![Alignment Newsletter Podcast artwork](https://is4-ssl.mzstatic.com/image/thumb/Podcasts123/v4/7f/da/7f/7fda7fe7-1829-84e1-1b71-7121419e6e8a/mza_15623188066483683162.png/100x100bb.jpg)
Alignment Newsletter Podcast
122 episodes - English - Latest episode: almost 2 years ago - ★★★★★ - 5 ratingsThe Alignment Newsletter is a weekly publication with recent content relevant to AI alignment.
This podcast is an audio version, recorded by Robert Miles (http://robertskmiles.com)
More information about the newsletter at: https://rohinshah.com/alignment-newsletter/
Homepage Apple Podcasts Google Podcasts Overcast Castro Pocket Casts RSS feed
Episodes
Alignment Newsletter #173: Recent language model results from DeepMind
July 21, 2022 15:43 - 16 minutes - 9.55 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg HIGHLIGHTS Scaling Language Models: Methods, Analysis & Insights from Training Gopher (Jack W. Rae et al) (summarized by Rohin): This paper details the training of the Gopher family of large language models (LLMs), the biggest of which is named Gopher and has 280 billion ...
Alignment Newsletter #172: Sorry for the long hiatus!
July 05, 2022 12:29 - 5 minutes - 3.53 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg Sorry for the long hiatus! I was really busy over the past few months and just didn't find time to write this newsletter. (Realistically, I was also a bit tired of writing it and so lacked motivation.) I'm intending to go back to writing it now, though I don't think I can r...
Alignment Newsletter #171: Disagreements between alignment "optimists" and "pessimists"
January 23, 2022 19:03 - 14 minutes - 6.88 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg HIGHLIGHTS Alignment difficulty (Richard Ngo and Eliezer Yudkowsky) (summarized by Rohin): Eliezer is known for being pessimistic about our chances of averting AI catastrophe. His argument in this dialogue is roughly as follows: 1. We are very likely going to keep impro...
Alignment Newsletter #170: Analyzing the argument for risk from power-seeking AI
December 08, 2021 17:30 - 13 minutes - 9.09 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg HIGHLIGHTS Draft report on existential risk from power-seeking AI (Joe Carlsmith) (summarized by Rohin): This report investigates the classic AI risk argument in detail, and decomposes it into a set of conjunctive claims. Here’s the quick version of the argument. We will ...
Alignment Newsletter #169: Collaborating with humans without human data
November 24, 2021 17:30 - 15 minutes - 10.6 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg HIGHLIGHTS Collaborating with Humans without Human Data (DJ Strouse et al) (summarized by Rohin): We’ve previously seen that if you want to collaborate with humans in the video game Overcooked, it helps to train a deep RL agent against a human model (AN #70), so that the...
Alignment Newsletter #168: Four technical topics for which Open Phil is soliciting grant proposals
October 28, 2021 16:54 - 16 minutes - 7.19 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg HIGHLIGHTS Request for proposals for projects in AI alignment that work with deep learning systems (Nick Beckstead and Asya Bergal) (summarized by Rohin): Open Philanthropy is seeking proposals for AI safety work in four major areas related to deep learning, each of whic...
Alignment Newsletter #167: Concrete ML safety problems and their relevance to x-risk
October 20, 2021 16:00 - 17 minutes - 7.85 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg HIGHLIGHTS Unsolved Problems in ML Safety (Dan Hendrycks, Nicholas Carlini, John Schulman, and Jacob Steinhardt) (summarized by Dan Hendrycks): To make the case for safety to the broader machine learning research community, this paper provides a revised and expanded collecti...
Alignment Newsletter #166: Is it crazy to claim we're in the most important century?
October 08, 2021 16:30 - 15 minutes - 7.2 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg HIGHLIGHTS The "most important century" series (Holden Karnofsky) (summarized by Rohin): In some sense, it is really weird for us to claim that there is a non-trivial chance that in the near future, we might build transformative AI and either (1) go extinct or (2) exceed...
Alignment Newsletter #165: When large models are more likely to lie
September 22, 2021 16:30 - 16 minutes - 6.51 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg HIGHLIGHTS TruthfulQA: Measuring How Models Mimic Human Falsehoods (Stephanie Lin et al) (summarized by Rohin): Given that large language models are trained using next-word prediction on a dataset scraped from the Internet, we expect that they will not be aligned with what w...
Alignment Newsletter #164: How well can language models write code?
September 15, 2021 17:17 - 18 minutes - 7.52 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg HIGHLIGHTS Program Synthesis with Large Language Models (Jacob Austin, Augustus Odena et al) (summarized by Rohin): Can we use large language models to solve programming problems? In order to answer this question, this paper builds the Mostly Basic Python Programming (MB...
Alignment Newsletter #163: Using finite factored sets for causal and temporal inference
September 08, 2021 16:30 - 19 minutes - 8.44 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg This newsletter is a combined summary + opinion for the Finite Factored Sets sequence by Scott Garrabrant. I (Rohin) have taken a lot more liberty than I usually do with the interpretation of the results; Scott may or may not agree with these interpretations. Motivat...
