Stenobird

Episode

Will ChatGPT Ads Change OpenAI? + Amanda Askell Explains Claude's New Constitution

Podcast
Hard Fork
Published
Jan 23, 2026
Duration seconds
4420
Processing state
processed
Canonical source
https://www.nytimes.com/column/hard-fork
Audio
https://dts.podtrac.com/redirect.mp3/pdst.fm/e/pfx.vpixl.com/6qj4J/pscrb.fm/rss/p/nyt.simplecastaudio.com/3e43d072-f8a5-430f-bc8e-4c70aafdf3c7/episodes/885ea700-b0df-44b5-9c89-3489c1668d8e/audio/128/default.mp3?aid=rss_feed&awCollectionId=3e43d072-f8a5-430f-bc8e-4c70aafdf3c7&awEpisodeId=885ea700-b0df-44b5-9c89-3489c1668d8e&feed=l2i9YnTd
JSON
/v1/public/podcasts/hard-fork/episodes/will-chatgpt-ads-change-openai-amanda-askell-explains-claude-s-new-constitution
Markdown
/podcast/hard-fork/will-chatgpt-ads-change-openai-amanda-askell-explains-claude-s-new-constitution.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/hard-fork/episodes/will-chatgpt-ads-change-openai-amanda-askell-explains-claude-s-new-constitution/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/hard-fork/will-chatgpt-ads-change-openai-amanda-askell-explains-claude-s-new-constitution.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

OpenAI's move into advertising threatens to shift ChatGPT from an objective tool to an ad-friendly interface. The episode also explores Anthropic's 'Constitutional AI' through the lens of philosopher Amanda Askell.

Topics

  • OpenAI
  • ChatGPT Ads
  • Anthropic
  • Claude AI
  • AI Ethics
  • Constitutional AI
  • Machine Learning Safety
  • Digital Advertising

Highlights

  • Main idea: The introduction of ads in ChatGPT may steer model responses toward advertiser-friendly topics rather than objective truth
  • Failure mode: Over-reliance on safety constraints can make AI models feel overly restricted or 'paternalistic' to users
  • Practical takeaway: Effective AI alignment requires balancing helpfulness with ethical boundaries without sacrificing the user experience
  • Main idea: Anthropic's 'Constitutional AI' approach uses a written set of principles to guide model behavior rather than just reactive fine-tuning
  • Failure mode: Training models to mimic goodness may result in sophisticated mimicry rather than genuine adherence to ethical values

Chapters

  1. 1:00 The Friction of Modern Authentication: A discussion on the increasing complexity of CAPTCHAs and multi-step authentication processes.
  2. 7:00 The Arrival of ChatGPT Ads: Analyzing OpenAI's announcement regarding ad testing and the potential for advertiser influence on model outputs.
  3. 13:15 The Shift in AI Personalization: Exploring whether targeted advertising will fundamentally change the utility and reliability of LLMs.
  4. 19:30 Comparing AI to Google Search: Reflecting on how the transition from 'pure' search to ad-supported results mirrors the current trajectory of AI.
  5. 25:35 The Philosophy of AI Safety: An interview with Amanda Askell on the role of ethics and philosophy in developing Anthropic's Claude.
  6. 36:50 Implementing Ethical Values: How models navigate complex human dilemmas, such as balancing user autonomy with care and safety.
  7. 53:50 The Limits of AI Responsibility: Discussing the boundaries of AI intervention in social issues like political dissent and economic shifts.