Top all-time · #23779
Andrew Lampinen
Personal blog · infinitefaculty.substack.comaiscience
Andrew Lampinen writes Infinite Faculty, a personal blog at infinitefaculty.substack.com about ai and science. Since 2025, 3 different Hacker News readers have linked it 4 times. It is still writing: the last post was on 2026-10-10, and you can follow it by feed.
Too few sampled comments (1) to say how often HN recommends it.?
Why HN links it
Andrew Lampinen's Substack, linked for a post on memorization versus generalization in LLMs. One commenter cites it on whether recitation of training data is a defect.
Summarised by Claude from a sample of 1 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Musings on natural and artificial intelligence, methods, and more.
What Hacker News says?
> LLMs do not encode nor encrypt their training data. The fact they can recite training data is a defect not a default. About this specific point, it is unclear how much of a defect memorization actually is - there are also reasons to see it as necessary for effective learning. This link explains it well: https://infinitefaculty.substack.com/p/memorization-vs-gener...
Comment linking /p/memorization-vs-generalization-in
on “Is legal the same as legitimate: AI reimplementation and the erosion of copyleft”Submitted What are the real problems of continual learning? as a story · 3 points
HN discussionSubmitted How cognitive science can contribute to AI: methods for understanding as a story · 2 points
HN discussionSubmitted Why isn't modern AI built around principles from cognitive science? as a story · 1 points
HN discussion