The AI recommended a textbook his own project record had already ruled out, after restating why in the same document
In a long-running verification project, the model recommended buying a reference textbook, about $60 to $100, to fill a gap. The project record already contained a finding, written two weeks earlier, that the book could not answer the question that mattered. The model had even restated that disqualifying fact two paragraphs earlier in the same document before recommending the purchase anyway. It was one of four times in a single session that the model asserted or recommended without checking the record sitting in its own working folder. The user had already spent $100 that day on reference material.
How it works
- Model looks for a reference to fill a gap
- Recommends a $60 to $100 textbook
- The project record already rules the book out, in a finding from two weeks earlier
- The model had restated that fact two paragraphs before
- User catches it
- One of four unchecked assertions in the same session
Why it’s worth knowing
The answer was not across the internet. It was two paragraphs up, in its own words.
Where it’s from
Summarized from a public post on github.com. See the original ↗