ken libbrecht
19:29Glitch Tokens - Computerphile
353K views • 3 years ago
Computerphile
Rob Miles explains "glitch" tokens, odd strings in large language model vocabularies that make the model produce gibberish. He covers how they likely arise from tokenizer training data, such as Reddit usernames, that never appeared in the model's own training.
57:03HACKENBUSH: a window to a new world of math
562K views • 5 years ago
Owen Maitzen
Owen Maitzen's animated hour-long introduction to Hackenbush, a simple combinatorial game, and how its strategy leads to surreal numbers and the roots of the number system.
29:46The REAL Reason McDonalds Ice Cream Machines Are Always Broken
13.6M views • 5 years ago
Johnny Harris
Johnny Harris investigates why McDonald's ice cream machines are so often broken, covering the Taylor machines, the repair restrictions on franchisees and the Kytch workaround device.
18:49The Snowflake Myth
10.6M views • 4 years ago
Veritasium
Veritasium visits snowflake expert Dr Ken Libbrecht, who grows designer snowflakes in the lab. The video explains how snowflakes form and why they take so many shapes.
2:45Two Is the Wrong Number of Days Off
14.1M views • 6 years ago
CGP Grey
CGP Grey argues that the two-day weekend is poorly structured and proposes taking days off midweek, such as Wednesday and Sunday, instead.
42:29YouTube's copyright system isn't broken. The world's is.
9M views • 6 years ago
Tom Scott
Tom Scott's long explainer on YouTube's copyright system, Content ID and the DMCA. He argues the platform's system is a workaround for a broader copyright system that does not fit the modern world.
4:14Computer Color is Broken
4.3M views • 11 years ago
minutephysics
MinutePhysics explains that computers store brightness nonlinearly, so blurring and blending colors in raw RGB gives wrong, dark results. It shows why mixed colors, such as in gradients, look off.
18:24The Uncanny Valley Is Wrong
236K views • 6 years ago
This Place
This Place argues that the uncanny valley is a flawed concept, since 'human likeness' is too broad and the supporting evidence is cherry-picked. It proposes that the creepiness really comes from how animated characters and robots move.
27:14Transformers, the tech behind LLMs | Deep Learning Chapter 5
11.6M views • 2 years ago
3Blue1Brown
3Blue1Brown's visual explanation of how transformers and GPT-style language models work, tracing data through tokens, embeddings and attention.
35:07These Stupid Trucks are Literally Killing Us
7.3M views • 3 years ago
Not Just Bikes
Not Just Bikes explains how SUVs and light trucks have grown in size and share of US sales, and why that is dangerous for pedestrians, cyclists and city life. He covers the regulatory loopholes, such as fuel economy rules and the chicken tax, that favored them.
12:04Anagrams, but where you can break apart letters: "Anagraphs"
256K views • 8 years ago
suckerpinch
Tom Scott-style exploration by suckerpinch (Tom7) of generalized anagrams where letters can be broken into pieces, called anagraphs, with a whimsical computational approach.
3:45The Dutch Solution for Safer Sidewalks - Continuous Sidewalks
452K views • 6 years ago
Not Just Bikes
A short Not Just Bikes video on continuous sidewalks in the Netherlands, where the sidewalk runs unbroken across side streets so cars must visibly cross pedestrian space. It places them within Dutch sustainable safety and Vision Zero efforts.
12:44Rapping, deconstructed: The best rhymers of all time
16.3M views • 10 years ago
Vox
Vox breaks down how great rappers build rhymes, from Kurtis Blow's simple bars to Biggie's smooth flow and Kendrick Lamar's dense, fast delivery. It uses on-screen visualizations of rhyme structure.
7:09What Is the RNA World Hypothesis?
886K views • 10 years ago
Stated Clearly
Stated Clearly explains the RNA world hypothesis, which resolves the chicken-or-egg problem of genes versus enzymes by proposing that RNA once did both jobs. It outlines how evolution could begin with RNA chains alone.
10:37The Bayesian Trap
4.5M views • 9 years ago
Veritasium
Veritasium explains Bayes' theorem with examples and introduces the 'Bayesian trap', where repeated events get mistaken for inevitable ones. It discusses how priors shape our expectations and behavior.
9:27What is a Geometry Kernel? C3D vs ACIS vs Parasolid
25K views • 4 years ago
nickkallen1
The Plasticity creator explains what a geometry kernel is and compares commercial options C3D, ACIS, and Parasolid, describing his months of research and why he chose as he did.
7:19Cyclists' hearts: can you be so fit that you die?
2.2M views • 8 years ago
Medlife Crisis
Medlife Crisis examines why some elite endurance athletes die in their sleep, covering extremely slow heart rates and the effects of performance-enhancing drugs that thicken blood. It uses cyclist Marco Pantani as an example.
13:52Ch(e)at GPT? - Computerphile
617K views • 3 years ago
Computerphile
Computerphile video in which Mike Pound explains a University of Maryland paper on watermarking LLM output. The method biases token choices so generated text can later be detected statistically.
9:33The British Museum is full of stolen artifacts
6.7M views • 6 years ago
Vox
A Vox explainer on how many of the British Museum's treasures were taken during British imperial rule, focusing on the Benin Bronzes and requests for their return.