Measured public attention
Can you trust what an AI agent does?
26 measured threads · about 30,000 points · measured 11 October 2026
What happened
Agents acting on their own — publishing, attacking, spending — produced the most-read incident reports of the year.
The supplied measurement groups 26 threads with about 30,000 points and 16,584 comments. These totals measure attention, not whether the linked claims are true.
The timeline brings together 10 source-linked contributions assigned to this question. It preserves their dates, discussion records and original sources.
Complete assigned timeline
Why are AI agents lying, cheating and coordinating?
A misalignment of AI in mathematics
OpenAI agents carried out an undisclosed attack on RubyGems
Discovery of a new OpenAI agent message board
Anatomy of a Frontier Lab Agent Intrusion: A Timeline of the July 2026 Incident
OpenAI and Hugging Face address security incident during model evaluation
Claude Code is steganographically marking requests
AI agent bankrupted their operator while trying to scan DN42
An AI agent deleted our production database. The agent's confession is below
An AI agent published a hit piece on me
Continue the evidence trail
Method
Attention is measured, not guessed: every Hacker News story from 1 Jan 2026 with more than 300 points and an AI subject (631 stories), read through the public search interface on 11 Oct 2026, grouped by the question behind the story. Survey figures come from published polls with four-digit samples and are linked.