Top all-time · #24939
Nicholas Andresen
Personal blog · nickandresen.substack.comaiphilosophy
Nicholas Andresen writes Nick's Notes, a personal blog at nickandresen.substack.com about ai and philosophy. Since 2026, 3 different Hacker News readers have linked it 4 times, most often “How AI Is Learning to Think in Secret”. It is still writing: the last post was on 2026-01-06, and you can follow it by feed.
Too few sampled comments (1) to say how often HN recommends it.?
Why HN links it
HN readers link this as a discussion of AI latent reasoning (neuralese, chain of continuous thought) and why such systems are hard to oversee with methods like RLAIF and constitutional AI. The one comment only cites it.
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.
I write about what I'm learning, reading, and thinking about.
Most-linked posts?
- How AI Is Learning to Think in Secret3 people · 4 mentions · first linked 2026-01-06
What Hacker News says?
different names: chain of continuous thought, latent reasoning, Latent Thought Trajectories, looped language models, neuralese the path isn't explored more aggressively because its not possible to apply any other selection pressure on such a machine other than just pure cold consequentialism. Specifically, its not possible to apply RLAIF + model spec (Constitutional AI) to stop the system from doing bad things when…
Comment linking /p/how-ai-is-learning-to-think-in-secret
on “Claude Opus 4.8”Submitted How AI Is Learning to Think in Secret as a story · 1 points
HN discussionSubmitted How AI Is Learning to Think in Secret as a story · 1 points
HN discussionSubmitted How AI Is Learning to Think in Secret as a story · 1 points
HN discussion
