macro-economic and venture-capital insights from the all-in podcast

A 50-basis-point opening cut historically signals the Fed sees recessionary pressure, not just a soft landing.

Since 1994 the Fed has opened six cutting cycles; only 2001, 2007 and the 2020 COVID emergency began with 50 bps. Markets fell 26-31 % in the two years following 2001 and 2007, while 2020 ripped 44 %—but 2020 had massive fiscal stimulus.

Why: Large initial cuts telegraph that the central bank is reacting to real-time weakness rather than preemptively fine-tuning.

"the only other two times where we've had a 50 basis point rate cut in modern history, it has been just before a recession — David Sacks"

AI will hollow out level-one customer support within two to three years because error tolerance and abundant training data align.

LLMs plus new voice APIs can be trained on product docs, email tickets and recorded calls; failures are acceptable at tier-one and can escalate exactly as human agents do today.

Why: Customer support is already tiered, so imperfect AI can substitute the first tier while escalating edge cases—minimizing blast radius.

"level one customer support is going to get replaced by AI — David Sacks"

Fund life must stretch to ~15 years because median time-to-liquidity has doubled since the 2010 era.

Companies that once exited in years 5-7 now take 11-13; without secondary sales most 2018-vintage funds show zero DPI by year seven.

Why: Longer gestation plus anti-M&A antitrust pressure compress IRR unless managers actively manufacture liquidity.

"they've always generated, if they've generated it at all, in years 11, 12 and 13 — Chamath Palihapitiya"

Deploying a fund in 18-24 months forfeits time diversification—the single best edge in venture.

Fast deployment cycles optimize for management fees and fund-raising velocity, not returns; small, patient funds (e.g., Benchmark, USV) preserve optionality to back late bloomers.

Why: Venture outcomes are power-law and time-uncorrelated; spreading investments across vintage years smooths exposure to macro shocks.

"when you have absolutely no time diversity in this business… you're giving away one of your best edges — Chamath Palihapitiya"

100 % accuracy in production is achievable for high-stakes workflows by chaining specialized models and deterministic guardrails.

An 8090 portfolio company now runs AI at 100 % accuracy for ten consecutive days in a regulated environment after iterating from mid-80 % to 100 % through engineering and chain-of-thought scaffolding.

Why: Error rates compound in regulated domains; engineering around model limitations (fail-over logic, deterministic checks) can drive observed error to zero.

"we've been running AI-powered software versus the old legacy deterministic solution… at 100 percent accuracy now for about 10 days — Chamath Palihapitiya"

Recession risk spikes when the yield curve re-steepens after inversion, not at the moment of inversion.

Short rates above long rates signal tight money; when the Fed slashes short rates to re-steepen the curve, it confirms they see real economic weakness.

Why: The curve inversion is a leading indicator; the de-inversion is the coincident signal that the Fed is reacting to realized slowdown.

"the recession doesn’t come when the yield curve inverts… it usually comes when the yield curve de-inverts — David Sacks"

Regulatory capture plus partisan retaliation can waste tens of billions even when superior private-sector solutions already exist.

$42 B rural-broadband and $7.5 B EV-charger programs have delivered zero connections and eight chargers after 1,000 days, while Starlink and private charging networks solved both problems faster and cheaper.

Why: Contract awards prioritize donor geography and ideological alignment over cost-effectiveness; media tribalism suppresses accountability.

"this is just pure… naked retaliation. The Biden-Harris administration doesn’t want to admit that Elon has the best solution — David Sacks"

Entry price is destiny: doubling valuations at constant exit multiples halves DPI.

Average VC fund targets ~2×; if 2021 vintage paid 2× historical entry prices, expected DPI falls to ~1× unless exit values also double.

Why: Venture returns are driven by multiple expansion and ownership percentage; overpaying at entry cannot be fixed by growth alone.

"if the entry prices were artificially double, then there goes your return right there — David Sacks"