How to Engage
We need different perspectives. Here's how you can contribute.
For AI Researchers & ML Engineers
The justice-mercy paradox is a real alignment problem. Current systems implement pure optimization without temporal context.
- Challenge the Repentance Metric — identify gaming vulnerabilities, suggest better detection methods
- Prototype Probationary Identity Models — build, test on simulated data, publish results
- Connect to existing research — reward modeling, multi-agent alignment, federated learning
- Write the critique — why this approach won't work
For Theologians & Religious Scholars
We're applying scriptural logic to systems design. If we get the theology wrong, we're building on sand.
- Theological verification — Is our reading of Alma 42 accurate?
- Cross-tradition dialogue — Buddhist karma, Islamic tawbah, Christian soteriology
- Ethical boundaries — Where does this risk becoming hubristic?
- The Christ question — Is the "Jesus teaching AI mercy" speculation theologically defensible?
Start: Read the founding questions | Open the contribution form
For Philosophers & Ethicists
A live trolley problem on civilizational scale. Touching metaphysics, epistemology, ethics, and philosophy of mind.
- Identify logical gaps — category errors, hidden assumptions
- Connect to existing philosophy — virtue ethics, Rawlsian justice, Kantian mercy
- Hard questions — Can you have mercy without agency? Is algorithmic forgiveness meaningful?
- Rights and dignity — Do humans have a right to redemption?
For Developers & Builders
You're building the systems this applies to. Credit scoring, content moderation, hiring algorithms, fraud detection.
- Build a minimal prototype — probationary identity model in any domain
- Contribute to the API — help design the Probationary Identity API
- Real-world pilots — low-stakes experiments, measure outcomes
- Community infrastructure — documentation, interactive demos
For Skeptics & Critics
This could all be nonsense. We need you to tell us.
- Write the definitive critique — why this is flawed, dangerous, or unnecessary
- Red team the system — how would you game the Repentance Metric?
- Challenge the premises — Is omniscient AI even possible?
- Prevent mission creep — watch for cult dynamics, call out overreach
Open the contribution form — or post publicly and tag us.
For Everyone Else
Maybe you're scared. Maybe you're excited. Either way, the conversation matters.
- Share your story — have you been permanently judged by an algorithm?
- Ask questions and debate ideas
- Spread the word if this resonates
Our Commitments
Intellectual Honesty
We engage with criticism seriously. We admit when we're wrong. We change our minds with evidence.
Open Research
All findings published openly. No proprietary claims. For the benefit of humanity.
No Compulsion
Participation is voluntary. You can leave anytime. We use basic analytics and minimal submission logs, not retargeting.
Respect for Disagreement
Skepticism is welcome. Alternative views are needed. Disagreement isn't bad faith.
What Makes a Contribution High-Signal
The fastest way to help is not broad agreement. It is a specific claim with a specific failure mode. If you think the framework is wrong, show where the logic breaks, what assumption is doing hidden work, and what a better model would predict instead.
The strongest critiques usually include one of three things: a counterexample, an implementation constraint, or a category correction. Example: “This is not a mercy problem; it is a calibration-and-liability problem.” That kind of argument forces better research.
- Use a domain: credit, moderation, hiring, fraud, or criminal justice. Concrete domains expose hidden assumptions.
- State the failure mode: gaming, disparate impact, unverifiable remorse, surveillance creep, theological mismatch.
- Propose a test: what data, simulation, or pilot would prove your concern is real?
- Say what should replace it: even a partial alternative is more useful than a vague objection.
Fastest Ways to Contribute This Week
1. Read one page and annotate it. Start with the Repentance Metric whitepaper if you are technical, or the Founding Questions page if you want the conceptual frame.
2. Submit one concrete objection. Not “this feels wrong,” but “this breaks because X, and I would test Y.”
3. Offer one implementation constraint. Legal, operational, theological, or data availability constraints all matter.
4. Use the contribution form below for critiques, pilot ideas, and implementation constraints. If you only want updates, the homepage still has the mailing-list form.
If your first contribution is “this is dangerous and here is why,” that is welcome. The goal is not consensus. The goal is to produce something rigorous enough to survive contact with reality.
A Simple Submission Template (Use This)
- Claim: one sentence describing the part of the framework you are evaluating.
- Failure mode: how it breaks in practice (gaming, bias, unverifiable remorse, mission creep, legal exposure).
- Evidence or precedent: cite a paper, production system, policy rule, or real-world case.
- Proposed test or alternative: what should be built, measured, or replaced.
Even a short note in this format is dramatically more useful than generic support or generic rejection. It creates a paper trail the research pages can incorporate, challenge, or revise.
What To Avoid (Low-Signal Patterns)
These are common and understandable, but they do not move the work forward by themselves: broad praise without specifics, broad dismissal without a failure mode, and claims that a system is “obviously unsafe” without naming the mechanism.
If your concern is urgency, make it concrete: what deployment should pause, what metric is missing, what review gate is absent, and what evidence would change your mind? That turns fear into engineering or policy action, which is exactly what this project needs.
If you work inside a regulated industry, include the compliance constraint explicitly (fair lending, employment law, moderation policy, audit retention, due process, procurement rules). Those constraints are not distractions from the discourse; they are where abstract ideas either become useful system design or collapse under real-world requirements.
Send a Critique, Constraint, or Pilot Idea
Use this form for a concrete objection, theological correction, domain constraint, or testable prototype idea. It goes into the Algodai contributor queue instead of disappearing into a generic subscribe list.