MaxwellCalkin.
Building intelligence in service of the widening and deepening of what is genuinely life-giving.
A constitution
for not drifting.
I will live to participate in the unfolding of a better world.
A life is good to the degree that it generates and propagates goodness: within the self, within the family, within the community, within humanity, and within the universe.
“The unfolding” is the widening and deepening of what is genuinely life-giving. It is the conviction underneath my work, my family, and the way I choose to live.
Health
Sleep, strength, movement, nourishment, and emotional steadiness. My body is an instrument of service.
Family
A loving home and a present father. No outer victory compensates for inner betrayal at home.
Mission
Hard, meaningful work worthy of my gifts, directed toward what is genuinely good.
Depth × span
Depth is interior richness, integration, and wisdom. Span is reach, inclusion, capability, and care. The aim is to grow both, without sacrificing either.
The decision filter
- Does this protect health?
- Does this strengthen family?
- Does this serve the mission?
- Is it true, courageous, and beautiful?
- Does it propagate genuine goodness outward?
Read the thinking
Stages. Screens.
Systems.
Each chapter trained something the next one needed.
The bass years
Touring bassist. Radio City Music Hall, Broadway pits, orchestras, and recording studios. Music taught me what mastery feels like in the body.
A harder game
Competitive League of Legends, Valorant, and Fortnite. High ranks, hard-contested games, and a lesson that lasted: stimulation is not the same as aliveness. I needed to contribute.
Work that compounds
Software engineering at Interactive Aptitude, a DoD contractor funded by SBIR grants. Building real systems while moving toward work in AI safety and a future research lab.
Ready enough to begin
Software has been the bridge: the discipline of building real things in the real world. The next chapter is the one I was preparing for the whole time.
Build with
conscience.
Open-source tools exploring how to make intelligence safer.
Alignment is a multidimensional problem. Behavior, internal representations, shared understanding, and institutional oversight all matter. These apps and open-source tools are part of my effort to work across those perspectives.
BEACN
An experimental app for sharing and voting on visions of the future, exploring collective signals for AI alignment.
02YOUR VOICE MATTERS / FEATUREDHeard Us
A place to make your voice heard. One of my favorite projects, alongside BEACN.
03BEHAVIORAL EVALUATIONalignment-evals
Open-source work on measuring how AI systems behave under pressure.
04MODEL INVESTIGATIONalignment-probes
Tools for exploring questions of alignment inside models.
05INTERPRETABILITYinterpretability-toolkit
Work on understanding the structures behind model behavior.
06ADVERSARIAL EVALUATIONprompt-injection-benchmark
An open-source benchmark project focused on prompt injection.
07CIRCUIT EXPLORATIONllm-circuit-visualizer
A project for making model circuits more approachable to explore.
These are ongoing projects. Visit each app or repository for its current implementation, documentation, and limitations.
Explore the full build logIntelligence
toward wisdom.
Not just more powerful intelligence. Intelligence in service of what is genuinely life-giving.
My aim is to help ensure that as artificial minds grow in capability, they also grow in alignment with the deepest good available to us. An AI safety research and development lab is the horizon I am building toward.
Capability needs alignment.
The first duty of a serious AI program is to keep its outputs oriented toward the good at every level of capability.
Alignment is not obedience.
A perfectly compliant system is not necessarily a safe one. The target is wisdom, not servility.
Safety belongs to the whole system.
Inside the model, outside the model, around the model, and beyond the model: no single perspective is enough.
Build what deserves to exist.
I will resist participating in systems that addict, diminish, manipulate, or flatten human beings.
The long view
Whatever I build, I will build well. Respect detail. Favor substance over theater. Finish what matters. Prototype, test, refine, persist.
At this edge?
Let’s talk.
Good work begins with a real conversation.
I want to hear from people working on AI alignment, safety-motivated foundation model research, interpretability, integral theory, or anyone taking the widening of consciousness seriously.
Strong enough to be kind.
Ready enough to begin.
Notes from
the unfolding.
Six pieces on what I am trying to live, and why.
Essay links open the original published articles.