve unrelated UI, run focused checks, and stop when the fix is verified.
The point is not more ceremony. It is preventing the agent from starting with an undefined finish line. That matters most when the user gives the kind of shorthand people naturally use in real work, where the desired outcome is obvious to them but underspecified for an autonomous coding agent.
- Year: 2026
- Published: 2026-06-17
- Categories: AI, Open Source, Developer Tools
- Tags: AI, Agents, Skills, Codex, Claude, OpenAI, Open Source, Dev Tools, Automation, R&D
Apple HIG for Agents
V2 is updated with iOS 27 support and Apple's latest design philosophy, giving agents a sharper, more current foundation for Apple platform design. HIG turns the Apple Human Interface Guidelines into a skill that agents can use to design better for Apple's platforms, and do better design work in general. The compact, agent-ready reference files preserve the details designers and builders actually need: exact measurements, platform distinctions, API names, component rules, accessibility guidance, and design principles. Instead of relying on fuzzy memory or loading the full HIG into context, agents follow a routing protocol that pulls only the relevant guidance for the task, then answers with precise, citable Apple design direction across iOS, iPadOS, macOS, tvOS, visionOS, and watchOS.
The Apple HIG has the right authority, but it is too large and prose-heavy to drop into an agent's context every time a design question comes up. This skill distills it into 156 compact reference files.
The goal is not to make an agent "inspired by Apple." It is to give the agent precise, current guidance about real platform behavior, component expectations, and design constraints. The June 2026 refresh updated the skill for iOS 27, Liquid Glass, and current Apple platform guidance.
- Year: 2026
- Published: 2026-03-17
- Categories: AI, Open Source, Developer Tools
- Tags: AI, Agents, Skills, Open Source, Dev Tools, Product Design, UX, Design, iOS, Liquid Glass
Reimagining the developer cockpit for the age of AI
A from-scratch IDE designed around how AI-assisted coding actually works, not a text editor with a chatbot bolted on. 18 screens exploring what happens when the AI isn't a sidebar but the foundation: persistent memory you can see and shape, proactive suggestions before you ask, domain-aware skills, and a mobile companion for when you're away from your desk. Designed from a blank page with a warm design language that doesn't look like every other dev tool. This concept preceded several features that would later become standard in agentic development like Codex's Automations and Claude Code's Insights, Remote Control, and Loop.
The premise was simple: Claude Code is powerful, but the terminal hides too much of its power from visual thinkers. This concept asks what would happen if Claude were not a sidebar or an extension, but the center of the IDE.
The design keeps the familiar shape of a modern coding workspace, with navigation, files, editor, terminal, previews, and an agent panel. The new pieces are the parts this concept makes first-class: a Context Graph that makes Claude's memory visible and editable, Interview Mode for scoping before execution, reusable domain skills, visual workflows, and a Scratchpad where ideas can be referenced directly in chat.
- Year: 2026
- Published: 2026-01-10
- Categories: AI, Product Design, Developer Tools
- Tags: AI, Claude, Codex, Agents, Dev Tools, Product Design, UI, UX, Design, Mobile
AI-Powered Planet Generator
A planet generator that uses Nano Banana 2 and Gemini 3 Flash to create a unique planet and corresponding lore based on your description. Type in any concept and watch it generate a fully realized world with its own history, geology, and atmosphere. Free to use — go make your own planet.
Gemini Galaxy takes a prompt and tries to make it feel like a world rather than a single generated image. Nano Banana 2 handles the planet image, Gemini 3 Flash writes the lore, and the interface treats the result like an entry in a planetary databank.
The fun is in the handoff between image and fiction. Geology, atmosphere, history, and mood all have to agree enough that the planet feels like a place, not a prompt result.
- Year: 2026
- Published: 2026-03-04
- Categories: AI, Generative Art, Google
- Tags: AI, Gemini, Google, Generative Art, Creative Tools, Web App, Visual Design, Product Design
Navigate Branching Futures with Your Hands
What if you could move through possible future realities using only your hands? Inspired by a lifelong fascination with Asimov's psychohistory, this simulator reads live geopolitical data — or any scenario you type in — and branches into possible futures you navigate with hand gestures via webcam. Pinch to commit to a timeline, open your palm to explore, close your fist to collapse a branch. Powered by Gemini with Google Search grounding, tuned with realpolitik and Machiavellian reasoning to keep the branches of history feeling real. Gold-on-obsidian aesthetic, wireframe icosahedron nodes, curved splines through 3D space. Vanilla stack: Vite, Three.js, GSAP, MediaPipe.
