This is great. I've been wondering what I'd be doing once SEO spam slop domains filled up my domain filtering list (which is still a long way away), but this seems like it will solve that issue....
This is great. I've been wondering what I'd be doing once SEO spam slop domains filled up my domain filtering list (which is still a long way away), but this seems like it will solve that issue. Crowdsourcing feels like a natural fit for Kagi, especially placing it in the same menu where you change the ranking of websites in general.
My only concern is that this is at the domain level. I know it's by far the easiest, but a similar issue happens with the AI image labelling in Kagi's image search. Stock image sites saw AI as an opportunity to not pay people for their work and so the same domain had both a large library of actual stock imagery and AI slop, so you couldn't really mark the entire domain as AI. It's certainly better than nothing, so I'm not really complaining.
I wonder what the effects of (eventually) releasing the dataset will be. Kagi is too small to have a noticeable impact on the traffic to websites (really only Google is big enough for that), but if it gets used by more people and projects then there is at least the possibility that something changes.
This half-assed thinking seems endemic to businesses and AI, and I find it amusing. If you're going to replace your product with output from AI, why would I buy that product when I can just use...
Stock image sites saw AI as an opportunity to not pay people for their work
This half-assed thinking seems endemic to businesses and AI, and I find it amusing. If you're going to replace your product with output from AI, why would I buy that product when I can just use the AI? The capitalist ouroboros consumes itself. (At least, until the burning dumpsters of investor money run out, then it'll cost an absurd amount to generate them.)
Never mind that copyright-eligibility of AI output has already had at least a couple of legal cases against it, hinging on the requirement for sufficient human artistic input to go into a result. That also means the business of selling rights to use the images is suspect, since they're inherently public domain if that precedent stands.
This is great. I've been wondering what I'd be doing once SEO spam slop domains filled up my domain filtering list (which is still a long way away), but this seems like it will solve that issue. Crowdsourcing feels like a natural fit for Kagi, especially placing it in the same menu where you change the ranking of websites in general.
My only concern is that this is at the domain level. I know it's by far the easiest, but a similar issue happens with the AI image labelling in Kagi's image search. Stock image sites saw AI as an opportunity to not pay people for their work and so the same domain had both a large library of actual stock imagery and AI slop, so you couldn't really mark the entire domain as AI. It's certainly better than nothing, so I'm not really complaining.
I wonder what the effects of (eventually) releasing the dataset will be. Kagi is too small to have a noticeable impact on the traffic to websites (really only Google is big enough for that), but if it gets used by more people and projects then there is at least the possibility that something changes.
This half-assed thinking seems endemic to businesses and AI, and I find it amusing. If you're going to replace your product with output from AI, why would I buy that product when I can just use the AI? The capitalist ouroboros consumes itself. (At least, until the burning dumpsters of investor money run out, then it'll cost an absurd amount to generate them.)
Never mind that copyright-eligibility of AI output has already had at least a couple of legal cases against it, hinging on the requirement for sufficient human artistic input to go into a result. That also means the business of selling rights to use the images is suspect, since they're inherently public domain if that precedent stands.