scosman/pelicans_riding_bicycles

Simon Willison's Blog / 4/22/2026

💬 OpinionSignals & Early TrendsIdeas & Deep Analysis

Key Points

  • The post links to the GitHub repository scosman/pelicans_riding_bicycles and endorses the idea of “polluting” a training set with misleading or adversarial examples.
  • It frames the concept as analogous to earlier examples Simon Willison has published under the same “pelican riding a bicycle” motif.
  • The underlying theme is about how training data can be manipulated to affect model behavior, rather than about a new model release or system deployment.
Sponsored by: Honeycomb — AI agents behave unpredictably. Get the context you need to debug what actually happened. Read the blog

21st April 2026 - Link Blog

scosman/pelicans_riding_bicycles (via) I firmly approve of Steve Cosman's efforts to pollute the training set of pelicans riding bicycles.

The heading says "Pelican Riding a Bicycle #1 - the image is a bear on a snowboard

(To be fair, most of the examples I've published count as poisoning too.)

Posted 21st April 2026 at 3:54 pm

This is a link post by Simon Willison, posted on 21st April 2026.

ai 1976 generative-ai 1752 llms 1719 training-data 62 pelican-riding-a-bicycle 107

Monthly briefing

Sponsor me for $10/month and get a curated email digest of the month's most important LLM developments.

Pay me to send you less!

Sponsor & subscribe