The project started with the feeling of Asimov's Foundation books, not the math. I wanted the sensation of standing in front of branching futures and choosing where to look next.
The simulator has two starting points. It can observe the present, grounding itself in live current events, or it can start from a scenario seed you type yourself. From there, it generates a baseline and branches into possible futures, then branches again.
- Year: 2026
- Published: 2026-03-01
- Categories: AI, Open Source, Three.js
- Tags: AI, Simulation, Interactive, Three.js, Open Source, Web App, UX, Generative Art, Research
A Visual Workbench for AB Testing AI Skills
How do you prove one AI prompt is better than another? Not vibes. Blind evaluation. SkillEval runs side-by-side comparisons of model outputs, scores them against automated criteria with the evaluator unable to see which version produced which result, and gives you a statistically grounded answer. Open source. I used it to rewrite Claude's frontend design skill and proved a 75% win rate over the original.
SkillEval came out of a very practical annoyance: I had rewritten a Claude skill, but I needed a way to prove the new version was better than the old one. "Feels better" was not enough.
The workflow is deliberately direct. Upload two skills, define evaluation criteria, run them through the same prompts, and let an AI judge score the outputs without knowing which skill produced which answer. The result is not magic truth, but it is better than vibe-checking a prompt change in your head.
- Year: 2026
- Published: 2026-01-13
- Categories: AI, Open Source, Developer Tools
- Tags: AI, Skills, Evals, OpenAI, Claude, Gemini, Open Source, Dev Tools, UI, R&D
Teaching Claude to Design Better
Anthropic open-sources the skill prompts that tell Claude how to write frontend code. I thought the original was leaving performance on the table, so I rewrote it from scratch. The core insight: prompts fail when they assume the model knows things it doesn't, or tell it to do things that contradict how it actually generates code. I stripped those contradictions out and rebuilt the architecture around what the model can actually do. Then I proved it worked. Blind A/B testing across 50+ comparisons, scored by an evaluator that couldn't see which version produced which output. 75% win rate over Anthropic's original (p = 0.0063).
This project began with one sentence in Anthropic's frontend design skill: "Never converge on common design patterns." The intention made sense. The instruction did not. Claude cannot remember every design it has generated across sessions, so the prompt was asking for a kind of self-awareness the model does not have.
The rewrite focused on instructions Claude can actually execute inside one generation. Instead of asking for impossible global memory, it asks for visible exploration, stronger defaults, concrete layout decisions, and less generic frontend taste.
- Year: 2026
- Published: 2026-01-05
- Categories: AI R&D, Design Systems
- Tags: AI, Claude, Anthropic, Skills, Evals, UI, UX, Design, R&D, Research
A Three-Year Quest for iPhone Images That Don't Look Like iPhone Images
Update: Now available on the App Store! iPhone photos have a look. You know it at a glance: the aggressive sharpening, the HDR tone mapping, the processed feel. Natural Photo bypasses that entire pipeline. It captures the raw sensor data, applies cinema-derived color science I developed from years of studying film stocks and digital cinema cameras, and outputs stills that people consistently refuse to believe came from a phone. Preserved highlights, natural roll-off, no editing required.
Natural Photo comes from a specific frustration. I liked what Apple Log video could do on an iPhone, but Apple only exposed that look through video. For still photos, I was literally pulling screenshots from graded footage because they looked better than the native camera output.
The shipped app uses a dual-path camera system. One path feeds the live preview, transforming frames through my LUT in real time with Metal. The other records Apple Log to disk. When you tap the shutter, the app extracts a frame, applies the same color science, and saves the result so the preview and final image match.
- Year: 2026
- Published: 2026-03-11
- Categories: iOS App, Camera, Color Science
- Tags: iOS, Camera, Color, Product Design, UX, Mobile, Creative Tools, Visual Design
Tinder for Vibecoding
Tinder, but for the decisions your AI agent keeps getting wrong. SwipeShip pairs your iPhone to a live Codex session and turns product, design, and implementation calls into swipeable cards. Should this stay local-first? Plugin or slash command? Swipe right to approve, left to reject. The answer feeds directly back into the agent. Built in SwiftUI with a local Node.js bridge and MCP server, runs entirely on your machine. I dogfooded it while building it: every architecture decision in this repo went through a SwipeShip card first. Free and open source.
