For life on Earth

Welcome to Project 3rd Rock.

An independent practice that gathers and shapes organizational expertise into systems that change what people can understand and do.

Start a conversation →

Practice

01Knowledge & discovery

02Intelligence & workflows

03Learning & participation

04Storytelling & understanding

The practice began in climate education, building course tools, resource libraries, and AI editorial workflows. It has since taken the same approach wherever expertise is hard to reach: structuring research so it can be searched and reasoned over, matching people by what they know, and building explainers people can explore.

Terra.do
Stanford University
Science Based Targets initiative
Option Zero
Global Sustainability Challenge
Meridian

The world knows more than it can use.

Research, language technology, product design, and teaching are brought to the same problem: expertise that exists but does not reach the people who need it. The practices below describe where that work tends to start, and most projects draw on several.

01 / Knowledge & discovery

Expertise is only useful when people can find it.

Language models, information retrieval, and ontology design turn unstructured documents, research, and expertise into knowledge people can search, navigate, and reason over. The design decides what belongs together, what matters for a given question, and what someone should see next.

In practice

NLP and a purpose-built skill taxonomy turned inconsistent mentor profiles into comparable expertise, and a matching algorithm paired each team with the mentors its work needed. A research library is organized around the questions its users ask while they build. aeoptimize, an independent project, studies which sources AI answers cite and which they leave out.

Core disciplines

NLP

Extracting meaning, identifying themes, and making unstructured language useful.

Ontology design

Structuring concepts and expertise so their relationships can be explored.

Information retrieval

Finding relevant sources and answers across a body of knowledge.

Algorithm design

Designing matching, ranking, and decision systems around a specific problem.

Constrained optimization

Finding workable solutions within competing priorities and practical limits.

02 / Intelligence & workflows

AI earns its place when people can trust what it produces.

Models, tools, and human judgment are designed together into workflows that carry research and content from source material to finished work. Context design and evaluation decide what each model sees and how its output is checked, so the judgment calls stay with people.

In practice

Multimodal workflows turn recorded talks into edited articles, social content, and video, with editorial review at each stage. A shared skill library encodes the voice and review standards those workflows run against, so quality holds whoever runs them.

Core disciplines

AI orchestration

Coordinating models and tools across tasks that depend on one another.

Context design

Bringing relevant knowledge, constraints, and standards into the work.

Model evaluation

Assessing quality and reliability against the needs of the task.

Multimodal systems

Working across language, images, audio, and other forms of information.

Product integration

Integrating AI into the tools and experiences people already use.

03 / Learning & participation

People learn most when they build something they can use.

Expertise becomes courses, interactive tools, and live builds that people practice with. Teaching, product design, and growth strategy are designed together, so individual learning turns into shared knowledge and a reason to return.

In practice

An AI-assisted playground for a climate communication course turns each lesson into a working tool, with model feedback on story structure, audience, and framing while participants build. The Hitchhiker’s Guide to AI teaches through concept-led material and live builds. Attested explores how people can carry verifiable recognition of their work beyond a program.

Core disciplines

Learning design

Building understanding through interactive experiences, curricula, and live exploration.

Productization

Turning expertise into tools and experiences people can use independently.

Growth strategy

Designing how people discover a product, find value, and return.

Community systems

Designing ways to contribute, exchange knowledge, and learn together.

Trust & recognition

Recognizing contributions and making achievements verifiable and shareable.

04 / Storytelling & understanding

Complex ideas land when people can explore the evidence themselves.

Writing, data, and interaction are combined so people can examine evidence and follow an argument at their own pace. Editorial systems keep meaning, voice, and uncertainty intact as an idea moves across audiences and media.

In practice

Vanishing Earth combines maps, timelines, and narrative so readers can see the pressures on five ecosystems in place and over time. For a live international event, talks and student ideas were rebuilt as editorial content that carried them to audiences who were not there.

View Vanishing Earth ↗

Core disciplines

Science communication

Making complex subjects understandable while preserving evidence and uncertainty.

Data storytelling

Explaining patterns through narrative and explorable visualizations.

Creative coding

Using code and interaction to explain ideas through experience.

Editorial systems

Maintaining voice, evidence, and consistency across a body of content.

Generative media

Exploring new forms of expression across text, images, sound, and motion.

Ideas need working systems.

Python
pandas
PostgreSQL
TypeScript
React
Next.js
D3

Structure & retrieval

Data and language structured for search, matching, and analysis.

Intelligence & process

Models and tools connected with context, evaluation, and human review.

Interfaces & interaction

Web tools and visualizations for exploration and creation.

Built to be maintained

Tool choices follow the question, who will use the result, and who maintains it.

The Observatory.

Independent projects and research on public knowledge, participation, and the living world. The previews show the questions, source material, and work completed so far.

Research in progress

Knowledge in the age of AI

How AI assistants cite their sources, read across the published audits: they mis-cite often, and few readers check. A method for an independent study is proposed.

Research in progress

The value of contribution

How shared work becomes trust between strangers, read across research that mostly comes from software, with a method to test it in climate, civic, and learning communities.

Preview

Climate Policy Countdown

Climate commitments from governments and the EU placed by target year, with the wording and source of each open for inspection.

Preview

NDC Tracker

National 2035 pledges set side by side with their baselines, coverage, and conditions, without ranking them.

Published

Vanishing Earth

An interactive story about five ecosystems and the pressures they face, developed as a Terra Studio capstone.

Explore the Observatory →

Bring a question you want to work on.

A body of knowledge people struggle to use, a workflow that needs rethinking, or an idea that could become something people can explore. Send a short introduction and a link to the work, if you have one.

Every engagement is run by Bodhi directly, from the first conversation to the finished work. That first conversation covers the question, what already exists, and whether this practice is the right fit for it.

Start a conversation →hello@project3rdrock.com