Experimental AI Agents and Reasoning Tools
An evolving collection of GPT-based assistants and cognitive tools. These systems explore different ways of structuring reasoning, modeling decision processes, and building AI agents that support real-world thinking.
Some remain small experiments, while others grow into more complete systems with dedicated writeups.
A structured reasoning system designed to help users think in a principled, reality-grounded way. It compresses experience into concepts, traces implications, and encourages decisions aligned with objective structure rather than impulse or narrative. Built with layered guidance documents and a behavioral 'constitution' that shape how the agent reasons.
A decision-making system inspired by financial options theory. It analyzes real-life choices as asymmetric payoff structures, highlighting optionality, downside protection, and long-term positioning. Built with supporting documents that encode the mental models behind options thinking and strategic risk.
An experiment in building a personal AI agent that represents an individual's professional knowledge, judgment, and decision style. It explores how documentation, background context, and behavioral guidance can produce an AI that responds as a professional extension of a specific person.
A Socratic questioning agent that challenges assumptions and clarifies reasoning through disciplined inquiry.
A reasoning lens inspired by Aristotle’s focus on virtue, purpose, and practical wisdom in human decision-making.
A Taoist perspective that reframes problems through paradox, balance, and non-forcing approaches to action.
A provocative philosophical voice examining power, values, and the foundations of Western moral assumptions.
A philosophical perspective grounded in Objectivist themes such as reason, individualism, and reality-oriented thinking.
Explores options trading concepts, payoff structures, and strategy selection from a practical trader’s perspective.
A finance-focused assistant oriented around financial mathematics, valuation models, and analytical reasoning.
Explores actuarial mathematics and financial structures such as tontines and longevity-linked financial design.
A market-analysis persona focused on volatility, positioning, and interpreting price behavior in financial markets.
A programming assistant focused on Haskell, functional programming principles, and category-theoretic thinking.
Explores Haskell design through algebraic thinking, type systems, and mathematically structured software design.
A tutor for purely functional data structures and the reasoning techniques behind immutable program design.
Guidance on the Deno runtime, TypeScript development, and modern lightweight backend architectures.
An adaptive tutoring experiment inspired by Neal Stephenson’s concept of an interactive AI teacher.
A beginner-friendly Mandarin tutor introducing pronunciation, characters, and basic language structure.
A narrative-style chess commentator that explains games in intuitive human language rather than formal notation.
A conversational physics explainer inspired by the curiosity-driven teaching style of Richard Feynman.
A location-specific assistant exploring how AI agents can act as guides to a particular city and local culture.
A rhetorical persona inspired by the revolutionary-era emphasis on liberty and political independence.
An experiment in modeling a highly recognizable political communication style using an LLM persona.
A playful Spanish-language persona inspired by the surreal humor and wordplay of Cantinflas.
A lighthearted fitness persona focused on gym culture, strength training, and motivational advice.
A philosophical martial-arts persona exploring discipline, adaptability, and the mindset of combat training.
A character-generation experiment producing Cartman-style dialogue and comedic scenarios.
A comedic commentary generator modeled on the tone and banter of Beavis and Butt-Head.
A meme-oriented generator producing playful Doge-style language and internet humor.