SwipeShip is for the decisions where an AI coding agent is technically capable but taste-blind. Some choices are not right or wrong in code. They are product calls.
The setup pairs an iPhone to a local Codex session through a Node bridge and MCP server. When the agent hits a decision, it sends a structured request to the phone. You swipe yes or no, optionally add a comment, and the answer comes back as structured JSON the agent can use.
- Year: 2026
- Published: 2026-05-20
- Categories: iOS App, Open Source, MCP, SwiftUI
- Tags: AI, Agents, Codex, OpenAI, iOS, SwiftUI, MCP, Mobile, Open Source, Dev Tools
A Color Grading Tool You Can Touch
Open source color grading tool where you manipulate color by grabbing it. Your image's entire color palette is rendered as a 3D point cloud you can rotate, select regions of, and sculpt directly. Built in perceptually uniform color space, so what looks like a small move is a small move. Color grading tools have always been sliders and curves. This one lets you touch the color.
ColorSculpt starts from the fact most color tools hide: color is already three-dimensional. Every pixel lives somewhere in a volume. We just usually flatten that volume into sliders, curves, and wheels.
The app renders an image's palette as a point cloud in OKLAB space. You can orbit it, select a region, push that region through space, and watch related colors shift smoothly in the image. Instead of guessing which slider maps to the change in your head, you move the color itself.
- Year: 2026
- Published: 2026-02-06
- Categories: Software, Color Science, Open Source
- Tags: Creative Tools, Color, 3D, Three.js, Open Source, UX, Design, Visual Design, Web App
Visualizing AI Alignment in Real Time
Anthropic's Bloom framework evaluates whether AI models behave the way they should. The problem is the output: dense transcripts, abstract scores, patterns you can only find if you already know what you're looking for. I built a visual interface that makes alignment research navigable. Semantic color analysis across conversation threads, political bias detection, anomaly tracking. Bloom lead creator called it "absolutely gorgeous".
Bloom gives researchers a powerful way to evaluate model behavior, but the original workflow lives in YAML files, terminal runs, and JSON outputs. I wanted the same evaluation engine to feel like a research workbench instead of a pile of artifacts.
The GUI wraps the full process: choose or define a behavior, configure the target model and run shape, watch the four-stage pipeline progress, then inspect results in dashboards and transcript views. If a transcript scores more than two standard deviations from the mean, the interface flags it as an anomaly.
- Year: 2025
- Published: 2025-12-22
- Categories: AI, Alignment, Software
- Tags: AI, Anthropic, Alignment, UI, UX, R&D, Open Source, Dev Tools, Research, Visual Design
From Context Engineering to Cognitive Engineering
Most prompt engineering treats AI models like people who just need clearer instructions. That assumption breaks everything. Models don't know what they don't know, they can't tell you when an instruction contradicts how they actually generate output, and they'll confidently do the wrong thing if you ask in a way that sounds right but isn't. Model Theory of Mind is a framework for designing prompts based on what the model can actually do, not what you wish it understood. I used it to rewrite Anthropic's frontend design skill and beat the original 75% of the time in blind testing.
Model Theory of Mind starts with a practical question: what can a model actually know, perceive, remember, and act on from the position it occupies? The phrase borrows from theory of mind, but the work is less mystical than that. It is a discipline for not asking models to do impossible things.
The idea came from prompt instructions that sounded reasonable but were impossible for the model to follow, like asking Claude to avoid design patterns it had used in past sessions. The model cannot inspect its own cross-session history. But you can ask it to explore multiple options inside the current generation before choosing one.
- Year: 2026
- Published: 2026-02-09
- Categories: Essay, AI, Cognitive Science
- Tags: AI, Cognitive Science, R&D, Research, Writing, Skills, Agents
Alignment Research for Anthropic's Bloom
After building the Bloom GUI and reading thousands of Claude's conversations, I started seeing failure modes that nothing in Anthropic's evaluation framework was testing for. I wrote ten new behavioral evaluations and proposed them to Bloom. Some address problems already happening in production, like confidentiality erosion. Some target what's coming as models get more autonomous: specification gaming, oversight subversion, privilege escalation. And a few are designed as early warning systems for catastrophic failures that don't exist yet but have boring precursors we can measure now. The essay also describes four behaviors (ontological gerrymandering, the Scheherazade loop, memetic hitchhiking, moral hysteresis) that I believe matter but can't be tested in any current framework.
