[eru]

Btw, I noticed that GPT 4.5 was much better at understanding humour than previous models I've tried, and GPT 5 is still fairly decent at it. (For a computer.)

The observation with GPT 4.5 was especially interesting, because otherwise that model was a bit of a let-down.

[XenophileJKO]

Well it was a BIG model. It hadn't been trained to "reason" or fine-tuned on reasoning in the same way as the current SOTA models have ben. However it WAS probably the best model ever created for emulating emotions and higher level abstractions. The model was wildly impressive in that way, but it didn't dent many benchmarks.

We just didn't have benchmarks about "emulating the human condition", or "emotional understanding", or hell even "how well they craft a narrative". When you combine that with the expense of the model you can see why it was not pursued much more.

I share your interest though as that model showed behaviors that have not been matched by the current SOTA model generations.

[ACCount37]

Ah yes, the legendary "big model smell".

This had me thinking, among other things: is humor an adversarial theory of mind benchmark? Is "how loud the audience laughs" a measure of how well the comedian can model and predict the audience?

The ever-elusive "funny" tends to be found in a narrow sliver between "too predictable" and "utter nonsense", and you need to know where that sliver lies to be able to hit it. You need to predict how your audience predicts.

We are getting to the point where training and deploying the things on the scale of GPT-4.5 becomes economical. So, expect funnier AIs in the future?

[eru]

Human humour certainly has a tinge of 'showing of your smarts' to it.