What Is Live Context?
Live context is decision context that is maintained continuously as events arrive: derived state — aggregates, counters, features, views — updates incrementally and converges in sub-second time, rather than being recomputed on a pipeline schedule or refreshed on a timer. A decision reading live context reads the world approximately as it is now, not as it was at the last batch run or cache refresh.
Updated
What is live context?
Live context is one of the three capabilities that define a Context Lake in Tacnode's vocabulary, alongside shared context and semantic context. It addresses the freshness half of the context gap: the pipeline lag between an event occurring and the derived state that decisions read reflecting it.
The defining mechanism is incremental propagation. Instead of periodically recomputing a derived value from scratch — the model of batch jobs and standard materialized-view refresh, whose cost and staleness grow with data volume — live context applies each change as it arrives, so maintenance cost is bounded by what changed and staleness is bounded by propagation time, not by a schedule. Incrementally maintained views are the canonical implementation, explained in incremental materialized views; the broader argument is the live context pillar and live context: the key that unlocks real-time AI.
Two honest boundaries keep the term precise. Live is not synchronous: derived state converges in sub-second time but is deliberately maintained asynchronously from the write path, to avoid coupling ingestion latency and lock contention to view maintenance. And live is not "always current": there is always some propagation interval; the claim is that the interval is sub-second and bounded, not zero.
How live context works
Compare the three maintenance models for one velocity counter:
- Batch recompute — a job rebuilds the counter hourly; staleness up to the full interval, cost proportional to all data scanned.
- Timer refresh — the counter's view is re-run every N seconds; staleness up to N plus query time, recompute cost paid whether anything changed or not.
- Incremental maintenance — each arriving event updates the counter directly; staleness equals propagation time (sub-second), cost proportional to the delta.
The incremental model is what makes narrow validity windows meetable for derived signals at all: a 200-millisecond authorization budget cannot be met by any schedule-based freshness, no matter how aggressive the schedule.
Why live context matters
Decisions inherit the age of the freshest context they can actually read. Models, rules, and agents evaluated against state seconds or minutes old are structurally blind to exactly the events that most need a response — the burst, the spike, the concurrent draw-down — because schedule-based derivation lags most when event volume is highest. Live context inverts that: propagation is event-driven, so the state decisions read tightens toward reality precisely when reality is moving fastest. For AI agents, whose multi-step plans compound on each read, the difference between schedule-fresh and live context is the difference between reasoning about the world and reasoning about a recording of it.
FAQ
Related terms
A Context Lake is real-time, multi-modal data infrastructure that gives AI agents and decision systems fresh, internally coherent context at decision time.
A context gap is the difference between the state of the world an automated decision acts on and the actual state at the moment the decision commits.
Derived context is state computed from events rather than stored directly — aggregates, velocity counts, features, embeddings — that decisions consume.
A validity window is the interval within which a decision's context remains an accurate basis for action — typically 10ms to 1s for automated decisions.
Shared context means every service and agent in a decision system reads one internally coherent snapshot of state, instead of private copies that diverge.
Semantic context is the meaning-bearing layer of decision context — aggregations, vector similarity, LLM-derived signals — computed from the same events.