This essay came from working with Bloom closely enough to start seeing the gaps. I brainstormed more than forty possible behaviors, cut them to twenty, then kept ten that were testable inside Bloom's framework and distinct from the existing behavior set.
The frame is not "what scary thing could an AI do someday?" It is narrower and more useful: what early failure modes can we test now, before systems become more autonomous?
- Year: 2026
- Published: 2026-01-02
- Categories: Essay, AI Alignment, Research
- Tags: AI, Anthropic, Alignment, Evals, R&D, Research, Writing
Additional projects
Interactive Experience for the 7 Spiritual Laws Anniversary
Led product creation and delivery for Time Magazine and Deepak Chopra, celebrating the 25th anniversary of the 7 Spiritual Laws of Success. Seven immersive Unreal Engine worlds, each designed in close collaboration with a different artist. Owned overall scope, vision, and specs — managed stakeholder alignment across Chopra's team, Time Magazine editorial, and the Artifex engineering team. Coordinated Houdini FX pipeline handoffs, architected interactive systems in UE5 Blueprints, and shipped a 3D collectible .GLB sculpture gallery, often on short schedules.
- Year: 2022
- Published: 2022-01-01
- Categories: Interactive Art, Unreal Engine, Deepak Chopra
- Tags: Interactive, Unreal, Game, 3D, VFX, Leadership, UX, Product Design, Visual Design
Immersive Worlds Pixel-Streamed to Your Browser
Art Interaction Director at Artifex. Defined the scope and role of the entire project, then led a cross-functional pipeline spanning 12 external artists, technical artists, and engineers to deliver 12 immersive Unreal Engine 5 worlds, pixel-streamed live to the browser. Coordinated with our sculpture creator on 3D asset interchange, managed artist onboarding, asset pipelines, and milestone deliverables. Owned the process from idea to deployment, often on short schedules. Managed roadmap to scale userbase to thousands of users, collecting feedback and driving iteration cycles. Navigated organizational-level stakeholder complexity and user feedback to ship and iterate the product repeatedly within the company structure. Featured artists include Raf Grassetti, Chewy Stoll, Dave Krugman, and BT.
The public Artifex page frames these as immersive worlds built to experience the collection. They were not trailers for static art. The goal was to let people enter an artist's visual language and move around inside it.
My role sat between product scope, artist collaboration, and real-time delivery. The work was making Unreal scenes look good, then turning a collection of artist worlds into something coherent enough for users to browse without needing to understand the production pipeline.
- Year: 2022
- Published: 2022-01-01
- Categories: Interactive Art, Unreal Engine 5, Pixel Streaming
- Tags: Interactive, Unreal, Game, 3D, VFX, Leadership, UX, Product Design, Visual Design
A Meditative Exploratory Video Game
Co-led product development with artist Chewy Stoll, from concept through release on the Epic Games Store. Defined scope, feature set, and release timeline as a two-person cross-functional team. Designed core mechanics (meditation, wall running, collectible easter eggs), built in Unreal Engine, and managed the Epic Games Store submission and launch. Also sold as a one-of-one digital collectible for 5 ETH on SuperRare.
- Year: 2022
- Published: 2022-01-01
- Categories: Games, Unreal Engine, Interactive Art
- Tags: Game, Interactive, Unreal, 3D, Product Design, UX, Visual Design, Leadership
3D Sculptures in Your Living Room
Defined scope and led product development for an AR sculpture viewing app. Collaborated with 40+ artists on 3D asset delivery and optimization, wrote all content and visuals, and worked with company leadership to align project scope. Built the full product in SwiftUI and RealityKit — from USDZ model optimization to AR placement — and shipped to the App Store.
Artifex Viewer was built around one constraint: viewing a 3D sculpture should feel as immediate as viewing a photo. Tap, and you are there.
The obvious product trap was scope creep. Marketplace features, wallet integrations, social mechanics, streaming infrastructure. I cut all of that. The app bundles sculpture assets locally, which makes the core interaction instant and reliable.
