explainer · HACKER NEWS RECOMMENDATION
Glitch Tokens - Computerphile
Computerphile · 2 HN recommenders · 2 recommendation months
Recommended by 2 Hacker News commenters across 2 months (2023–2023).
SignalTube analysis
Among 14 other explainer videos tagged LLMs & generative AI, the mean is 3.0 distinct HN recommenders. This video has 2 distinct HN recommenders, below that mean. Recommendations span 2 distinct months; the peer average is 2.5 months. This is a data-derived comparison, not a human review or quality rating.
29% of its 7 labeled comments are marked as recommendations or strong recommendations. HN mention years are 2023–2023; those years describe discussion history, not upload date. Its verified runtime is 19 minutes 29 seconds. YouTube lists its publication date as 2023-03-07.
- Channel / speaker
- Computerphile
- Verified runtime
- 19m 29s
- YouTube upload date
- 2023-03-07
- Hacker News mention years
- 2023–2023
SAME TOPIC AND FORMAT
Compare with related videos
The comparison cohort uses other explainer videos in LLMs & generative AI. “Shared recommenders” counts HN users who recommended both videos in the source graph.
| Video | HN recommenders | Recommendation months | Positive label share | Shared recommenders | Runtime | Upload date |
|---|---|---|---|---|---|---|
| Glitch Tokens - ComputerphileCurrent video | 2 | 2 | 29% | — | 19m 29s | 2023-03-07 |
| Transformers, the tech behind LLMs | Deep Learning Chapter 5explainer · 3Blue1Brown | 13 | 10 | 81% | No recorded overlap | 27m 14s | 2024-04-01 |
| Attention in transformers, step-by-step | Deep Learning Chapter 6explainer · 3Blue1Brown | 5 | 4 | 63% | No recorded overlap | 26m 10s | 2024-04-07 |
| Attention Is All You Needexplainer · Yannic Kilcher | 4 | 4 | 80% | No recorded overlap | 27m 7s | 2017-11-28 |
THREE SOURCE-BASED CONTEXT POINTS
Why Hacker News recommends it
Each note combines the dataset’s stance label and short rationale; comment bodies are not reproduced. Open the HN source for the original discussion.
The archive classifies this as a contextual reference in “What Reddit Got Wrong”. Its short rationale: Discusses glitch tokens that are Reddit usernames.
The archive classifies this as a contextual reference in a Hacker News discussion. Its short rationale: Cited to explain LLMs are statistical models, not sentient.
The archive classifies this as a recommendation in “Understanding GPT Tokenizers”. Its short rationale: Good introduction to glitch tokens.