The Minions of Avian Intelligence
-
@SmartmanApps@dotnet.social @Stem@framapiaf.org @davidrevoy@framapiaf.org It’s not a random sample…
@breizh @davidrevoy @Stem
"It’s not a random sample" and where EXACTLY was it directed at only a certain group?
-
@SmartmanApps @AmyIsCoolz @Stem
> "[...] we know there's no bias in the Mastodon poll [...]"
Best joke of the year, imo.@davidrevoy @AmyIsCoolz @Stem
"Best joke of the year, imo" - it's not a joke. Exact same results as the web poll of more than 175,000 people. -
@davidrevoy @AmyIsCoolz @Stem
"Best joke of the year, imo" - it's not a joke. Exact same results as the web poll of more than 175,000 people.@SmartmanApps @AmyIsCoolz @Stem Do you really think, that the people using Mastodon (or the Fediverse) is a totally representative sample of the real population ideas and opinions?
-
@SmartmanApps @AmyIsCoolz @Stem Do you really think, that the people using Mastodon (or the Fediverse) is a totally representative sample of the real population ideas and opinions?
@davidrevoy @AmyIsCoolz @Stem
"is a totally representative sample of the real population ideas and opinions?" - did you miss that it got the exact same results as the web poll of more than 175,000 people? -
@davidrevoy @breizh @Stem
"in my local reality" - which is no longer a random sample."just the result of active link shared in communities informed about AI issues" - which you have no evidence to support
"175K is not a lot" - as I said, in a random sample you only need *25* to get an accurate result.
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org “In my local reality” is not a random sample, nor are online circles.
The Fediverse isn’t a random sample of the world population.
DDG users aren’t a random sample of the world population.
Yeah, the link was shared in a lot of places, but these places were not random. When you share a link, you’re doing it in your own “local reality”.
-
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org “In my local reality” is not a random sample, nor are online circles.
The Fediverse isn’t a random sample of the world population.
DDG users aren’t a random sample of the world population.
Yeah, the link was shared in a lot of places, but these places were not random. When you share a link, you’re doing it in your own “local reality”.
@breizh @davidrevoy @Stem
""In my local reality" is not a random sample" - that's right, that's what David said"nor are online circles" - go ahead and run a poll of people on the street, I guarantee you'll get the same results.
How many people EXACTLY in the population do you think are NOT online? 
Spoiler alert: if they're not online, then they have no access to AI online either"The Fediverse isn’t a random sample of the world population" - the statistics tell you otherwise
-
@breizh @davidrevoy @Stem
""In my local reality" is not a random sample" - that's right, that's what David said"nor are online circles" - go ahead and run a poll of people on the street, I guarantee you'll get the same results.
How many people EXACTLY in the population do you think are NOT online? Spoiler alert: if they're not online, then they have no access to AI online either"The Fediverse isn’t a random sample of the world population" - the statistics tell you otherwise
@breizh @davidrevoy @Stem
"these places were not random" - says person failing to point out which specific group it was directed at"When you share a link, you’re doing it in your own "local reality"" - not if it's shared widely. I ran a poll recently that nearly 1000 people answered, and I only have 390 followers, so it went WELL beyond my local reality! That's what happens when it gets shared by someone with a lot of followers
-
@breizh @davidrevoy @Stem
""In my local reality" is not a random sample" - that's right, that's what David said"nor are online circles" - go ahead and run a poll of people on the street, I guarantee you'll get the same results.
How many people EXACTLY in the population do you think are NOT online? Spoiler alert: if they're not online, then they have no access to AI online either"The Fediverse isn’t a random sample of the world population" - the statistics tell you otherwise
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org Being online does not mean accessing to everything everywhere on the net.
“The Fediverse isn’t a random sample of the world population” - the statistics tell you otherwise
Does it? Do you have any study or numbers to prove it?
The fediverse have way much french people in proportion than the world have, that’s already a big bias…
And there is more LGBT people in proportion than other social medias too, that’s another bias.
There is more people that care about privacy than in the general population, too.