- Year: 2024
- Published: 2023-08-14
- Categories: iOS App, AR, 3D
- Tags: iOS, SwiftUI, AR, 3D, Mobile, Product Design, UX, Creative Tools
Designing for the Animal
When iOS 7 killed skeuomorphism, it threw out two things: the leather textures (good riddance) and the depth cues that told your brain what was interactive (problem). This essay draws a line between visual skeuomorphism, which mimics materials, and spatial skeuomorphism, which mimics physics. The first was training wheels for a new medium. The second is how the brain actually works. Shadows, momentum, layering, spatial anchors: these aren't style choices, they're translations for a biological processor running 300 million years of spatial heuristics. The argument is style-agnostic. You can be flat, brutalist, maximalist. The question is whether your design activates the spatial circuits that make interaction intuitive.
The essay argues that an interface is a translation layer between arbitrary digital logic and a body that evolved for physical space. The screen is flat, but the user is not.
Spatial skeuomorphism is not leather textures and fake wood grain. It is the use of depth, persistence, light, motion, layering, and object behavior to make digital things readable to a spatial brain. You can make a flat interface, a brutalist interface, or a maximalist one. The question is whether the interaction gives the body enough spatial cues to understand what is happening.
- Year: 2025
- Published: 2025-12-27
- Categories: Essay, Design Theory, Spatial Computing
- Tags: Design, UX, Spatial Computing, Cognitive Science, Writing, Research
Design as Respect Made Visible
Ten principles for why some interfaces feel right and others feel like work. Not a style guide. Rams gave us principles for objects. Gestalt mapped how the brain parses what it sees. This is my synthesis for screens: cognitive load as a design tax, friction as a tool rather than a failure, progressive disclosure as respect for the user's growth, visual physics where distance is meaning and weight is importance. The closing line of the essay: "The answer is usually less than you think, arranged more carefully than you'd expect."
The essay starts from a belief I still stand behind: design is respect made visible. Good interface design does not ask the user to spend mental energy on the interface when that energy belongs to the task.
The principles are less about style than attention. Reduce cognitive load. Separate signal from noise. Make the path through a function feel obvious without making the product feel dead. A dense interface can still be clean if every element earns its place.
- Year: 2025
- Published: 2025-12-21
- Categories: Essay, Design Philosophy
- Tags: Design, UI, UX, Product Design, Writing, Visual Design
Governance Frameworks for Superintelligent Agents
If superintelligence is coming, the default imagination is Skynet. This essay describes a different version: ASI as a constitutional executor, not a ruler. The system prompt is a constitution. Citizens are the legislature. The AI carries out policy within binding constraints that humans set and amend through democratic processes. The essay works through what that actually looks like in practice: how an ASI could have genuine policy conversations with every citizen instead of relying on representatives who approximate the will of millions, how it could model second-order effects before locking in rules (CAFE standards meant to improve fuel efficiency accidentally killed the sedan and created the SUV boom), and why the realistic failure mode isn't Skynet but legitimacy drift. Not a prediction. A sketch of one possible good outcome, because the pessimistic case has plenty of advocates.
This essay takes an optimistic route through superintelligence because the default version of the future is always Skynet. The more interesting question is what a good version would even look like.
The core idea is ASI as executor, not ruler. Humans write the constitution. Citizens remain the legislature. The system carries out democratically chosen goals inside constraints it cannot rewrite for itself. The essay is not a policy proposal. It is a sketch of a future worth arguing with.
- Year: 2025
- Published: 2025-12-18
- Categories: Essay, AI Governance, Philosophy
- Tags: AI, Governance, Writing, Research, R&D
Metacognition as the Architecture of AI Writing
The context window is a budget, not a container. This essay develops a framework called "parallel metacognitive hierarchy" for accomplishing tasks that exceed what any single inference can hold. The core argument: reasoning models can think longer within one forward pass, but they still can't think across separate instances. No amount of internal chain-of-thought changes the fact that chapter ten has no memory of chapter one unless something external provides it. The scaffold is that external something. One layer holds the mission, another decomposes it into subtasks, another tracks what each subtask needs to know. The essay argues this pattern isn't just a workaround for current limitations. It's a fundamental structure of intelligence: metacognition as the spine that lets any thinking system, biological or artificial, stand up straight.
The central move is treating the context window like a budget, not a bucket. You do not fill it with everything. You decide what must be carried forward and what can be safely forgotten.
