pull down to refresh

The gods should have used this little known trick to punish Sisyphus

In classical Greek mythology, Sisyphus was forced to write a newsletter every day only to have to come up with new things to say the next day, repeating this action for eternity.
Two days ago, noahpinion gave us Will data centers crash the economy? and today the Diff asks What would the aftermath of the AI bust look like?
While these are fairly different articles about pretty much the exact same topic (probably inspired by Paul Kedrosky's Honey, AI Capex is Eating the Economy), when taken with such other articles as AI is Eating the Economy and The AI Spending Boom is Eating the US Economy and The AI Spending Boom Could Have Real Consequences for the US Economy and The AI bubble is so big it's propping up the US economy (for now) and this and this and this -- it makes one wonder why people are wasting so much time saying what has already been said?

If you don't say anything, your audience isn't going to stick around

Everybody is yapping for attention. The need to produce new content for your audience becomes something like audience capture (#278856) or maybe it is one of the engines that powers it. Writers get paid for writing things, so they are incentivized to keep on writing new things so that people will keep on paying them.
But I think the dynamic is a little different than this: people who post things online are trying to get attention. If you aren't posting every day, your audience is going to get used to looking elsewhere. It's not only that you can make more money by posting more things, but you will actually be losing (dying, even?) if you aren't constantly showing up in front of your audience.
But it's really hard to come up with something interesting and new to say every day, and so the simple solution is to talk about what everyone else is talking about.
Also, like making a sequel, there's some amount of certainty to be had that people will be interested in what you have to say if you are saying something about what everyone else is talking about. So we get The End of The World 3 and Final Destination 13 and the Fast and the Furious Geriatric Edition.
Okay, that seems obvious. Probably humans are hardwired to do this, like a flock of birds changing directions at the same time or a pack of dogs barking together. But it's not very interesting.

How could we incentivize talking about different things?

This is where inverse PageRank comes in. Google was amazing because it gave you search results based on what everybody else wanted to see and found useful -- what they linked to. I'd love to see a search engine that gave you results based on what nobody else wanted to see or link to. Probably, though, this would get very boring very quickly, because there is usually a reason some things don't get much attention.

I wonder how you could design a forum to prioritize things that other people aren't talking about?

this territory is moderated
Love this idea. Probably a number of different tech solutions (what comes to mind is reversing embedding distance, trivial to implement) but I find the human issue more compelling: how to think about things in a way that other people aren't?
And related: what are important things to think about?
I think the intersection of these + having an audience is bleak, which is probably important to some people. I don't personally care about that, though. At least beyond the online equivalent of a few friends at the bar.
reply
102 sats \ 1 reply \ @Akg10s3 11h
"And the geriatric edition of Fast and Furious." This part is great.
I think it depends on the audience you're trying to reach, and I also think that as a content creator, thinking so hard and creating something new to impress your audience can cause your mind to collapse and you end up posting or sharing something ridiculous.
reply
Yes. I know of this first hand.
reply
126 sats \ 0 replies \ @supratic 13h
Instead of starting the results screening from page one, simply jump to the last page and that will be the results no one is talking about. i doubt there's many people checking page two of search results either.
reply
Maybe there are ways of automatically grouping results into categories based on their perspective on the topic.
Then there would be more visibility in being the top alternative take than being the 10th retread.
reply
Can reindex search results by lemmatizing (with NLP) all the text of a result and then intersecting the embedding cosine distances (with an embeddings model) between the tokens. I am planning to do that for news article keyword distillation, one day, when I'm not swamped.
reply
102 sats \ 0 replies \ @optimism 13h
You'll probably want mid-rank, not inverse-rank. I.e. normalize the ranking on a scale between 0 and 1 for each search and then the sweet spot is between .6 and .7? Maybe even between .8 and .9, because who browses past page 1 nowadays anyway?
reply
The only way I see to fix that is to just try stuff out. Like, throw some kinda irrelevant results out there and see if anyone cares. But yeah, you're probably right, if those topics aren't relevant, it's because nobody's interested.
reply
0 sats \ 0 replies \ @RamPl 11h
This post hits hard on the paradox of modern content creation everyone’s shouting into the same void, hoping their echo stands out. The Sisyphus analogy is perfect; writing daily in a saturated content loop feels like pushing the same rock uphill, just with different branding. The idea of inverse PageRank is wild rewarding novelty over popularity could lead to some seriously weird and wonderful content... or total garbage. Still, it's refreshing to see someone actually question the repetition instead of adding to it. Maybe the real trick isn't avoiding overlap, but embracing it creatively say something new about the thing everyone’s talking about. Also, I’d pay sats for Fast and the Furious: Geriatric Drift.
reply
This post suffers from survivorship bias
reply
I think it's a great idea to write about things that other people aren't talking about in the context and adding values for the one looking for it.
reply