The questions the podcast didn't fully answer.
8 surfaced · 4 investigated by Zerve agents · 4 queued.
Will AI eliminate the seed-stage VC associate role entirely within five years?
For the Geeks · sources & method▾
Logistic regression with L2 regularisation
Outcome: prediction-resolves-true. Features: speaker base rate, topic, confidence stated, time horizon.
n = 1,847 historical predictions
- 20VC episode transcript (verbatim)Diarised + speaker-attributed via WhisperX + pyannote 3.1.
- SEC EDGAR — 10-K / S-1 filings ↗Revenue, segment data, risk factor disclosures.
- OECD AI Policy Observatory ↗Cross-country AI adoption + labour exposure indices.
Last agent run · 2026-06-08
Can amplification networks defend against open-weight commoditisation, or are they just another data moat illusion?
For the Geeks · sources & method▾
Structured expert elicitation (IDEA protocol)
4 independent domain experts, Delphi-style two rounds, aggregated via geometric mean of probabilities.
n = 4 experts · 2 rounds
- Acemoglu & Restrepo (2023) — Tasks, Automation & Wage Effects ↗NBER Working Paper 28257.
- Stanford AI Index 2026 ↗Compute, model performance, investment, training cost.
- US Bureau of Labor Statistics — OEWS ↗Occupational employment & wages, 2020–2025 panels.
Last agent run · 2026-06-08
What happens to LP capital flows if model-layer write-downs hit Q4 2026 as Hoffman implies?
For the Geeks · sources & method▾
Bayesian update over base-rate priors
Prior built from 10-year historical base rate; likelihood from cited evidence weighted by source reliability score.
n = 14 evidence units · 9 supportive, 5 contradictory
- US Bureau of Labor Statistics — OEWS ↗Occupational employment & wages, 2020–2025 panels.
- Harmonic.ai — Founder GraphFounder pedigree, prior exits, team formation signals.
- Goldman Sachs (2024) — Generative AI & Labour Markets ↗Sector-level exposure model.
Last agent run · 2026-06-08
If open-weight models hit frontier parity, does regulatory regime arbitrage become the dominant venture strategy?
For the Geeks · sources & method▾
Monte Carlo simulation (10,000 runs)
Inputs: inference cost decay curve, model release cadence, enterprise switching cost distribution.
10,000 iterations · 95% CI reported
- Epoch AI — Notable Models DB ↗Training FLOPs, parameter counts, release cadence.
- Federal Reserve FRED ↗Macro series: rates, M2, employment, GDP deflator.
- PitchBook — Global Venture Report Q1 2026 ↗Stage-by-stage deal counts, median round size, dry powder.
Last agent run · 2026-06-08
What is the right confidence interval on 'AI creates more jobs than it destroys' given this cycle's cognitive focus?
For the Geeks · sources & method▾
Logistic regression with L2 regularisation
Outcome: prediction-resolves-true. Features: speaker base rate, topic, confidence stated, time horizon.
n = 1,847 historical predictions
- SEC EDGAR — 10-K / S-1 filings ↗Revenue, segment data, risk factor disclosures.
- US Bureau of Labor Statistics — OEWS ↗Occupational employment & wages, 2020–2025 panels.
Last agent run · 2026-06-08
Can private markets remain attractive at current valuations without sustained ZIRP-era liquidity?
For the Geeks · sources & method▾
Logistic regression with L2 regularisation
Outcome: prediction-resolves-true. Features: speaker base rate, topic, confidence stated, time horizon.
n = 1,847 historical predictions
- Federal Reserve FRED ↗Macro series: rates, M2, employment, GDP deflator.
Last agent run · 2026-06-08
Will the emerging manager class of 2020–2022 ever raise Fund III?
For the Geeks · sources & method▾
Structured expert elicitation (IDEA protocol)
4 independent domain experts, Delphi-style two rounds, aggregated via geometric mean of probabilities.
n = 4 experts · 2 rounds
- PitchBook — Global Venture Report Q1 2026 ↗Stage-by-stage deal counts, median round size, dry powder.
- Artificial Analysis — Inference Pricing ↗Per-million-token price history across frontier APIs.
- Federal Reserve FRED ↗Macro series: rates, M2, employment, GDP deflator.
Last agent run · 2026-06-08
What does the right hybrid pricing model look like for AI-augmented enterprise software?
For the Geeks · sources & method▾
Counterfactual time-series (synthetic control)
Synthetic control built from weighted donor pool of comparable markets pre-2023.
Donor pool: 11 markets · MSPE ratio 8.4
- Acemoglu & Restrepo (2023) — Tasks, Automation & Wage Effects ↗NBER Working Paper 28257.
- OECD AI Policy Observatory ↗Cross-country AI adoption + labour exposure indices.
- Federal Reserve FRED ↗Macro series: rates, M2, employment, GDP deflator.
Last agent run · 2026-06-08