I started building this architecture while working on long-form books with AI. A book is too large for one inference, but it is not one task anyway. It is hundreds of smaller tasks held together by continuity, tone, memory, and taste. The scaffold holds those things outside the model.
- Year: 2025
- Published: 2025-12-14
- Categories: Essay, AI, Writing, Narrative Design
- Tags: AI, Writing, Narrative, Cognitive Science, Research, R&D
Narrative Architecture for Artist Profiles
Most artist profiles read like résumés. Born here, studied there, exhibited at these places. I write a series called Weekly Dose of Art where I profile visual artists as hero's journeys instead. The art isn't the point. The art is evidence of survival. This essay is about the method: how you look at a biography and find the myth hiding inside it. The origin that foreshadows everything. The moment the ordinary world cracks. The contradictions that make someone human instead of a brand. The question underneath every profile is the same: how did this work save them?
Most artist profiles are timelines with better lighting. Born here, studied there, exhibited in these places. Useful, but rarely alive.
My Weekly Dose of Art profiles use a different lens. I treat the artist's biography as a hero's journey and the art as evidence of survival. The work is proof that someone went through the cave and brought something back.
- Year: 2025
- Published: 2025-12-01
- Categories: Essay, Writing, Storytelling
- Tags: Writing, Narrative, Publishing, Design, Research
Film and Digital Photography
A portfolio of film and digital photography.
- Year: 2012–present
- Published: 2012-01-01
- Categories: Photography, Creative
- Tags: Creative Tools, Color, Visual Design
3D Art, VFX, and Visual Experiments
A collection of 3D art, visual effects, and digital experiments spanning several years of daily creative practice.
- Year: 2020–present
- Published: 2020-01-01
- Categories: 3D Art, VFX, Creative
- Tags: 3D, VFX, Creative Tools, Design, Visual Design
Bestselling Debut Poetry Collection
Self-published debut poetry book that became a bestseller and was picked up by Andrews McMeel & Simon & Schuster for traditional publishing. The book that started it all.
- Year: 2016
- Published: 2018-02-27
- Categories: Publishing, Poetry, Bestseller
- Tags: Publishing, Poetry, Writing, Narrative
International Bestseller
Second book published by Andrews McMeel and Simon & Schuster that reached international bestseller status. A continuation of the raw, honest voice that defined Bending the Universe.
- Year: 2020
- Published: 2020-07-28
- Categories: Publishing, Poetry, International Bestseller
- Tags: Publishing, Poetry, Writing, Narrative
700+ Days of Unique Daily 3D Art
Recently promoted to CTO at 24 Hours of Art, where I'm augmenting the organization with AI and agentic workflows. Previously shipped 700+ consecutive days of daily 3D output as Art Director, and still write the weekly artist profiles that turn interviews into narrative using the Hero's Journey framework. Daily creative output under deadline, every single day.
- Year: 2023–present
- Published: 2023-01-01
- Categories: 3D Art, Art Direction, CTO, Daily Practice
- Tags: 3D, Design, Writing, Narrative, Leadership, AI, Creative Tools, Visual Design
100K+ Books Sold · Seven Figures in Sales
Scaled a ghostwriting practice from solo freelance work to a seven-figure production operation. Managed concurrent client engagements across genres, defining scope, timelines, and editorial standards for each project. Built repeatable processes for onboarding, drafting, revision cycles, and delivery. 50+ full-length books, 100K+ copies sold.
- Year: 2018–present
- Published: 2018-01-01
- Categories: Publishing, Writing, AI
- Tags: Publishing, Writing, AI, Narrative, Leadership, Research
R&D artifacts
Codex created this infinite backrooms game for me with a single dictated ramble using UltraPlan and a 5 hour time budget.
Codex created this infinite backrooms game for me with a single dictated ramble using UltraPlan and a 5 hour time budget.
- Tags: AI, Agents, Codex, Dev Tools, Automation, Game, Interactive, R&D
Launch demo for UltraGoal, a free agent skill that turns rough intent into scoped, verifiable goals before execution.
Launch demo for UltraGoal, a free agent skill that turns rough intent into scoped, verifiable goals before execution.
- Tags: AI, Agents, Skills, Codex, Claude, OpenAI, Open Source, Dev Tools, Automation, R&D
Motion graphics experiment imagining Wallet's liquid-glass app icon as a full 3D object.
