corpus.blog
56,966 blogs · [ { "id": "01a0de3a-b09b-731f-b281-9cced64fe829", "title": "Writing in public is still underrated", "url": "https://newsletter.danielpaleka.com/p/writing-in-public-is-still-underrated", "published_at": "2025-12-01T05:55:30+00:00" }, { "id": "01a0de3a-b09b-731f-b281-9cced74befe6", "title": "Underrated reasons to be thankful: AI research edition", "url": "https://newsletter.danielpaleka.com/p/underrated-reasons-to-be-thankful", "published_at": "2025-11-28T05:29:27+00:00" }, { "id": "01a0de3a-b09b-731f-b281-9cced7711777", "title": "Even superhuman AI forecasters are only as good as your questions", "url": "https://newsletter.danielpaleka.com/p/even-superhuman-ai-forecasters-are", "published_at": "2025-11-27T07:52:41+00:00" }, { "id": "01a0de3a-b09b-731f-b281-9cced7e01824", "title": "What’s the deal with RL and forecasting?", "url": "https://newsletter.danielpaleka.com/p/whats-the-deal-with-rl-and-forecasting", "published_at": "2025-11-20T07:45:59+00:00" }, { "id": "01a0de3a-b09b-731f-b281-9cced8c85623", "title": "Contra Dwarkesh on RL sample-efficiency via information theory", "url": "https://newsletter.danielpaleka.com/p/contra-dwarkesh-on-rl-sample-efficiency", "published_at": "2025-11-18T22:24:03+00:00" }, { "id": "01a0de3a-b09b-731f-b281-9cced8e4bb3f", "title": "16 questions about reinforcement learning", "url": "https://newsletter.danielpaleka.com/p/16-questions-about-reinforcement", "published_at": "2025-11-18T02:30:36+00:00" }, { "id": "01a0de3a-b09b-731f-b281-9cced9ba1b47", "title": "Obvious ways RL can fail", "url": "https://newsletter.danielpaleka.com/p/obvious-ways-rl-can-fail", "published_at": "2025-11-15T05:16:24+00:00" }, { "id": "01a0de3a-b09b-731f-b281-9ccedaacc5d0", "title": "The two types of LLM preferences", "url": "https://newsletter.danielpaleka.com/p/the-two-types-of-llm-preferences", "published_at": "2025-11-10T07:44:57+00:00" }, { "id": "01a0de3a-b09b-731f-b281-9ccedac5de38", "title": "October 2025 AI safety news: Adaptive attacks, Tokenization, Impossible tasks", "url": "https://newsletter.danielpaleka.com/p/october-2025-ai-safety-news-adaptive", "published_at": "2025-11-07T02:20:31+00:00" }, { "id": "01a0de3a-b09b-731f-b281-9ccedb324013", "title": "You are going to get priced out of the best AI coding tools", "url": "https://newsletter.danielpaleka.com/p/you-are-going-to-get-priced-out-of", "published_at": "2025-11-05T04:48:57+00:00" }, { "id": "01a0de3a-b09b-731f-b281-9ccedb9c6b42", "title": "A/B testing could lead LLMs to retain users instead of helping them", "url": "https://newsletter.danielpaleka.com/p/ab-testing-could-lead-llms-to-retain", "published_at": "2025-11-02T03:20:19+00:00" }, { "id": "01a0de3a-b09c-7163-ae35-9e25830d70f2", "title": "Memetic optimization #1: brainrot", "url": "https://newsletter.danielpaleka.com/p/reflections-on-brainrot", "published_at": "2025-07-10T18:16:15+00:00" }, { "id": "01a0de3a-b09c-7163-ae35-9e25836ab440", "title": "March-April 2025 safety news: Antidistillation, Cultural alignment, Dark patterns", "url": "https://newsletter.danielpaleka.com/p/march-april-2025-safety-news-antidistillation", "published_at": "2025-05-16T15:42:07+00:00" }, { "id": "01a0de3a-b09c-7163-ae35-9e258432c4ee", "title": "GPT-4o draws itself as a consistent type of guy", "url": "https://newsletter.danielpaleka.com/p/gpt-4o-draws-itself-as-a-consistent", "published_at": "2025-03-31T07:02:23+00:00" }, { "id": "01a0de3a-b09c-7163-ae35-9e2584fa88f2", "title": "January-February 2025 safety news: Emergent misalignment, SAE sanity checks, Utility engineering", "url": "https://newsletter.danielpaleka.com/p/january-february-2025-safety-news", "published_at": "2025-03-09T16:01:56+00:00" }, { "id": "01a0de3a-b09c-7163-ae35-9e25859779de", "title": "You should delay engineering-heavy research in light of R&D automation", "url": "https://newsletter.danielpaleka.com/p/you-should-delay-engineering-heavy", "published_at": "2025-01-06T23:28:56+00:00" }, { "id": "01a0de3a-b09c-7163-ae35-9e2585fb6e4f", "title": "September/October 2024 safety news: Jailbreaks on robots, Breaking unlearning, Forecasting evals", "url": "https://newsletter.danielpaleka.com/p/septemberoctober-2024-safety-news", "published_at": "2024-10-31T19:06:11+00:00" }, { "id": "01a0de3a-b09c-7163-ae35-9e25864bdd87", "title": "July/August 2024 safety news: Tamper resistance, Fluent jailbreaks, Scaling limits", "url": "https://newsletter.danielpaleka.com/p/julyaugust-2024-safety-news-tamper", "published_at": "2024-08-31T19:00:18+00:00" }, { "id": "01a0de3a-b09c-7163-ae35-9e2587419603", "title": "May/June 2024 safety news: Out-of-context reasoning, Sparse autoencoders, Interpreting CLIP", "url": "https://newsletter.danielpaleka.com/p/mayjune-2024-safety-news-out-of-context", "published_at": "2024-07-01T09:11:32+00:00" }, { "id": "01a0de3a-b09c-7163-ae35-9e25876dc5ad", "title": "March/April 2024 safety news: Latent training, Emergent abilities, Instruction hierarchy", "url": "https://newsletter.danielpaleka.com/p/marchapril-2024-safety-news-latent", "published_at": "2024-04-30T22:35:50+00:00" } ] posts Claim your blog
Back to newsletter.danielpaleka.com
Blog · corpus.blog/blogs/newsletter.danielpaleka.com/posts

