You run the same content check through several independent calls and block the content only if most agree. What is this called?
Why A is the answer
Parallelization with voting
Parallelization has two forms: sectioning splits work into independent parts, and voting runs the same task several times to compare the answers.
In detail
- Parallelization has two forms: sectioning and voting.
- Voting runs the same task several times and compares the answers.
- Voting costs a call per vote, so keep it for decisions that justify it.
Use case
An illustrative scenario, not a real incident.
A forum runs a content check three times. A post is blocked only if at least two calls flag it, which cuts false alarms from one odd result.
See it in different ways
mindmap
root((Patterns))
Prompt chaining
Fixed steps
Routing
Classify first
Parallelization
Sectioning
Voting
Orchestrator workers
Subtasks found at run time
Evaluator optimizer
Feedback loopmindmap
root(("Parallelization"))
n1["Sectioning"]
n2["Split into independent parts"]
n3["Voting"]
n4["Same task several times"]
n5["Compare the answers"]Key point: Voting runs one task several times and compares.
Why the other options are wrong
- B. Prompt chaining
- Chaining passes output from one step to the next.
- C. Orchestrator-workers
- Workers handle different subtasks, not the same check.
- D. Routing
- Routing sends an input down one path.

