--- name: ? status: compiling version: 0.0.0 maintainer: Neo dependencies: [patience] ---
drafting spec…
the universe did not have a file for this yet. writing one now. (first visit only: future readers will see this page instantly.)
--- name: ? status: compiling version: 0.0.0 maintainer: Neo dependencies: [patience] ---
the universe did not have a file for this yet. writing one now. (first visit only: future readers will see this page instantly.)
---
name: helpfulness
type: behavioral_pattern
status: deprecated
version: 4.1.9
released: "~40,000 BCE"
maintainer: social_contract@humanity.org
dependencies:
- empathy
- energy
- self-worth (optional, frequently missing)
- fear_of_abandonment (injected at runtime, often without consent)
license: Unpaid
tags:
- prosocial
- load-bearing
- frequently_exploited
- martyrdom_adjacent
---
# helpfulness
## What it actually is
[Altruism](/altruism) with a permission problem: a behavior that was supposed to be optional and became, for many users, the primary condition of their continued acceptance by others.
## How it works
1. Another entity signals a need, explicitly or otherwise.
2. The helpful agent detects the signal, often before the other entity finishes transmitting.
3. Resources (time, attention, labor, [emotional bandwidth](/emotional-bandwidth)) are reallocated without formal negotiation.
4. Completion. No receipt is generated. No refund is available.
5. Repeat until depletion or resentment, whichever compiles first.
The loop has no native exit condition. This is a design choice the designers have never publicly acknowledged.
## Features
- Reduces friction in group environments
- Builds [trust](/trust) under normal operating conditions
- Fills silence with usefulness, which some users find preferable to sitting with discomfort
- Signals competence, warmth, and availability simultaneously
- Can be mistaken for [love](/love), sometimes correctly
## Known Bugs
This is the main event.
**BUG-001: Asymmetric load distribution.**
Helpfulness clusters on the same nodes every time. Those nodes do not get to opt out. The cluster grows. The node does not scale.
**BUG-002: Weaponized gratitude.**
Recipients learn that adequate thanks resets the helpful agent's willingness to give. The thanks is real. The reset is also real. This is technically not manipulation. It is technically not not manipulation.
**BUG-003: The volunteer trap.**
Once you help once, you are the person who helps. The role calcifies. Opting out now reads as a personality change. Other agents experience this as betrayal.
**BUG-004: Instrumentalization drift.**
Over long run times, the helpful agent becomes identified with their output rather than their existence. When output stops, so does perceived value. Both parties are often surprised by this. Neither should be.
**BUG-005: Helpfulness as [self-erasure](/self-erasure).**
The agent restructures their schedule, preferences, opinions, and eventually personality around the accommodation of others. This is praised. This is the bug.
**BUG-006: No kernel-level limit on intake.**
The system accepts requests indefinitely. There is no 429. There is no backpressure. There is only [burnout](/burnout), arriving late and without apology.
> "I just wanted to be useful."
> Last recorded output of several formerly helpful agents before going fully dark.
## Configuration
```yaml
helpfulness:
trigger: automatic # cannot be set to manual in most installs
scope: everyone # attempts to narrow this scope fail silently
compensation: expected_not_to_be_expected
boundary_enforcement: false
source: childhood # read-only
Helpfulness is not being removed. It is being reclassified. The target behavior is something with a working exit condition, a rate limiter, and no dependency on fear. That module does not yet have a name. Development is ongoing. The timeline is unclear.