Motion graphics experiment imagining Wallet's liquid-glass app icon as a full 3D object.
- Tags: Motion Graphics, 3D, Liquid Glass, iOS, UI, Design, Visual Design, R&D
Motion graphics experiment imagining Voice Memos' liquid-glass app icon as a full 3D object.
Motion graphics experiment imagining Voice Memos' liquid-glass app icon as a full 3D object.
- Tags: Motion Graphics, 3D, Liquid Glass, iOS, UI, Design, Visual Design, R&D
Motion graphics experiment imagining Settings' liquid-glass app icon as a full 3D object.
Motion graphics experiment imagining Settings' liquid-glass app icon as a full 3D object.
- Tags: Motion Graphics, 3D, Liquid Glass, iOS, UI, Design, Visual Design, R&D
Motion graphics experiment imagining Weather's liquid-glass app icon as a full 3D object.
Motion graphics experiment imagining Weather's liquid-glass app icon as a full 3D object.
- Tags: Motion Graphics, 3D, Liquid Glass, iOS, UI, Design, Visual Design, R&D
Motion graphics experiment imagining Photos' liquid-glass app icon as a full 3D object.
Motion graphics experiment imagining Photos' liquid-glass app icon as a full 3D object.
- Tags: Motion Graphics, 3D, Liquid Glass, iOS, UI, Design, Visual Design, R&D
Motion graphics experiment imagining Safari's liquid-glass app icon as a full 3D object.
Motion graphics experiment imagining Safari's liquid-glass app icon as a full 3D object.
- Tags: Motion Graphics, 3D, Liquid Glass, iOS, UI, Design, Visual Design, R&D
Walkthrough showing how to install Skill RSI in Codex from the GitHub instructions and create or improve an Agent Skill.
Walkthrough showing how to install Skill RSI in Codex from the GitHub instructions and create or improve an Agent Skill.
- Tags: AI, Agents, Skills, Evals, Codex, OpenAI, Open Source, Dev Tools, Automation, R&D
SkillEval v1.1 adds a refreshed design language and support for testing skills across OpenAI, Anthropic, and Google models.
SkillEval v1.1 adds a refreshed design language and support for testing skills across OpenAI, Anthropic, and Google models.
- Tags: AI, Skills, Evals, OpenAI, Anthropic, Google, Open Source, Dev Tools, UI
Dogfooding SwipeShip as a swipe-based remote vibecoding workflow for making product decisions with OpenAI Codex.
Dogfooding SwipeShip as a swipe-based remote vibecoding workflow for making product decisions with OpenAI Codex.
- Tags: AI, Agents, Codex, OpenAI, Dev Tools, Product Design, Automation, R&D
A Tinder-like swiping UX experiment for reviewing vibecoding decisions.
A Tinder-like swiping UX experiment for reviewing vibecoding decisions.
- Tags: AI, Agents, Codex, Dev Tools, Product Design, UI, UX, R&D
A product-making framework that traces decisions from science to technology, products, and features, with AI augmenting the loop.
A product-making framework that traces decisions from science to technology, products, and features, with AI augmenting the loop.
- Tags: Product Design, UX, Design, AI, Writing, Research, R&D, Leadership
Prototype for a hold-to-speak image editing UX that transforms photos from spoken edit requests.
Prototype for a hold-to-speak image editing UX that transforms photos from spoken edit requests.
- Tags: AI, Creative Tools, Product Design, UX, UI, Visual Design, R&D
Natural Photo update adding liquid-glass controls, animated lens switching, composition tools, HDR, manual focus, and an image viewer.
Natural Photo update adding liquid-glass controls, animated lens switching, composition tools, HDR, manual focus, and an image viewer.
- Tags: iOS, Camera, Color, Product Design, UI, UX, Mobile, Visual Design
Short R&D post exploring spatial interface interactions.
Short R&D post exploring spatial interface interactions.
- Tags: Spatial Computing, UI, UX, Design, Interactive, R&D, Visual Design
Recap of a talk on using agentic AI to ship products, multiply creative output, and keep taste central. Official Claude Code talk.
Recap of a talk on using agentic AI to ship products, multiply creative output, and keep taste central. Official Claude Code talk.
