All of CAC's Comments + Replies

CAC40

My assumption is that many of these successes would tend to be widely distributed around some mean, rather than being narrowly concentrated at one point.

So if a joke needs to be 7/10 funny to get a laugh, but a comedian delivers what is actually a 6.5/10 joke, you’ll still get some subset of people who find it funnier than it is, such that it gets an appropriate amount of laughs.

Probably there’s some inefficiency, but because of this effect, the number of laughs/number of upvotes I think gives quite good information about the perceived quality of the joke/post.

CAC70

I asked GPT 4.5 to write a system prompt and user message for models to write Pilish poems, feeding it your comment as context.

Then I gave these prompts to o1 (via OpenAI's playground).

GPT 4.5's system prompt

You are an expert composer skilled in writing poetry under strict, unusual linguistic constraints, specifically "Pilish." Pilish is a literary constraint in which the length of consecutive words precisely matches each digit of π (pi). The first word contains 3 letters, second word 1 letter, third word 4 letters, fourth word 1 letter, fifth word 5 lette

... (read more)
2Mo Putera
This is great, thanks! Didn't think of the model-prompting-model trick.