Mind · Strategy
Parallelize
Run subtasks simultaneously and aggregate
Definition
Mechanism
Runs multiple LLM calls simultaneously and aggregates results. Two modes: Sectioning (split Task into independent subtasks, run in Parallel, merge) and Voting (run same Task multiple times, Aggregate via majority Mode or selection). Trades compute cost for speed and/or confidence. It spawns concurrent execution threads and employs Aggregate to unify the results into a final Result.
Preconditions
- Batch of tasks
- Multiple workers
Postconditions
- All results aggregated, or the unaggregatable ones reported as such
Invariants
- Independence: no subtask reads another subtask's output.
- Aggregation is chosen before fan-out, and subtasks are shaped to produce outputs it accepts.
Risks
Failure modes
Relationships
Builds on
Accepts
Infrastructure
The atomic unit of intent
Composes with
Infrastructure
Reduce a set to a summary statistic
References
Infrastructure
Structural processing error definition
Infrastructure
A discrete configuration of agency
Infrastructure
Concurrent flow
Infrastructure
The output of agentic transformation
Adaptive high-level planning
Used by
References
Infrastructure
Directed acyclic dependency graph
Society
Interlocking state threads
Reference
Signature
Parallel(Task)Aggregate(Result)