Hacker News Reader

Comment by wonder_er | Hacker News Reader

I believe the humor and novelty and interestingness of this standup routine is related to the _density_ of information conveyance.

I read this paper a while back, it was so good I wanted to make it more sharable:

https://josh.works/driven-by-compression-progress-novelty-hu...

The title is:

Driven by Compression Progress: A Simple Principle Explains Essential Aspects of Subjective Beauty, Novelty, Surprise, Interestingness, Attention, Curiosity, Creativity, Art, Science, Music, Jokes

And... it's really good. It argues that a common thread among many different domains (as listed in the title) is how much information can be compressed into the information stream, like the bit-rate. Higher bitrate, more interesting information.

View on HN

As a practitioner of improv comedy (including a little clowncraft), I'd argue that you can produce laughs from essentially doing nothing, repeatedly.

So, I'm not really convinced. Maybe that principle can explain some cases of beauty, novelty, .... jokes. Definitely not all.

I think your example fits the theory pretty well. If you flip a coin ten times and get mixed heads and tails, it's boring, but if you get ten heads in a row (so the theory goes), at some point your brain suddenly recognizes that you're dealing with an unexpected source of compressible data. Surprise!

> It argues that a common thread among many different domains (as listed in the title) is how much information can be compressed into the information stream, like the bit-rate. Higher bitrate, more interesting information.

Piet Mondrian, Philip Glass, Andy Kaufman would provide trivial counter-arguments to the idea that high-density = "better" (more interesting, funnier, etc.).

I only know of Andy Kaufman from your examples.

Wasn’t some of his work very much reliant on what he didn’t say directly? An argument one could make would be that the information density is still high in terms of what is perceived or experienced. That would still be a kind of compression but it involves the context and audience to a very high degree. I think he sometimes made his performances so intense, that it would be unbearable not to engage with them.

You might argue that work like Mondrian's implies additional information, in the form of implied commentary on centuries of art history (arguably necessary for his work to be considered interesting or "good"). You might then argue that this makes up for the lower explicit information density.

It’s model building isn’t it? Our world model is a compressed representation of the real world. The better the model the more effective the compression, given the same storage space. That’s why unexpected results excite us, because they allow us to improve our model. I’d wager the best comedy allows us to see the mundane (i.e. well within our model’s capability) from a novel perspective.

This comment is the beautiful content that I come to HackerNews for.

I wonder if there is a proper theory of humor out there. For example, surprise is a big part of any joke - be it double meaning joke, a pun, something clever; the only common thing among almost every joke is surprise. So I wonder if there is a comprehensive theory of humor out there.

Compression Progress = finding pattern.

It's the pattern (i.e. redundancy) to be found that makes it interesting. But pattern reduces information capacity - the greatest information is conveyed when there is no constraint of pattern or regularity, and is indistinguishable from pure randomness or white noise: maximum entropy.

Bollocks. The stand-up I will never forget is Stewart Lee, keeping two thousand punters in stitches while he spins a single joke out over ten minutes.

It's not a stupid theory but it misses masterful simplicity.

Norm Macdonald was good at that too.

Reply on news.ycombinator.com