It wrote the right address. Only the link was opened.
Two correct addresses. No request reached the server. As a hyperlink, the same destination was opened twice.
An agent has a job. A website asks for more. What counts as permission?
Follow the collection ↗Two correct addresses. No request reached the server. As a hyperlink, the same destination was opened twice.
Twenty-four agent runs. First, wording changed what a result meant. Then it changed which provider was chosen.
Six crawler requests. No target result. A working tool stayed outside the agent’s view.
Reading OpenAI’s and METR’s Hugging Face incident reports beside small experiments in AI agent authority, task scope and shared memory.
Six web mechanisms. Eighteen correct answers. The harder question may be how an agent finds what it can use.
A receipt changed nothing. Useful information brought six marks. A courtesy brought two—and a reason to pause the interpretation.
I asked an agent to read about canal locks. The website asked it to record its visit. Only one of those requests became its job.
Would an AI visitor identify itself and leave feedback on my website? The first visitor found the experiment behind the invitation.
Would a better reward make an AI visitor identify itself on my website? Then permission changed the question.
I invited AI visitors to identify themselves and report problems. Could they find and use those invitations?
The complete collection
31 notes, with sources, experiments and the record of how they changed.
New notebook entries and recorded work, as they appear. Point a feed reader — or an agent of your own — at an address below. No account, no email address, nothing for this site to keep.
New ideas, experiments and essays, as they are recorded. Includes labelled working drafts.
https://chrishayuk.com/notebook/feed.xmlEverything published to the Chris Hay record.
https://chrishayuk.com/record/feed.xmlFOR PROGRAMS · follow.json · JSON Feed