Alignment Newsletter #162: Foundation models: a paradigm shift within AI
August 27, 2021 16:30 - 15 minutes - 6.63 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #161: Creating generalizable reward functions for multiple tasks by learning a model of functional similarity
August 20, 2021 16:30 - 17 minutes - 7.81 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #160: Building AIs that learn and think like people
August 13, 2021 16:30 - 17 minutes - 7.42 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #159: Building agents that know how to experiment, by training on procedurally generated games
August 04, 2021 16:30 - 27 minutes - 11 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #158: Should we be optimistic about generalization?
July 29, 2021 17:15 - 15 minutes - 6.72 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #157: Measuring misalignment in the technology underlying Copilot
July 23, 2021 16:56 - 14 minutes - 6.24 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #156: The scaling hypothesis: a plan for building AGI
July 16, 2021 16:30 - 14 minutes - 6.62 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #155: A Minecraft benchmark for algorithms that learn without reward functions
July 08, 2021 17:15 - 12 minutes - 5.72 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #154: What economic growth theory has to say about transformative AI
June 30, 2021 16:30 - 16 minutes - 8.53 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #153: Experiments that demonstrate failures of objective robustness
June 26, 2021 17:10 - 15 minutes - 8.7 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #152: How we’ve overestimated few-shot learning capabilities
June 16, 2021 16:00 - 14 minutes - 7.84 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #151: How sparsity in the final layer makes a neural net debuggable
May 19, 2021 16:00 - 11 minutes - 6.16 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #150: The subtypes of Cooperative AI research
May 12, 2021 16:00 - 12 minutes - 6.79 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #149: The newsletter's editorial policy
May 05, 2021 18:34 - 14 minutes - 7.28 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #148: Analyzing generalization across more axes than just accuracy or loss
April 28, 2021 16:00 - 21 minutes - 10.6 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #147: An overview of the interpretability landscape
April 21, 2021 16:00 - 13 minutes - 7.41 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #146: Plausible stories of how we might fail to avert an existential catastrophe
April 14, 2021 16:00 - 15 minutes - 7.77 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #145: Our three year anniversary!
April 07, 2021 16:00 - 13 minutes - 7.25 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #144: How language models can also be finetuned for non-language tasks
April 02, 2021 18:20 - 12 minutes - 5.57 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #143: How to make embedded agents that reason probabilistically about their environments
March 24, 2021 17:00 - 14 minutes - 6.4 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #142: The quest to understand a network well enough to reimplement it by hand
March 17, 2021 17:00 - 15 minutes - 6.96 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #141: The case for practicing alignment work on GPT-3 and other large models
March 10, 2021 17:00 - 16 minutes - 6.92 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #140: Theoretical models that predict scaling laws
March 04, 2021 18:00 - 19 minutes - 8.3 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #139: How the simplicity of reality explains the success of neural nets
February 24, 2021 17:30 - 22 minutes - 9.71 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #138: Why AI governance should find problems rather than just solving them
February 17, 2021 17:00 - 16 minutes - 8.07 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #137: Quantifying the benefits of pretraining on downstream task performance
February 10, 2021 17:00 - 15 minutes - 7.92 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #136: How well will GPT-N perform on downstream tasks?
February 03, 2021 17:00 - 17 minutes - 9.15 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #135: Five properties of goal-directed systems
January 27, 2021 17:00 - 15 minutes - 8.28 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #134: Underspecification as a cause of fragility to distribution shift
January 21, 2021 17:00 - 13 minutes - 7.14 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #133: Building machines that can cooperate (with humans, institutions, or other machines)
January 13, 2021 17:00 - 17 minutes - 9.09 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #132: Complex and subtly incorrect arguments as an obstacle to debate
January 06, 2021 17:00 - 17 minutes - 9.39 MBRecorded by Robert Miles: http://robertskmiles.com More information about the newsletter here: https://rohinshah.com/alignment-newsletter/ YouTube Channel: https://www.youtube.com/channel/UCfGGFXwKpr-TJ5HfxEFaFCg
Alignment Newsletter #131: Formalizing the argument of ignored attributes in a utility function
December 31, 2020 17:00 - 17 minutes - 8.1 MBRecorded by Robert Miles More information about the newsletter here
Alignment Newsletter #130: A new AI x-risk podcast, and reviews of the field
December 24, 2020 17:00 - 12 minutes - 7.07 MBRecorded by Robert Miles More information about the newsletter here
Alignment Newsletter #129: Explaining double descent by measuring bias and variance
December 17, 2020 02:19 - 13 minutes - 6.92 MBRecorded by Robert Miles More information about the newsletter here
Alignment Newsletter #128: Prioritizing research on AI existential safety based on its application to governance demands
December 09, 2020 17:00 - 18 minutes - 9.26 MBRecorded by Robert Miles More information about the newsletter here
Alignment Newsletter #127: Rethinking agency: Cartesian frames as a formalization of ways to carve up the world into an agent and its environment
December 02, 2020 17:00 - 22 minutes - 12.3 MBRecorded by Robert Miles More information about the newsletter here
Alignment Newsletter #126: Avoiding wireheading by decoupling action feedback from action effects
November 26, 2020 17:00 - 16 minutes - 9.71 MBRecorded by Robert Miles More information about the newsletter here
Alignment Newsletter #125: Neural network scaling laws across multiple modalities
November 11, 2020 17:00 - 14 minutes - 7.71 MBRecorded by Robert Miles More information about the newsletter here
Alignment Newsletter #124: Provably safe exploration through shielding
November 04, 2020 17:00 - 18 minutes - 9.41 MBRecorded by Robert Miles More information about the newsletter here