I mean, the fact that there is a lot more of people on BlueSky than on the Fediverse is already a proof that the Fediverse isn’t a random sample (BlueSky isn’t either, for what it’s worth, but they are populated by differents persons that will have different opinions on some subjects).
Like, if what you say is right, then we couldn’t find a subject where BlueSky, X and the Fediverse have different results.
But there is some…
-
@breizh @davidrevoy @Stem
"these places were not random" - says person failing to point out which specific group it was directed at"When you share a link, you’re doing it in your own "local reality"" - not if it's shared widely. I ran a poll recently that nearly 1000 people answered, and I only have 390 followers, so it went WELL beyond my local reality! That's what happens when it gets shared by someone with a lot of followers
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org Not, it didn’t go well beyond your local reality, only by a little. People following the people that follow you have way more in common with you that you’d think (the language is one of them, first. I never responded to a dutch or greek poll AFAIK, only french and english ones…).
-
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org Being online does not mean accessing to everything everywhere on the net.
“The Fediverse isn’t a random sample of the world population” - the statistics tell you otherwise
Does it? Do you have any study or numbers to prove it?
The fediverse have way much french people in proportion than the world have, that’s already a big bias…
And there is more LGBT people in proportion than other social medias too, that’s another bias.
There is more people that care about privacy than in the general population, too.
I mean, the fact that there is a lot more of people on BlueSky than on the Fediverse is already a proof that the Fediverse isn’t a random sample (BlueSky isn’t either, for what it’s worth, but they are populated by differents persons that will have different opinions on some subjects).
Like, if what you say is right, then we couldn’t find a subject where BlueSky, X and the Fediverse have different results.
But there is some…
@breizh @davidrevoy @Stem
"Being online does not mean accessing to everything everywhere on the net" - but still provides a random sample of the population"It does? Do you have any study or numbers that prove it" - yep, the web poll of 175,000 people, as I already said
"fediverse have way much french people in proportion than the world have, that’s already a big bias" - only if the French have a different attitude to the rest of the world. Go ahead and provide any evidence you have
-
@breizh @davidrevoy @Stem
"Being online does not mean accessing to everything everywhere on the net" - but still provides a random sample of the population"It does? Do you have any study or numbers that prove it" - yep, the web poll of 175,000 people, as I already said
"fediverse have way much french people in proportion than the world have, that’s already a big bias" - only if the French have a different attitude to the rest of the world. Go ahead and provide any evidence you have
@breizh @davidrevoy @Stem
"there is more LGBT people in proportion than other social medias too, that’s another bias" - ditto"There is more people that care about privacy than in the general population, too" - ditto
"lot more of people on BlueSky than on the Fediverse is already a proof that the Fediverse isn’t a random sample" - no it isn't. Not sure what you're not understanding about the word "random", nor only 25 random votes needed
"there is some" - X is a Nazi Bar, so of course
-
@breizh @davidrevoy @Stem
"there is more LGBT people in proportion than other social medias too, that’s another bias" - ditto"There is more people that care about privacy than in the general population, too" - ditto
"lot more of people on BlueSky than on the Fediverse is already a proof that the Fediverse isn’t a random sample" - no it isn't. Not sure what you're not understanding about the word "random", nor only 25 random votes needed
"there is some" - X is a Nazi Bar, so of course
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org And the Fediverse is a privacy-friendly bar.
If you aknowledge that X and the Fediverse have different people that have differents opinions, then you’re already destroyed your own point.
-
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org Not, it didn’t go well beyond your local reality, only by a little. People following the people that follow you have way more in common with you that you’d think (the language is one of them, first. I never responded to a dutch or greek poll AFAIK, only french and english ones…).
@breizh @davidrevoy @Stem
"it didn’t go well beyond your local reality, only by a little" - it went beyond by a factor of at least 150%, and that's even assuming ALL of my followers voted, which I know they didn't"I never responded to a dutch or greek poll AFAIK, only french and english ones" - my poll was a Maths poll, so language isn't an issue. Also there are translation tools
-
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org And the Fediverse is a privacy-friendly bar.
