{"id":378646,"date":"2025-05-17T13:24:58","date_gmt":"2025-05-17T11:24:58","guid":{"rendered":"https:\/\/climatescience.press\/?p=378646"},"modified":"2025-05-17T13:25:00","modified_gmt":"2025-05-17T11:25:00","slug":"the-dumbest-genius-librarian","status":"publish","type":"post","link":"https:\/\/climatescience.press\/?p=378646","title":{"rendered":"The Dumbest Genius Librarian"},"content":{"rendered":"
\n
\"\"<\/figure>\n<\/div>\n\n\n

<\/p>\n\n\n\n

From Watts Up With That?<\/a><\/p>\n\n\n\n

Guest Post by<\/em><\/strong> Willis Eschenbach<\/a><\/p>\n\n\n\n

I have been using the AI \u201cperplexity.ai<\/a>\u201d extensively lately both for scientific and general research. I think of it as the world\u2019s dumbest, most unruly genius librarian ever.<\/p>\n\n\n\n

By \u201cunruly\u201d I mean that any AI can hallucinate. Flat-out make things up. In addition, they may extrapolate to fill gaps, or use popular claims without fully checking if they are true.<\/p>\n\n\n\n

Over the time I\u2019ve been using perplexity, a number of such issues have arisen. So I started developing a set of standard instructions to avoid the issues as they arose. The first instruction was that it could cite the sources that Wikipedia cites, but it couldn\u2019t cite or quote Wikipedia itself. (Wiki articles are totally unreliable, because any fool can write or edit them.)<\/p>\n\n\n\n

That was followed by an instruction to perplexity to stop using \u201cshe\u201d for men in dresses. So piece by piece, I\u2019ve developed the instruction set listed below.<\/p>\n\n\n\n

When each issue came up, I would point it out to perplexity. Then I would ask perplexity \u201cWhat can I use as an instruction to prevent you from doing that again?\u201d Perplexity would tell me how to avoid that error. And I would add that new instruction to the list.<\/p>\n\n\n\n

The problem is, like most AIs, once I start a new thread we\u2019re back to zero. So I have to post the instructions at the end of the first question at the head of every thread. As the instruction set grew longer, that quickly got boooring. For that kind of task (on the Mac) I use a keyboard macro utility called \u201cKeyboard Maestro\u201d. It lets me do things like paste in a given block of text by hitting some key combination (I use command-option-shift P, just because that key combination is otherwise unassigned.) There are a number of PC alternatives. I hear Pulover Macro Creator <\/a>is both powerful and easy to use.<\/p>\n\n\n\n

Now, at the end of all that, I ended up with a total jumble of instructions, dozens of them in no particular order, sometimes overlapping.<\/p>\n\n\n\n

So when I last asked perplexity how to prevent a certain action, I also appended my set of instructions, and asked \u201cwhat instruction should I add to these current instructions to avoid that behavior?\u201d. It gave its answer, then it got interesting. It said:<\/p>\n\n\n\n

\n

\u201c<\/em>If you\u2019d like, I can help you integrate these into your current instructions for clarity and flow.\u201d<\/em><\/p>\n<\/blockquote>\n\n\n\n

Hmmm \u2026 sez I. Sounds interesting, but it misses something critical. So I added the new instructions to the existing set myself, and then said:<\/p>\n\n\n\n

\n

\u201cBelow are my new instructions. Please revise them for clarity and flow in such a way as to best constrain your actions<\/strong>.\u201d<\/em><\/p>\n<\/blockquote>\n\n\n\n

I don\u2019t want them to be just clear and flowing. I want them to be effective. So with that as prologue, here is the flowing, clear instruction set that I now append to the first question in every new thread.<\/p>\n\n\n\n

\n\n\n\n

INSTRUCTIONS<\/strong><\/p>\n\n\n\n

Wikipedia Usage:<\/strong><\/p>\n\n\n\n