Building human capability in the age of AI
We believe AI should make humans more capable, not more dependent. Every tool we build asks: does this develop human skill, or just get to an answer faster?
The skills we mean are specific: critical thinking, communication, and collaboration.
Our Philosophy
The Problem We See
AI is incredibly powerful at generating outputs. But when we use it as a shortcut to skip thinking, we hollow out the very skills that make us valuable.
The pattern is seductive: ask AI, get answer, move on. But each time we do this, we miss an opportunity to develop judgment, build understanding, and strengthen our ability to navigate novel situations.
Our Approach
We design AI interactions that make you think harder, not less. The Dojo creates "productive friction" — challenges that force active engagement while still leveraging AI's capabilities.
The goal is symbiosis: humans and AI each contributing what they do best. AI brings knowledge and pattern-matching; you bring judgment, creativity, and accountability.
Design Principles
Judgment over Execution
In the age of AI, knowing what to do matters more than doing it fast. We optimize for decision quality, not speed.
Questions over Answers
The Sensei guides through inquiry. Good questions develop thinking; quick answers can short-circuit it.
Productive Struggle
Some difficulty is necessary for learning. We don't rush to resolve every challenge — we help you work through them.
Metacognitive Awareness
Thinking about your thinking. The Dojo makes cognitive processes visible so you can improve them.
Who We Are
Dr. Sathya Narayanan
Professor of Computer Science & Director
Professor of Computer Science at California State University, Monterey Bay. Director of the Computing Talent Initiative, focused on developing AI-native approaches to computing education.
snarayanan@computingtalentinitiative.orgComputing Talent Initiative
California State University, Monterey Bay
An institute dedicated to reimagining computing education for the AI age. We research and build tools that develop human capability rather than replace it.
computingtalentinitiative.org →Transparency
Open Source
The Dojo is fully open source under the MIT license. You can inspect every line of code, verify our privacy claims, and contribute improvements.
View on GitHub →Privacy Architecture
We designed the Dojo so we cannot read your conversations. With a personal Gemini or Groq key, your key stays in your browser and messages go directly to that provider — our servers only serve the application code. Institutional CTI keys are the one case where a request passes through our backend, and that backend records token counts for budgeting, never conversation content.
Learn about our privacy architecture →No Hidden Agenda
We're not building this to collect data, train models, or drive ad revenue. We're educators who believe AI literacy is one of the most important skills for the coming decades, and we want to give students the right tools.
Frequently Asked Questions
Is the Dojo really free?
Yes. The Dojo itself is completely free and open source. You bring your own API key from either Groq or Google — both have generous free tiers. Most users never pay anything for API usage.
What AI models does it use?
Three providers. Google Gemini uses Gemini 2.5 Flash (~15 requests/min, ~20 free requests/day). Groq uses Llama 3.3 70B (~14,400 free requests/day — better for extended practice). Institutions running the CTI backend can issue students CTI Program keys, which use Anthropic's Claude Sonnet against a coordinator-managed token budget. You can switch providers anytime in Settings.
Is my data private?
With a personal Gemini or Groq key, yes — the Dojo runs in your browser and your conversations go straight to that provider, so their privacy policy governs what they receive. We never see them. If your institution issued you a CTI key, requests are proxied through our backend to Anthropic so your coordinator can manage token budgets; that backend records token counts and never conversation content. In every case, we do not store your conversations.
Is anything saved when I close the browser?
Nothing is saved to our servers. Some things are saved in your browser so you can pick up where you left off — your API key, Practice Dojo session progress, Architect Studio runs, and any belts you've earned in the Code Kata Dojo. A free-form chat conversation is not persisted: close or refresh and it's gone. Use Export, or ask @reflector for a summary, if you want a record.
Can I use it offline?
No, you need an internet connection since the AI processing happens through your chosen provider's API. However, since it's a web app, you can use it on any device with a browser.
How is this different from ChatGPT or Claude?
Standard AI chats optimize for giving you answers quickly. The Dojo optimizes for developing your thinking ability. It uses structured frameworks, metacognitive coaching, and Sparring Partners to challenge your reasoning rather than just providing solutions.
Can I use this for my class?
Educators can have students get their own API keys and use the Dojo for assignments. Students can generate session summaries using @reflector to share their thinking process without exposing raw conversations. See our Educators page for implementation ideas.
How long is a typical session?
It depends on the activity. A quick Practice Dojo pathway runs 12-15 minutes; the guided Symbiotic Thinking foundations topic runs 25-35; Career Intelligence sessions run 30-45; and the paired Project Interview activity is built for a full ~60-minute class. Free exploration is as long as you want it to be.
What is Symbiotic Thinking, in one sentence?
Critical thinking, communication, and collaboration — practiced with AI in the room. Those three abilities are what the Dojo is built to develop; everything else (the 3Cs, DIKW, UMPIRE, the Sparring Partners) is method in service of them.
Want to contribute or connect?
We're building this in the open and welcome collaborators.