At Uber, If You Don't Perform, You're Out

Jevons Paradox applied to tech

When a product becomes radically more convenient or cheaper, the market expands beyond how it was calculated. The old market math is the wrong denominator. "The size of Uber today is way beyond the original marketplace of black cars and or taxis. The company today is a result of Jevons paradox."

When to use: Sizing markets under technology transitions; evaluating adoption curves; pricing defensibility assessments.

When not to use: Mature, friction-saturated markets where physical or regulatory constraints cap expansion.

Attributed to: William Stanley Jevons (1865, coal) · applied to tech transitions by Dara Khosrowshahi via the Uber black-car → global mobility case.

Source-over-summary information discipline

When investigating an issue, go to the originator — not the manager's summary. Every layer of retelling degrades signal. Skip levels deliberately; random direct channels kept live regardless of level.

When to use: Diagnosing operational problems at scale; reset moments (turnarounds, crises); when team responses feel filtered.

When not to use: Routine reporting where summary is the product; small teams where the CEO is already primary filter.

Attributed to: Barry Diller (IAC) → Dara Khosrowshahi.

Transparency-as-self-defense (four-step)

Operating procedure for CEOs who want real data: (1) model honesty first — tell the team what's actually happening, including hard stuff. (2) Skip levels regularly. (3) Go to the source, not the summary. (4) Pick the mistake you'd rather make — too blunt or too soft? Decide explicitly.

When to use: Large orgs with filtered information; turnarounds; suspected good-news bias.

When not to use: Small teams where the CEO is the primary information synthesiser.

Attributed to: Dara Khosrowshahi, synthesised from Barry Diller's information discipline and applied across Expedia + Uber turnarounds.

They-won-we-lost-next loss framework

Three-move protocol for losses: (1) recognise and name it explicitly ("we lost"). (2) Analyse once to extract the learning. (3) Move on — don't paper over, don't obsess. Neither suppression nor post-mortem rumination.

When to use: Post-loss team communication; competitive defeats; failed bets; any moment where the team is tempted to either hide or wallow.

When not to use: Strategic debriefs that require structured analysis over days, not hours.

Attributed to: Barry Diller (IAC, after losing Paramount tender) → Dara Khosrowshahi.

What do the 9 million Uber drivers and couriers do when autonomous vehicles handle the majority of trips?

Based on: Dara forecasts ~20M autonomous vehicles on Uber in 15-20 years; 70-80% of intellectual work replaceable by AI in 10 years.

Why unresolved: Dara explicitly declines to answer: "I don't know." Platform is expanding into shoppers / AI-training work, but the velocity comparison with automation is unresolved.

Can national-scale retraining infrastructure be built fast enough to absorb 70–80% job displacement in 10 years?

Based on: Dara names the absence of retraining machinery as the gap. "That is a core capability I don't see any of our countries investing in."

Why unresolved: No current national-scale programme exists. AI itself may become the retraining vehicle, but that's unproven and timing-dependent.

Where will displaced workers find meaning when the job is gone?

Based on: Every UBI test has produced worse outcomes than the control group — Dara's reading. Work creates self-worth, not just income.

Why unresolved: Dara frames it as an open question with evidence that money-alone doesn't substitute for the meaning work provides. Steven links it to his father's own meaning-collapse after losing the family business.

When will AI learn in real time — and what changes when it does?

Based on: Current models learn in training, not during use. Dara names this as the gate keeping humans irreplaceable.

Why unresolved: Dara: "At this point I haven't seen a model that can learn in real time. When they can, that's the point at which I'm going to think we are all replaceable."

Is a non-hard-worker ever reshaped into a hard worker?

Based on: Dara and Brian Chesky (Airbnb) both say they have never seen it happen.

Why unresolved: Dara pauses mid-answer looking for a counterexample and cannot produce one — but acknowledges the sample is his own lifetime in elite orgs.