Distribution
A statistical notion describing the range of features present in an AI's training data; a system reliably recognizes new instances within that range but fails on inputs outside it (out-of-distribution)
As defined in Melanie Mitchell on Artificial Intelligence
Other definitions
In valuation analysis, the range of possible outcomes from a Monte Carlo simulation, showing the probability of different valuation results rather than a single number.
Valuation in Four Lessons | Aswath Damodaran | Talks at Google
Where it comes up · 25
- A conversation between Nassim Nicholas Taleb and Stephen Wolfram at the Wolfram Summer School 202111 claims
- #278 Peter Thiel6 claims
- Alex Rampell - Investing in Operating Systems - [Invest Like the Best, EP. 248]4 claims
- Amadeus: The IT Backbone of Travel3 claims
- Building the Third Place2 claims
- Thesis Driven Investing2 claims
- Gaurav Misra & Dwight Churchill - Building Captions - [Invest Like the Best, EP.405]2 claims
- Zach Dell - Powering the Future - [Invest Like the Best, EP.434]2 claims
and 17 more
Most important claims · 10 of 51Search all claims for “Distribution” →
- Most people do not shop for home insurance; they buy it as a side product of buying a home, because mortgages require proof of insurance, so Hippo built distribution through mortgage, title, real estate, and homebuilder partners.Assaf Wand - Innovation in Static Industries – [Founder’s Field Guide, EP. 36]
- Silicon Valley engineers are skeptical of advertising, but advertising works on nerds and on everyone else by embedding subtle impressions that drive later sales, and anyone who denies its effect on themselves is doubly deceived.#278 Peter Thiel
- Connectivity and transportation costs are the primary determinants of tail risk in economic distributions: as connectivity increases (fewer barriers to goods transport, fewer geographic monopolies), previously-local income distributions become winner-take-all (fat-tailed), transforming an opera singer's income distribution from thin-tailed (local monopoly) to fat-tailed (global competition where the perceived-best takes most income).A conversation between Nassim Nicholas Taleb and Stephen Wolfram at the Wolfram Summer School 2021
- In the digital world the scarce asset is distribution and audience, which are hard and expensive to build, so customer acquisition cost has become the new rent.Better, Cheaper, Faster: Why Companies that Reduce Friction Win
- Asset management, especially alternatives, consolidates around winners that deploy more resources, so mid-sized managers get squeezed and must find a niche or specialization to survive.Ted Seides - Investment Industry Paradigms - [Invest Like the Best, EP.390]
- A trained classifier can recognize new instances within the statistical distribution of its training data (a new photo of a dog), but cannot recognize out-of-distribution cases (a cartoon or painting of a dog) or transfer knowledge to a genuinely new domain—so what AI does is not learning in the human sense, since human learning implies using knowledge in novel situations.Melanie Mitchell on Artificial Intelligence
- P-values are random variables themselves; observing a p-value of 0.12 means the true parameter could produce p-values ranging widely, so approximately 53% of repeated experiments could yield p < 0.05 even though the true p-value is 0.12, making thresholds like p < 0.05 arbitrary.A conversation between Nassim Nicholas Taleb and Stephen Wolfram at the Wolfram Summer School 2021
- Fat tails emerge from a Gaussian distribution when volatility itself is variable; by constructing a mixture of Gaussian distributions with different variance parameters, the resulting composite distribution exhibits heavier tails and higher shoulders than a single Gaussian.A conversation between Nassim Nicholas Taleb and Stephen Wolfram at the Wolfram Summer School 2021
- Startups can build distribution faster than ever because AI-powered products feel magical, but incumbents with existing distribution that deploy AI are even harder to compete with, and their main constraint is organizational size.Ravi Gupta - AI or Die - [Invest Like the Best, EP.411]
- Inequality measured as a static snapshot (Gini index) is misleading; social mobility requires measuring the probability that someone in a high-income state exits it, or tracking whether the same families remain in top quantiles across generations, not comparing income distributions at two points in time.A conversation between Nassim Nicholas Taleb and Stephen Wolfram at the Wolfram Summer School 2021