Like many others, I am worried about the seemingly ever-increasing flood of slop, both on crates.io and more generally across the entire web. I wanted to post this as a random thought collection as the start for discussions, because I think we should be looking for mechanisms that incentivize "real" contributions and prevent "slop cannons" from spamming their stuff and hogging resources.
I think there is a reason for why AI tools are predominantly used in programming, writing, generating images/videos, deriving mathematical proofs, writing spam messages etc: there is no cost to trying and no cost to failing. If someone who uses a bot to post 1000 messages per day would need to pay 1ct per post, it would cost them a significant amount of money to continue doing so, while people who post a few times a day only pay a few ct. Instead, posting text/images/videos, trying to prove theorems, uploading crates or publishing low-effort clones can be done without fearing any losses, even more, is reinforced by external rewards (ad revenue, attention, internet points,...). If instead every attempt would take a small amount of (artificially induced) effort, these "intelligent spam campains" would cease to be viable if they are done just for fun/just because they can be done.
If we create a separate crate registry where you have to pay a small fee to be able to publish a crate onto it, and add a website where people who do not have enough money can advertise their crate and look for "sponsors", we would be able to create a curated list of crates where someone at some point at least found each entry worthwhile to pay for. "money spam attacks" could be detected (i.e. someone who randomly vets for 1000 packages or often vets for packages that noone else vetted for) to some degree. The resulting money could be used to provide the infrastructure that the registry is running on, or to people who work through the contributions and add their opinions on each one.
I know that it is not a perfect solution, but to me something like this "proof of worth" seems to be the only way around more or less helplessly drowning in slop. If anyone has other ideas or ideas for improvements, I would love to hear it. I think we need to have more discussions on these topics.
On a sidenote, this concept applied to free Steam games - where you have to pay to publish (and maybe get the money after your game reaches 1000 players?) - could also prevent a significant amount of slop-clones on steam, no? I know that it is completely against the concept of the free, democratic, internet to create such hurdles (we did it with blockchains and cryptocurrency, but using a proof of work algorithm gives the resources used to the energy providers instead of the institutions actually providing the service), but it seems like that concept needs to grow some teeth to fight back and survive in the age of AI.
1 post - 1 participant
