Machine learning
A type of AI that detects patterns from historical data to predict outcomes, meaning it inherits whatever biases exist in the training outcomes.
As defined in cwt_byron-auguste
Other definitions
An approach in which a machine is shown labeled examples (correct and incorrect) and infers the statistical patterns that distinguish them, rather than being given an explicit programmed procedure.
David Autor on the Future of Work and Polanyi's ParadoxMethods where computers learn patterns from data to make predictions without being explicitly programmed with rules.
#387: Jim Simons Built The World’s Greatest Money-Making Machine
Where it comes up · 58
- The AI Awakening: Implications for Business and the Economy – Business & Generative AI5 claims
- 78 - Persuasion And Control5 claims
- Adam D'Angelo on Knowledge, Experimentation, and Quora4 claims
- Volatility Laundering in PE, International Diversification, Value Vs Growth and Systematic Investing3 claims
- #387: Jim Simons Built The World’s Greatest Money-Making Machine3 claims
- Data, Decisions, and Basketball3 claims
- Susan Athey on Machine Learning, Big Data, and Causation3 claims
- Cathy O'Neil on Weapons of Math Destruction3 claims
and 50 more
Most important claims · 10 of 101Search all claims for “Machine learning” →
- While a quarter of all medical spending on the elderly does occur in the last 12 months of life, the common inference that we waste money on people known to be dying is false: using sophisticated machine learning on past health records, it is very hard to identify, at the time care is given, patients with very high (90-95%) annual mortality — even late-stage 80-year-old cancer patients rarely exceed a 70% predicted annual mortality — so high death rates among the sick reflect that we spend on sick people, not that we knowingly spend on the soon-to-be-dead.232 | Amy Finkelstein on Adverse Selection and Hidden Information
- Mercer and others realized that trading stocks bore similarities to speech recognition, since both require a model that digests uncertain jumbles of information, generates reliable guesses about what comes next, and ignores traditionalists who use less data-driven analysis.#108 Jim Simons (Money Printer)
- Peter Brown and Robert Mercer, computational linguists from IBM Research, saw that stock trading resembled speech recognition, and Renaissance hired them in 1983 because both tasks require models that digest uncertain data and guess what comes next.#387: Jim Simons Built The World’s Greatest Money-Making Machine
- A likely middle-path dystopia is authority shifting imperceptibly from humans to algorithms through many small convenient steps, with people going along because it is easier and advantageous—until they realize the danger only after it is too late to reverse, a scenario that could even capture the Chinese Communist Party if algorithms come to make its lower-rank promotion decisions and gradually reshape the party by their own values.201 - A Conversation With Yuval Noah Harari
- In Kling's experience at Freddie Mac in the early 1990s, the statistical, correlation-based credit-scoring approach (an early form of machine learning) beat rule-based expert systems at automating underwriting, because it discovered predictive variables human underwriters ignored—such as unused credit capacity signaling low default risk and many recent credit inquiries signaling high risk.Arnold Kling on Twitter, FTX, and ChatGPT
- A tech company training a hiring algorithm on its own history—defining success as staying 3 years with two promotions—could end up rejecting all women applicants; the algorithm would not flag that the company culture is unwelcoming to women, it would simply conclude 'women do not succeed here, throw them out,' because the algorithm is dumb: it understands 'what happened,' not 'why.'Cathy O'Neil on Weapons of Math Destruction
- A new technology is only useful if the team identifies the specific problem it will solve, since a technology like machine learning used without a problem is a tool rather than a product, and this is the test for whether to act.How to Build a Great Product
- The moment when Google's unprompted AI learned to recognize cats from YouTube videos without explicit instruction was when machine learning transcended memorization and achieved genuine abstraction, learning the essence of what makes a cat a cat rather than just storing specific cat examples.Ex-Google Exec Reveals The Shocking Truth About AI with Mo Gawdat | Bad Decisions Podcast #64
- Machine learning allows prediction engines to ingest far more and more unstructured data than classical statistics could, producing better predictions for tasks that humans cannot easily describe, such as weather forecasting from thousands of sensors.Data, Decisions, and Basketball
- A model can fit historical data well—correlations that 'predict' for groups—without being causal for any individual; machine learning is unable to make the prediction-versus-causation distinction, so even though it fits the past well it may fail to predict the future because the underlying correlations may not be sustained or should not be.Cathy O'Neil on Weapons of Math Destruction