20 talks · released on a rolling basis · free for 24h after each airs
The new playbook for sourcing venture deals from public engineering data.
Five days, twenty talks, the entire methodology behind the SSRN panel of 219 venture-backed startups, taught chapter by chapter by the data nerds who wrote it. Each talk free for 24 hours after it airs. Anonymous-by-design: no founder face, no founder voice, no real names. Synthetic narration, Remotion-rendered slides, the same pipeline that ships the Acceleration Watch.
20 talks, released on a rolling basis · register to get each one as it drops · €0 for 24 hours after each talk airs · All-Access Pass for lifetime replays + transcripts + 219-startup panel dataset · €97 one-time
Register your Free Pass
Talks air on a rolling basis. Register and we email you the moment each one unlocks, free to watch for 24 hours after it airs. No code required to follow along. Unsubscribe in one click. We use the same published methodology for the talks as for the dashboard, no marketing spin, no founder face, just the work.
By registering, you opt in to summit emails (one each time a talk unlocks) and the weekly Acceleration Watch digest. Both unsubscribe in one click. Anonymity-preserving, we never publish attendee lists.
The single belief the summit hinges on
If GitHub commit-velocity acceleration is the most leading public signal in venture capital, then every other deal-flow source - pitch decks, AngelList, Crunchbase, warm intros, is a lagging indicator.
The whole investing thesis falls or stands on whether that single belief is true. Day 1 of the summit makes it falsifiable. The next four days operationalise it.
The schedule
Day 1, Foundations · The Acceleration Era
2026-05-19
- All-Access
22 min · The Data Nerd
The Core Claim: Engineering Acceleration as a Leading Signal for VC
If commit-velocity is the most leading public signal in venture capital, every other deal-flow source is a lagging indicator.
- All-Access
18 min · Methodology Lead
Public Over Private: The CC BY 4.0 Methodology Walkthrough
If we can't publish the methodology, we don't deserve the price. Here's the walkthrough.
- All-Access
16 min · Quantitative Architect
From Rolodex to Regression: Killing Warm-Intro Roulette
The sourcing system that rewards proximity to the right rolodex is forty years old. Public data is the first credible challenge.
- All-Access
14 min · Signal Taxonomist
The Six Atomic Signals: A Glossary for the Acceleration Era
Commit velocity, contributor influx, repo expansion, star detachment, issue cadence, dependency adoption, defined and decision-ruled.
Day 2, Methodology · Inside the Engine
2026-05-20
- All-Access
20 min · Quantitative Architect
Computing Commit Velocity: The 14-Day Window Decision Rule
Why 14 days, not 7 or 30. Why two-period confirmation. Why z-score normalization beats raw counts.
- All-Access
16 min · Methodology Lead
The Bus-Factor Signal: Contributor Diversity as Risk Premium
If one person leaves and the org collapses, we mark it. Here's the formula and the decision rule.
- All-Access
18 min · The Data Nerd
Repository Expansion: New Repos as Leading Hiring Indicator
Companies create infrastructure repos before they hire the engineers who will fill them. Here's how we detect it.
- All-Access
19 min · The Data Nerd
False Positives, Honestly: The 38% That Doesn't Raise (And Why)
Some flagged organizations don't raise. Here's how we'll report precision and the false-positive breakdown honestly on the scorecard, and what it tells you about the signal.
Day 3, Sector Deep-Dives · Where the Signal is Hottest
2026-05-21
- All-Access
22 min · Sector Analyst
AI Infrastructure 2026: The 47-Day Lead Time on the Series A Wave
AI-infra Series A announcements are clustering around code-side acceleration spikes from 47 days earlier. Here are the tells.
- All-Access
18 min · Sector Analyst
Developer Tools: Why GitHub Sees Their Acceleration First
Dev-tools companies live on GitHub by definition. The signal is purer here than anywhere else.
- All-Access
16 min · Sector Analyst
Climate Tech: Engineering Velocity in a Capital-Heavy Sector
Climate-tech is the hardest sector to read on engineering signals, but the lead time is also the longest. Here's the calibration.
- All-Access
17 min · Sector Analyst
Fintech & Payments: Decoding Regulated Engineering Cadence
Fintech codebases are mostly closed. Public infrastructure repos are the tell.