If you aknowledge that X and the Fediverse have different people that have differents opinions, then you’re already destroyed your own point.
@breizh @davidrevoy @Stem
"the Fediverse is a privacy-friendly bar" and 90% of them are anti-AI"then you’re already destroyed your own point" - no I haven't. You haven't shown it wasn't shared on X as well
-
@breizh @davidrevoy @Stem
"the Fediverse is a privacy-friendly bar" and 90% of them are anti-AI"then you’re already destroyed your own point" - no I haven't. You haven't shown it wasn't shared on X as well
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org If the pro-AI/anti-AI position of people on X is the same, that would be a coincidence, not a proof that X is representative of the population either.
You are saying that people can be representative on certain subjects and not others, which don’t make any sense.
Either is representative on all subjects, ore none.
If X have more people accepting facsism than the Fedivers, then none of them are representating on any other subject, there’s already a bias.
-
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org If the pro-AI/anti-AI position of people on X is the same, that would be a coincidence, not a proof that X is representative of the population either.
You are saying that people can be representative on certain subjects and not others, which don’t make any sense.
Either is representative on all subjects, ore none.
If X have more people accepting facsism than the Fedivers, then none of them are representating on any other subject, there’s already a bias.
@SmartmanApps@dotnet.social @Stem@framapiaf.org @davidrevoy@framapiaf.org Maybe that bias isn’t important, but you have to prove it, while you’re assuming it don’t have any importance until proven false, which is a really bad way to do statistics…
-
@SmartmanApps@dotnet.social @Stem@framapiaf.org @davidrevoy@framapiaf.org Maybe that bias isn’t important, but you have to prove it, while you’re assuming it don’t have any importance until proven false, which is a really bad way to do statistics…
@breizh @davidrevoy @Stem
"If the pro-AI/anti-AI position of people on X is the same, that would be a coincidence" - or simply proof that we had a random sample to begin with"not a proof that X is representative of the population either." - a random sample of *25* people is representative of the population
"You are saying that people can be representative on certain subjects and not others" - never said anything of the sort
-
@breizh @davidrevoy @Stem
"If the pro-AI/anti-AI position of people on X is the same, that would be a coincidence" - or simply proof that we had a random sample to begin with"not a proof that X is representative of the population either." - a random sample of *25* people is representative of the population
"You are saying that people can be representative on certain subjects and not others" - never said anything of the sort
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org A random sample of 25 people, maybe, but you have to prove it’s a random sample… my point is precisely that you can’t assume the Fediverse or the DDG poll was done on a random sample (because people sharing a DDG poll or using the Fediverse aren’t random, they are people that already have a bias toward privacy).
-
@breizh @davidrevoy @Stem
"If the pro-AI/anti-AI position of people on X is the same, that would be a coincidence" - or simply proof that we had a random sample to begin with"not a proof that X is representative of the population either." - a random sample of *25* people is representative of the population
"You are saying that people can be representative on certain subjects and not others" - never said anything of the sort
@breizh @davidrevoy @Stem
"there’s already a bias" - so you're saying that you can't ask fascists what they have for breakfast, because somehow being a fascist means they have different breakfast preferences to everyone else?"Maybe that bias isn’t important" - you're finally catching on
"you have to prove it" - no I don't. YOU'RE the one claiming it's biased, YOU are the one who has to prove that claim
"you have to prove it’s a random sample" - nope, you have to show it was biased
-
@SmartmanApps@dotnet.social @davidrevoy@framapiaf.org @Stem@framapiaf.org A random sample of 25 people, maybe, but you have to prove it’s a random sample… my point is precisely that you can’t assume the Fediverse or the DDG poll was done on a random sample (because people sharing a DDG poll or using the Fediverse aren’t random, they are people that already have a bias toward privacy).
@SmartmanApps@dotnet.social @Stem@framapiaf.org @davidrevoy@framapiaf.org I think that you should learn about filter bubble.
Local reality exist online the same way as they do offline. If you’re saying that a local reality isn’t a valid random sample, then a poll shared on the web isn’t either. Its results will depend on where it was shared, and who have shared it.
