Issue #90 - Chinese Whispers

February 1, 2025
Generated with FLUX 1 and ComfyUI

In This Issue

  • Editorial
    • Chinese Whispers.
  • What Caught My Attention This Week
    • When Capital No Longer Needs Labor, How Does Labor Gain Capital?
    • US Copyright Office declared AI does not undermine copyright.
    • OpenAI confirms GPT-5
  • The Tools of the Trade
    • OpenAI o3 mini doesn’t reason either.
  • The Way We Work Now
    • $2,500 to let AI train on your bestseller.

Editorial

Chinese Whispers

In Issue #89 – (Un)Reasonable Mass Delusion, I showed you how DeepSeek R1 only pretends to reason by mimicking the inner monologue of a human as he/she seeks a logical path to answer a question.

All it took was one prompt (but it’s easy to debunk the notion of reasoning by looking at other “reasoning” outputs published online).

You need a paid membership to read this content.