Skip to content

Columbia Business School

Columbia University's graduate business school in New York City, whose faculty research covers economics, finance, technology, and consumer behavior.

Current stories

leadershipOne report1 publisher

Columbia adjunct estimates Uber drivers in six of 16 cities earn below the federal minimum wage after costs

Columbia Business School's Len Sherman estimates Uber's median pay on active rides is $19.30 an hour in Orlando and $46 in Los Angeles. Uber, which has said drivers earn more than $30 an hour, calls the analysis flawed.

Reality

Evidence40
Adoption
Not scored
Hype gap−5
Incentives55
Confidence50
investConfirmed2 publishers

Columbia economist sets a $3.7 trillion annual revenue bar for the $10.3 trillion US AI buildout

Columbia's Stijn Van Nieuwerburgh estimates the US AI build at $10.3 trillion through 2032, needing $3.7 trillion a year in revenue to pay off. The bar is 37 times the roughly $100 billion OpenAI and Anthropic book today, a gap the model gives about seven years to close.

Reality

Evidence55
Adoption
Not scored
Hype gap+15
Incentives
Not scored
Confidence50
buildOne report1 publisher

Five agents per engineer makes review the only human contact with Daytona's code

Ivan Burazin told Business Insider that Daytona's best agent operators are former people managers, and the org he described runs about 80 agents behind 16 engineers who no longer write code themselves.

Publishers:runtimewire.com

Reality

Evidence42
Adoption20
Hype gap+30
Incentives80
Confidence58
leadershipOne report1 publisher

Edinburgh delivery riders are assembling their own evidence against pay algorithms

A rider-run charity with a decade of funding, a field test in Fife and one Deliveroo courier's spreadsheet are being pressed into service as evidence, because the party holding the pay logic is not the party being asked to explain it.

Publishers:theguardian.com

Reality

Evidence45
Adoption40
Hype gap+20
Incentives72
Confidence50
scienceOne report1 publisher

Digital twins built from 500-question profiles flatten the people they were meant to copy

LLM surrogates of more than 2,000 real respondents were wrong about a quarter of the time and landed roughly where a chatbot given demographics alone landed, with errors that pull toward stereotype.

Publishers:sciencenews.org

Reality

Evidence68
Adoption24
Hype gap−12
Incentives58
Confidence64