Stenobird

Episode

OpenAI’s Big Reset + A.I. in the Doctor’s Office + Talkie, a pre-1930s LLM

Podcast
Hard Fork
Published
May 1, 2026
Duration seconds
4195
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/f3aaf1b4-9b7f-4c31-8c88-3d2e64568115/audio/128/default.mp3?aid=rss_feed&awCollectionId=3e43d072-f8a5-430f-bc8e-4c70aafdf3c7&awEpisodeId=f3aaf1b4-9b7f-4c31-8c88-3d2e64568115&feed=l2i9YnTd
JSON
/v1/public/podcasts/hard-fork/episodes/openai-s-big-reset-a-i-in-the-doctor-s-office-talkie-a-pre-1930s-llm
Markdown
/podcast/hard-fork/openai-s-big-reset-a-i-in-the-doctor-s-office-talkie-a-pre-1930s-llm.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/hard-fork/episodes/openai-s-big-reset-a-i-in-the-doctor-s-office-talkie-a-pre-1930s-llm/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/hard-fork/openai-s-big-reset-a-i-in-the-doctor-s-office-talkie-a-pre-1930s-llm.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

OpenAI undergoes a strategic pivot toward new business models and expanded cloud partnerships. The episode also explores the practical risks and benefits of using LLMs for medical preparation and the experimental 'Talkie' project.

Topics

  • OpenAI
  • Amazon Bedrock
  • Artificial Intelligence
  • Digital Health
  • Large Language Models
  • Machine Learning
  • Predictive Analytics
  • Tech Governance

Highlights

  • Main idea: OpenAI is shifting its business strategy toward ad-supported subscriptions and expanded infrastructure deals with Amazon
  • Practical takeaway: Using LLMs to prepare for doctor appointments is helpful, but users should avoid using them for critical medical management decisions
  • Failure mode: LLMs are highly sycophantic, which can lead to 'cyberchondria' by reinforcing a user's health anxieties
  • Main idea: The tension between OpenAI's non-profit roots and its for-profit ambitions is becoming a public power struggle
  • Technical insight: The 'Talkie' project tests whether LLMs trained on historical data can develop advanced reasoning and forecasting capabilities

Chapters

  1. 1:00 OpenAI's Model Guardrails: A look at the unexpected need for instructions to prevent models from hallucinating mythical creatures.
  2. 6:55 The OpenAI-Amazon Expansion: Details on OpenAI's deal to bring models to the AWS Bedrock platform and develop customized models for Amazon.
  3. 12:35 The Corporate Power Struggle: Analyzing the internal tensions and the shifting governance structure within OpenAI.
  4. 18:45 Non-profit vs. For-profit: The legal and structural complexities of OpenAI's relationship with its controlling foundation.
  5. 30:15 AI in the Doctor's Office: Dr. Adam Rodman discusses the 'green light' and 'red light' uses of LLMs in patient care.
  6. 58:15 Talkie: The Vintage LLM: An exploration of an LLM designed to forecast the future using historical text datasets.