It's adorable that people are only slowly realizing that Google search at least fed sites traffic, while chat AI thingies slurp up and summarize content, which they anonymize and feed back, leaving the slurped sites traffic-less and dying. But, innovation.

It is, in a way, a tragedy of the commons problem, with no easy way to police "over grazing" of the information commons, leading to automated over-usage and eventual ecosystem collapse.

There is, btw, a terrific new updating of Ostrom and others' work out on the ToC, arguing that in the real world it can lead much faster to what it calls a mass of catastrophic poverty (many have-nots) and oligarchs (a few haves).

See:

Generic catastrophic poverty when selfish investors exploit a degradable common resource https://royalsocietypublishing.org/doi/full/10.1098/rsos.221234?af=R#d20123427e6469s

@paul I’d been thinking of ChatGPT as a potential new layer of structure - and hadn’t focused on the one-directional flow of information - which definitely emphasizes the extractive relationship. Possible also that ‘the new ecosystem’ is yet incomplete (?)

@paul @danyork I actually beg to differ on your analysis (though I agree with your starting fact).

This isn't a tragedy of the commons, we don't have people overgrazing the Web's information (which they couldn't if they tried), it's the opposite: a tragedy of capture.

There is a link between the two in that the failure boils down to collective action problems. Publishers would benefit from boycotting Google, but that only works if they all do.

@paul @danyork Google is particularly adept at taking a shared resource and organising prisoners' dilemmas that didn't exist previously. PDs make cooperation fragile and so the bigger party wins.

The next logical step after AI in search is that they'll have to produce content directly for it. No one else can be sustainable. I hear that is already the case for some of what shows up there.

@paul I don't see how the chat model beats search and I don't understand why people think it will. I'm always going to want to be presented with different perspectives for most searches and I'm always going to prefer to know the source of the assertions being made.
@paul Another problem with AI-based chat is how it incentivizes bad actors to publish content to just shape the answers the AI is giving users rather than to make money off of ads. For example, business industry groups might fund lots of websites that discuss the evils of minimum wage and child labor laws. Even though they run the websites at a loss, the sites are training the AI that minimum wage and child labor laws are bad, so that's what the AI will tell its users.
@paul Hrm, good thought. I'd hope the AIs would answer and then say something like "For more details see this site ..." or even better "Here's a list of forums/people/orgs as additional resources".

@paul I learned a lot from what you said here. It's a powerful metaphor.

I hope you can take this as constructive feedback because I'd like this idea to spread: I hesitate to share it because of the "it's adorable" framing. When people are coming around to your way of thinking it's probably not the best to berate them as infantile for not doing it sooner