newsletter.danielpaleka.com

newsletter.danielpaleka.com

2025

Writing in public is still underratedoriginal ↗
1 Dec 2025
Underrated reasons to be thankful: AI research editionoriginal ↗
28 Nov 2025
Even superhuman AI forecasters are only as good as your questionsoriginal ↗
27 Nov 2025
What’s the deal with RL and forecasting?original ↗
20 Nov 2025
Contra Dwarkesh on RL sample-efficiency via information theoryoriginal ↗
18 Nov 2025
16 questions about reinforcement learningoriginal ↗
18 Nov 2025
Obvious ways RL can failoriginal ↗
15 Nov 2025
The two types of LLM preferencesoriginal ↗
10 Nov 2025
October 2025 AI safety news: Adaptive attacks, Tokenization, Impossible tasksoriginal ↗
7 Nov 2025
You are going to get priced out of the best AI coding toolsoriginal ↗
5 Nov 2025
A/B testing could lead LLMs to retain users instead of helping themoriginal ↗
2 Nov 2025
Memetic optimization #1: brainrotoriginal ↗
10 Jul 2025
March-April 2025 safety news: Antidistillation, Cultural alignment, Dark patternsoriginal ↗
16 May 2025
GPT-4o draws itself as a consistent type of guyoriginal ↗
31 Mar 2025
January-February 2025 safety news: Emergent misalignment, SAE sanity checks, Utility engineeringoriginal ↗
9 Mar 2025
You should delay engineering-heavy research in light of R&D automationoriginal ↗
6 Jan 2025

2024

September/October 2024 safety news: Jailbreaks on robots, Breaking unlearning, Forecasting evalsoriginal ↗
31 Oct 2024
July/August 2024 safety news: Tamper resistance, Fluent jailbreaks, Scaling limitsoriginal ↗
31 Aug 2024
May/June 2024 safety news: Out-of-context reasoning, Sparse autoencoders, Interpreting CLIPoriginal ↗
1 Jul 2024
March/April 2024 safety news: Latent training, Emergent abilities, Instruction hierarchyoriginal ↗
30 Apr 2024