Post #2640990 · Elektrine

2026-01-17 00:31 UTC

@notthatdelta@furry.engineer @Orb2069@mastodon.online @lexfeathers@tenforward.social Sure, but - that's not my point? "It's unprofitable" sure, yeah, that's obvious "It's unpopular" does _not_ necessarily follow, given that they _lose money on each of their users_. 15k isn't a nothingburger of a sample pool, but - that's why I'm curious about how representative it is. My thought process here is something along the lines of, 'If you actually like AI, _you wouldn't be using DDG to begin with_, you'd be fine with google or chatgpt or whatever so you wouldn't even notice this and you're not going to care to vote in it'. Hence: is this a representative sample _of fedi_? _Of DDG users_ as a whole? Or of the _general public_?

Replies (1)

@notthatdelta@furry.engineer 2026-01-17 00:47

@pixx@merveilles.town @Orb2069@mastodon.online @lexfeathers@tenforward.social whoops, I actually meant to reply to Orb and replied to your reply - but you raise a good point. I definitely think the sample is biased, probably toward fedi users (for anecdotal evidence, I use DDG as my primary search engine at work and for personal use, and I never saw *anything* about this via DDG). But even assuming this sample set is primarily DDG users (not just DDG-fedi users) it's inherently biased, because someone going to the effort of switching to a non-default search engine is probably in the top 10 or 5 percent of computer users by skill level, and therefore more likely to be concerned about their privacy or cognizant of the ethical and environmental issues around LLM usage.

Open ##2640991