Monthly Insider Drop · 1 live · 11 scheduled
The Sunday digest is the free cadence. The Dashboard is the daily tool. The Insider Drop is the thing that makes the paid tier a continuity programme — a brand-new sector deep-dive, methodology release, founder essay, or shipping tool on the first Tuesday of every month. Public gets the abstract. Members get the artefact.
The cadence rule
First Tuesday of every month, 09:00 UTC. Email to Insider list at the same time. Telegram thread opens for the drop the same morning.
Drops are dated, numbered, and pre-scheduled. The next twelve months are visible on this page. If a drop slips by more than 48 hours past its publish date, every Insider gets an automatic credit (one month free) — that’s how serious the commitment is to the cadence.
The four-format rotation
Variety is the point. A Methodology release answers a different question than a Sector Deep-Dive, which answers a different question than a Founder Essay, which answers a different question than a Tool Release. The four-month rotation means every quarter you get one of each.
Sector Deep-Dive
25-page PDF on one specific sector — top 25 ranked orgs, contributor maps, three pre-Crunchbase breakouts, raw CSV.
Methodology Release
Formal release of a methodology update — regression code, paper revision, falsifiability test, and the production-ready replacement for the prior version.
Founder Essay
4-6K-word analytical post-mortem from The Data Nerd. Looks back at the prior cycle, names what worked, what missed, and what changed in the engine.
Tool / Artefact Release
New shipping artefact — MCP tool, API endpoint, chart pack, or webhook template — ready to drop into a member's stack on the day it lands.
The twelve-month calendar
Live drops link to the full essay + member artefact. Scheduled drops show the topic + format so you know what’s coming. We commit forward to twelve months at a time so the cadence is auditable, not aspirational.
v2 of the Insider API exposes a GraphQL endpoint alongside the existing REST surface, plus bulk historical pulls of every watchlist + Sunday list since the engine's first publish.
Members get a GraphQL endpoint at /api/v2/insider/graphql with the full schema (Org, Sector, Period, Signal, Watchlist, ScorecardEntry). Plus bulk historical CSV pulls — every Sunday list, every watchlist, every grade — since the engine's first public publish.
Every weekly pick from 2026, graded at T+60 and T+90, with the lead-time distribution, the false-positive count, and the calibration boundary the year actually drew.
The annual post-mortem. Every Acceleration Watch pick from 2026, T+60 + T+90 grade, lead-time distribution, false-positive boundary, and the structural revision the founder is making for the 2027 methodology release.
v2 of the stage-inference model uses team-size velocity, investor-handle signals, and infra-repo creation patterns to separate Seed-stage from Series-A-stage acceleration.
Current stage inference is a single-class call — Pre-Seed / Seed / A / B / Later — based on team size + investor handle when available. v2 introduces a continuous-stage estimator that lets members filter the Insider list by stage-velocity-fit (e.g. 'Series A acceleration patterns inside what looks like a Seed team').
Top 25 ranked orgs in climate-side developer tools — energy modelling, grid simulation, MRV, carbon accounting, satellite data ingest, ESG SDKs.
January deep-dive on a sector most generalist VCs don't track as one — climate-side developer tools. Energy modelling SDKs, grid simulation runtimes, MRV pipelines, carbon-accounting infrastructure, satellite data ingest, ESG SDK layers. Top 25 ranked orgs.
v2 of the Insider webhooks adds custom threshold triggers, retry-safe JSON payloads, and Slack/Discord block-kit formatting templates.
v2 webhooks let members fire on custom threshold triggers — not just 'any acceleration above N percentile' but 'this specific watchlist crossed N for two consecutive periods'. Retry-safe payloads with idempotency keys. Pre-built Slack and Discord block-kit templates so the message renders cleanly without a recipient-side adapter.
The mid-cycle post-mortem. What the engine learned about itself in six months of production. What the cadence taught us. What the cohort taught us.
Six months in. The semi-annual post-mortem from The Data Nerd. The hits, the misses, the methodology revisions, the cohort behavioural patterns, and the one structural shift the founder is making to the engine for the next six months.
v3 of the contributor-quality filter introduces three new heuristics for distinguishing high-signal contributor influx from green-square gardening.
The current contributor-quality filter catches 91% of bot accounts and ~70% of low-signal mass-PR contributors. v3 adds three new heuristics — code-review depth, issue-thread participation, and historical commit pattern — that push the low-signal filter to ~88%. Members get the heuristic code, the test set, the regression results.
Top 25 ranked orgs in agent infrastructure — frameworks, runtimes, observability, evaluation, sandbox, memory layer.
September deep-dive on the agent-infra sector. Where the visible AI-agents conversation is in product layer, the engineering acceleration is two layers down — frameworks, runtimes, observability stacks, evaluation harnesses, sandbox primitives, agent-memory infrastructure. Top 25 ranked orgs with the full PDF + CSV bundle.
A new MCP tool wired into the existing @gitdealflow/mcp-signal package. Query the 24-hour Insider lead directly from your IDE, with API-key auth.
Members get a new tool exposed in the MCP server — query_insider_lead(sector?: string, limit?: number) — that returns the same Sunday list 24 hours before the public Monday publish, inside the buyer's IDE. Same auth model as the Insider API. Drop-in upgrade for any existing MCP host install.
A 5K-word post-mortem on the first 90 days of Insider watchlists. What hit, what missed, what the calibration boundary looked like in production.
The honest read on the first 90 days. We graded every Insider watchlist pick at T+60 and T+90. This essay walks the hits, the misses, the false positives that taught us most, and the one structural revision the founder made to the engine because of what the data showed.
Formal release of the v2 confirmation window — wider sample, sharper false-positive boundary, and the regression code to replicate it on your own panel.
The current two-period confirmation window filters out 73% of single-spike noise but also blunts the signal on genuine 14-day acceleration regimes that resolve in 12 days. v2 introduces a sliding-window confirmation against a contributor-quality threshold rather than a fixed 14-day pair. Members get the regression notebook, the panel code, and the falsifiability test.
Top 25 ranked orgs by 14-day commit-velocity acceleration, three pre-Crunchbase breakouts, and the contributor-influx pattern that distinguishes infrastructure from theatre.
The May 2026 inaugural Insider Drop. Verifiable compute is the engineering backbone of every credible AI-native devtools claim shipped this quarter — yet it sits at #11 in tracked sector mindshare and #2 in 14-day commit-velocity acceleration. This drop names the gap, ranks the 25 orgs leading it, and gives you the three pre-Crunchbase breakouts the engine flagged before any deck circulated.
Public — free forever
Members — Insider Circle (€97/mo) and above
The continuity gap this closes
A subscription that gives you the same dashboard every month is a tool subscription. A subscription that delivers a brand-new artefact every month — sector deep-dive, methodology release, founder essay, or shipping tool — is a continuity programme. The difference is whether you anticipate the next renewal or just tolerate it.
The Free Acceleration Watch is the cadence-builder. The Dashboard is the daily-rhythm tool. The State of GitHub is the public monthly anchor. The Insider Drop — this surface — is the paid-tier’s anticipation engine. First Tuesday of every month, something net-new lands.