cavendish

Share graph · provenance running forward

OG

Ollie Grant

Product engineer · firm

What they dropped, which clusters it landed in, what got elected, and which experiments and findings descended from it. Everyone’s is visible to everyone, symmetrically. Show outcomes, not counts — no totals, no rankings, no rollups.

Dropped

34

signals with this person attached

Landed in

24

distinct clusters

Elected

21

of those, now fields

Descended

34

experiments and published items

Drops

What descended

recommendationTested

Use a structured episodic store with summarised recall, not a raw vector memory

recommendationTested

Route through a gateway you control; do not standardise on a vendor's

recommendationTested

Prompt caching: use for stable prefixes over 2k tokens; expect 30–45%, not 60%

recommendationTested

Every LLM judge ships with a human agreement score or does not ship

recommendationTested

Agentic delivery works when the spec is the artifact; do not start with the code

recommendationTested

Not yet: realtime voice for AU contact centres above tier-1 triage

recommendationTested

Distil to a small model only after the frontier baseline is measured on the same eval

recommendationTested

Put backpressure on agent fan-out before you put it on the model

recommendationTested

Agents act under delegated, scoped, expiring authority — never a service account

recommendationTested

Measure ROI from telemetry and cycle time, not surveys

recommendationAssessed

Give citizen developers a paved road and a retention policy, not a review board

standing answerTested

What does inference actually cost right now?

standing answerTested

Retrieval or fine-tuning for this?

standing answerTested

Which memory layer should a new agent use?

standing answerAssessed

What eval tooling do we use?

standing answerAssessed

When do we need to move to post-quantum crypto?

positionAssessed

The direction of agentic development

positionAssessed

Learning without weights: where continual learning actually lands

experiment · concludedValidated

Memory layer bake-off on a 40-session support corpus

experiment · concludedValidated

Cost-aware routing across three model gardens

experiment · concludedValidated

Token cost ledger across six client patterns

experiment · concludedRefuted

Judge calibration against human panel

experiment · concludedAbandoned

Voice agent latency floor for AU telco

experiment · measuring

Release canary suite v2

experiment · measuring

Edge SLM for in-store classification

experiment · measuring

Agentic QA on a regression-heavy codebase

experiment · running

Spec-first agentic delivery on a live internal repo

experiment · running

Nightingale measurement run: coding-assistant ROI

experiment · running

Delegated-authority broker for agent tool calls

experiment · running

Compiling agent experience into a persistent skill wiki

experiment · voting

Backpressure patterns for agent fan-out

experiment · voting

Temporal decision memory for a claims agent

experiment · voting

Shared-context AI surface for a delivery pod

experiment · proposed

Non-weight-bound learning via retrieval-updated skills

Follow this person’s finds

Following someone whose drops are consistently good is the internal version of the external voice watchlist, and often a better source than any detector.