interpolation
Reusing a melody from an existing song (re-recorded rather than sampled) within a new song, requiring songwriting credit.
As defined in Lex Fridman Podcast #492 - Rick Beato
Where it comes up · 15
- 248 | Yejin Choi on AI and Common Sense4 claims
- 280 | François Chollet on Deep Learning and the Meaning of Intelligence3 claims
- Francois Chollet | Why abstraction is the key to intelligence, and what we’re still missing2 claims
- The AI Reset Is Here: Search, Jobs, and Everything Else w/ Anish Acharya, Dave Blundin, Salim Ismail2 claims
- Adam Brown — Bubble universes, space elevators, & AdS/CFT2 claims
- Lex Fridman Podcast #492 - Rick Beato2 claims
- 184 | Gary Marcus on Artificial Intelligence and Common Sense2 claims
- Mustafa Suleyman: The AI Pioneer Reveals the Future in 'The Coming Wave' | Intelligence Squared1 claim
and 7 more
Most important claims · 10 of 25Search all claims for “interpolation” →
- When deep learning makes an error, it is not behaving randomly but learning a systematic, lawful rule that simply differs from how humans see the universe—because rules are underdetermined by data (like '2,4,6,8' could continue many ways), and deep learning's output-node-driven computation favors interpolation over the rule-seeking extrapolation that humans are built for.184 | Gary Marcus on Artificial Intelligence and Common Sense
- Financial models are mostly models rather than theories—glorified ways of interpolation that take an analogy between markets and something better understood (smoke, hydraulics) and make predictions on that basis; they work in a narrow range when the world doesn't change much, but in a crisis the model becomes fundamentally wrong and inapplicable as assumptions like price independence and market liquidity drop away when people panic.Emanuel Derman on Theories, Models, and Science
- Computing the probability of even 10-tuple sequences of language would require more data than there are atoms in the universe, which means language models must perform a kind of clever interpolation beyond mere statisticsCHATGPT won't ever be the same again after this
- Because LLM transformations are vector programs, you can average or interpolate between them—asking it to talk like a pirate and like Shakespeare simultaneously yields a Shakespearean pirate—which is something you cannot do with an explicit logic program, giving the impression of creativity through interpolation within the convex hull of training representations.280 | François Chollet on Deep Learning and the Meaning of Intelligence
- Because LLMs are big differentiable parametric curves fit to a data distribution, you cannot input symbolic rules valid for whole categories of problems; you can only supply individual data points the model interpolates between, so fixing errors requires labor-intensive pointwise patches that address one query at a time rather than entire categories of issues.280 | François Chollet on Deep Learning and the Meaning of Intelligence
- It is commonly claimed that large language models can only interpolate within their training data and not extrapolate or be creative like humans, but it is unclear whether humans actually extrapolate beyond their experience or also merely interpolate, raising the question of by what mechanism genuine extrapolation would occur.AMA | August 2024
- Using short music clips within commentary, interviews, or breakdown analysis qualifies as fair use, and YouTube Content ID claims against such use can be fought and won — Beato's lawyer fought 4,000 Content ID claims across his catalog and won every single one.Lex Fridman Podcast #492 - Rick Beato
- Large language models are generative only in a shallow sense—they mine the past and can never produce anything truly outside the dataset they were trained on; the human brain can mine novelty from the future, and there is a difference between genuine novelty and mere interpolation that current AI cannot cross.Lee Cronin
- Large language models only interpolate between existing points in vector space, whereas genetic algorithms can discover entirely novel points; combining an LLM trained on human art with a genetic-algorithm randomizing component raises the question of whether the result could be called creative.Ask Me Anything | July 2024
- When prompted, an LLM does not understand or simulate the query as a human would; it fetches a memorized program or interpolates between memorized programs and reapplies it to the input, which works well for familiar patterns seen many times in training but fails on unfamiliar inputs—and crucially the LLM has no way to tell the difference, since it does the exact same thing whether correct or completely off.280 | François Chollet on Deep Learning and the Meaning of Intelligence