Day 4, Operationalizing · From Signal to Term Sheet
2026-05-22
- All-Access
20 min · The Data Nerd
Building a Sourcing Workflow: From Weekly Digest to First Reply
A 15-minute Sunday workflow that turns the weekly digest into 3-5 outbound conversations a week.
- All-Access
16 min · The Data Nerd
ICP Engineering: Defining a Top-100 by GitHub Behavior
A top-100 ICP list, but defined by code-side behavior instead of demographic guessing.
- All-Access
18 min · Agent Architect
Agent-Native Sourcing: MCP, x402, and Per-Call Pricing
Autonomous agents source 24×7. Here's the protocol surface that makes them fast, cheap, and credible.
- All-Access
14 min · The Data Nerd
From Signal to First Reply: The 4× Reply-Rate Outbound Play
First-sentence proof, second-sentence thesis, third-sentence specific question. No deck. No calendar link.
Day 5, Future · The founder talk
2026-05-23
- All-Access
21 min · The Data Nerd
The Next Five Years: When Code Eats the Pitch Deck
Five years from now, the pitch deck is a confirmation document, not a sourcing artifact. Here's how that transition plays out.
- All-Access
16 min · Methodology Lead
Reproducibility as a Moat: Why We Open-Sourced Everything
Closed methodology compounds error. Open methodology compounds trust. Here's the case for radical transparency.
- All-Access
19 min · The Data Nerd
The Founder Talk: Falsifiable Predictions in a Hype-Driven Industry
Once a quarter, we make a single falsifiable prediction in front of the public record. Here's the next one.
- All-Access
20 min · The Data Nerd
Closing Keynote: A Manifesto for Data-First Venture Capital
Seven pillars, one named enemy, and the line we don't cross to grow. The summit closes here.
What's in the All-Access Pass
Each talk is free for 24 hours after it airs. After the window closes, every talk is locked behind the All-Access Pass, €97 one-time, lifetime access to all 20 talks, full transcripts, slide decks, and the 219-startup panel dataset that the panel was built on. No subscription, no expiration.
- ✓
20 talks, released on a rolling basis, synthetic narration, anonymous-by-design
€0 free for 24h after each talk airs
- ✓
Full transcripts of every talk (PDF + markdown)
€197 with All-Access Pass
- ✓
Slide decks and chart packs from every talk
€197 with All-Access Pass
- ✓
219-startup panel dataset, the same CC BY 4.0 file published free at /api/signals.csv; paying gets you the cleaned import-ready build and the talk walkthroughs
€297 standalone, included with All-Access
- ✓
Lifetime replays of every talk after the free window closes
€197 with All-Access Pass
- ✓
Methodology vault: SSRN preprint, Zenodo dataset, regression code
Free, always
30-day refund, no questions asked. Reply REFUND to any email.
FAQ
Why is everything anonymous?
Manifesto pillar #4, methodology over personality. The product is a dataset, not a personality. The work has to stand on whether the signal is real, not on whether the person delivering it is charismatic. Speakers are pseudonymous data-nerd roles - Methodology Lead, Quantitative Architect, Sector Analyst, Agent Architect, narrated by the same synthetic Cartesia voice that ships the Acceleration Watch.
How is each talk delivered?
Remotion-rendered slides + synthetic Cartesia narration. The same anonymous-by-design pipeline behind the weekly Acceleration Watch and the monthly founder talk on /state-of-github. Every talk has a full text transcript so you can read or skim if you'd rather not watch.
What's the catch with the Free Pass?
No catch. Each talk is free for 24 hours after it airs. After that the talk locks behind the All-Access Pass. The free window creates a real reason to attend live (or watch within 24 hours); the All-Access Pass funds the production cost of running the summit and producing the transcripts. We never gate the underlying methodology, that's published openly on /methodology under CC BY 4.0, free forever, and neither is the dataset: the panel itself is published free at /api/signals.csv under the same licence.
Can I get a refund on the All-Access Pass?
30 days, no questions asked. Reply REFUND to any email. We process inside two business days. The signal is reproducible from public data, we'd rather refund a buyer who didn't get value than retain a buyer who didn't.
See also: State of GitHub address · 12-min walkthrough · Manifesto · €7 First Look