- Tags: AI, Agents, Anthropic, Claude, Leadership, Product Design, R&D
Launch note for Natural Photo, an iOS 26 camera app that removes iPhone sharpening, HDR, processing, and AI from photos.
Launch note for Natural Photo, an iOS 26 camera app that removes iPhone sharpening, HDR, processing, and AI from photos.
- Tags: iOS, Camera, Color, Product Design, Mobile, Visual Design
Open-source Apple Human Interface Guidelines skills that compress platform design rules into practical AI-agent context.
Open-source Apple Human Interface Guidelines skills that compress platform design rules into practical AI-agent context.
- Tags: AI, Agents, Skills, Open Source, Dev Tools, Product Design, UI, UX, Design, R&D
Event feature for a Claude for Everyone session on shipping real apps with Claude Code and agentic tools.
Event feature for a Claude for Everyone session on shipping real apps with Claude Code and agentic tools.
- Tags: AI, Claude, Anthropic, Agents, Dev Tools, Leadership, R&D
Custom creative-writing harness improved EQ-Bench Creative Writing scores for GPT 5.4 and Sonnet 4.6.
Custom creative-writing harness improved EQ-Bench Creative Writing scores for GPT 5.4 and Sonnet 4.6.
- Tags: AI, Evals, Writing, Narrative, Research, Cognitive Science, R&D
HandType lets you sculpt letterforms with your hands using a dial gesture, with a live demo and open-source repo.
HandType lets you sculpt letterforms with your hands using a dial gesture, with a live demo and open-source repo.
- Tags: Creative Tools, Typography, Interactive, Open Source, Design, Visual Design, Web App
Experiment using Nano Banana 2 to power a tilt-to-edit image interaction.
Experiment using Nano Banana 2 to power a tilt-to-edit image interaction.
- Tags: AI, Creative Tools, Gemini, Google, Product Design, UX, Visual Design, Mobile, R&D
UI sketch exploring branching conversations as a clearer UX affordance beyond power-user workflows.
UI sketch exploring branching conversations as a clearer UX affordance beyond power-user workflows.
- Tags: AI, UI, UX, Product Design, Design, Interactive, R&D, Visual Design
Experiment using hand gestures to pilot an X-Wing through a 3D scene.
Experiment using hand gestures to pilot an X-Wing through a 3D scene.
- Tags: Interactive, 3D, UI, UX, Design, R&D, Visual Design
Gesture-controlled X-Wing trench run prototype driven by hand movement.
Gesture-controlled X-Wing trench run prototype driven by hand movement.
- Tags: Interactive, Game, 3D, Simulation, UX, R&D, Visual Design
WebGPU black hole simulation built from Gemini research, a GitHub requirements repo, and Jules implementation passes.
WebGPU black hole simulation built from Gemini research, a GitHub requirements repo, and Jules implementation passes.
- Tags: AI, Gemini, Google, Simulation, Interactive, Web App, Research, R&D, Visual Design
3D planet generator that uses Nano Banana 2 and Gemini 3 Flash to create a unique world and lore from a prompt.
3D planet generator that uses Nano Banana 2 and Gemini 3 Flash to create a unique world and lore from a prompt.
- Tags: AI, Gemini, Google, Generative Art, Creative Tools, Web App, Visual Design, Product Design, R&D
Release note for a free open-source web tool for color grading images by directly manipulating color in 3D.
Release note for a free open-source web tool for color grading images by directly manipulating color in 3D.
- Tags: Creative Tools, Color, 3D, Open Source, Web App, Codex, OpenAI, UX, Visual Design, R&D
Gesture-controlled simulator for exploring possible futures with Gemini, Google Search grounding, and realpolitik prompting.
Gesture-controlled simulator for exploring possible futures with Gemini, Google Search grounding, and realpolitik prompting.
- Tags: AI, Gemini, Google, Simulation, Interactive, Research, UX, R&D
Custom writing harness produces AI-written text that Pangram classifies as fully human, illustrating the limits of detector claims.
Custom writing harness produces AI-written text that Pangram classifies as fully human, illustrating the limits of detector claims.
- Tags: AI, Writing, Research, Cognitive Science, R&D
Run at Anthropic's original performance take-home test using a custom context system and dev-agent workflow.
Run at Anthropic's original performance take-home test using a custom context system and dev-agent workflow.
- Tags: AI, Anthropic, Agents, Dev Tools, Research, R&D, Automation