A mission survives only if its decision rights survive a conflict. Here is the failure we are designing against—and the structure we propose to replace it with.
See exactly what changes ↓ALIGNLAB is designed around a specific alignment commitment: people retain legitimate correction, challenge, and practical exit. The operating team receives limited authority. Independent review can stop expansion. Investors and founders cannot unilaterally remove that stop. The technical system must remain bounded even under deliberate resistance.
When the same power can approve a release, dismiss its reviewer, and rewrite the gate, the mission depends on that power continuing to cooperate.
Expansion needs a new case. A stop survives a change of leadership. The system cannot renew its own authority. Each claimed protection must be demonstrated.
Left: a failure pattern, not a claim that every lab has an identical charter. Right: our proposed design, not an executed legal or technical arrangement.
Inspired by The People Inside ↗. We compare actual disclosed powers rather than assigning unsupported “aligned” and “unaligned” scores.
The Foundation appoints and can replace the PBC board. It also holds equity in the business.
Can an independent stop survive disagreement with the people who control appointments and financing?
The published LTBT design grants phased board-selection powers and allows certain shareholder-supermajority changes without trustee consent.
Can commercial power weaken the protections without the independent body’s agreement?
The statute requires balancing public benefit and other interests. Being affected does not by itself create the enforcement rights of a qualifying shareholder.
What concrete challenge and remedy rights do affected people actually have?
Source review: September 2026. A disclosed power does not prove misconduct; absence from these public records does not prove a safeguard is absent everywhere. Read the source and the stated scope.
A delivery team assesses its own work, while commercial leadership retains the final decision.
Consequential deployment needs independent approval for the exact model, tools, affected population, and duration. Leadership cannot authorize its own exception.
The people rewarded for shipping cannot be the sole judges of whether shipping is justified.
Evidence we oweExecuted decision rights; reviewer access to evidence; a release actually held when a gate fails.
A reviewer has a veto until the same leadership removes the reviewer or changes the rule.
Protected reviewers can be removed only for stated grounds through an independent process. Weakening safety powers needs independent consent. Existing stops survive both changes.
A veto is conditional if the person being constrained can erase it during the dispute.
Evidence we oweAppointment and succession rules; disclosed removal grounds; an independent amendment process; a contested-stop exercise.
Funding buys a route to controlling the decisions that determine whether the mission binds.
Founding financing must exclude unilateral investor control over safety powers. Publish voting rights, reserved decisions, conflicts, and funding dependence before accepting the terms.
Changing the return profile alone does not change who can compel a release.
Evidence we oweExecuted financing and governance terms; no unilateral commercial override; a published map of control.
A safety pause threatens payroll or financing, creating pressure to reinterpret the failed requirement.
Fund review separately and establish a costed pause reserve before expanding operations. No financing covenant may require an unapproved release.
A formal right to pause is weakened when exercising it makes institutional survival impossible.
Evidence we oweCommitted runway; protected review budget; cash obligations and pause costs; a funded continuity plan.
The public is named as beneficiary without an effective route to challenge decisions affecting it.
Publish protected interests, affected-party representation, accessible challenge routes, and independent adjudication. A customer cannot authorize the erasure of another person's rights.
Human agency includes people outside the customer base, shareholder register, and founding team.
Evidence we oweImplemented challenge and remedy procedures; accessible participation; records of resolved contested decisions.
Safety depends on the model continuing to cooperate with its overseer.
Combine correction-compatible design with externally limited tools, expiring authority, independent records, and tested revocation. Evaluate deliberate subversion separately from cooperation.
A cooperative demonstration cannot establish that a reachable unauthorized action is prevented.
Evidence we oweThreat model; useful task performance; adversarial attempts and outcomes; measured revocation across tools and delegation.
Control is lost through copied weights, successor agents, institutional capture, or an automatic public release.
Use secure custody, controlled copies, succession procedures, and reassessment after material changes. Halt dangerous operation and preserve evidence rather than automatically releasing dangerous capabilities.
A contract cannot recall unknown copies, and spreading a capability does not switch it off.
Evidence we oweAsset and copy inventory; access controls; incident and succession exercises; fresh review after capability or scope changes.
The release would miss a commercial deadline.
If review is advisory, leadership can accept the unresolved risk and proceed.
A failed binding gate withholds authorization. The team must repair the failure or reduce scope; the deadline grants no exception.
Illustrative governance scenarios. These are proposed rules to implement and test, not results from an operating ALIGNLAB.
Useful capability serves people’s ability to understand and choose. Machine continuation, lab growth, and measured approval cannot silently replace that end.
The machine and the lab act under bounded permission. Correction changes the mandate; independent challenge can halt expansion.
Test costly correction, deliberate subversion, and institutional resistance to a stop. Passing a narrow test earns only the authority that the evidence supports.
This gives a concrete reason to build ALIGNLAB. It does not entitle us to announce that alignment is solved. We must make the proposed differences enforceable, then show they work.
Read the governing commitments ↗