|
Casey Finnigan posted:bro old men used to marry like 12 year old girls and it was normal very first post is "well actually"-ing the OP, goons never change
|
#
?
Dec 16, 2022 12:58
|
|
|
|
| # ? May 2, 2024 15:25 |
|
|
add to that particle pollution in the city, pesticide pollution in the countryside, lead everywhere, microplastics, phthalates, pfas, ultraviolet radiation, chronic stress, wars, famines, natural disasters, lol lmao it's enough to drive someone to religion. gently caress the 21st century. Fuvk the computer
|
|
#
?
Dec 16, 2022 12:59
|
|
|
buddy,
|
|
#
?
Dec 16, 2022 12:59
|
|
|
Zodium posted:buddy,
|
|
#
?
Dec 16, 2022 13:00
|
|
|
im reactin' here
|
|
#
?
Dec 16, 2022 13:00
|
|
|
i hate the computer. and yet?? i love the computer
|
|
#
?
Dec 16, 2022 13:01
|
|
|
a strange fowl posted:i hate the computer. and yet?? i love the computer you wanna gently caress it?
|
|
#
?
Dec 16, 2022 13:02
|
|
|
|
|
#
?
Dec 16, 2022 13:03
|
|
|
help i 'm turning into a pixel on an endless screen
|
|
#
?
Dec 16, 2022 13:04
|
|
|
Weka posted:Doing the lords work. Have you seen that bit about the racist medical scan AI? Thanks fellow poster! It’s been a while since I’ve looked at literature about medical applications. I did interview someone who worked on a system for mammogram classification but that was years ago and I was much dumber than I am now, if you can believe it. The medical system is deeply racist of course so training a model on any data produced by that system is pretty lol worthy. There are a number of techniques that could theoretically be applied to minimize the racism but capitalism, etc. Anyway, I’d be interested in seeing a reference if you have one to hand. a strange fowl posted:when i was growing up there was still widespread resistance to screentime for kids. The most hosed up part about this is that all of the people in my peer group are super strict about screen time for their kids so this is mostly a problem for poors and so nobody cares and society will collapse before we can establish the impacts because who cares about poor people. Also something about feedback loops making it even harder to transcend generational wealth. Cuttlefush posted:you wanna gently caress it? Good(?) news, “AI powered” sex toys now exist. I’m pretty sure it is just a gimmick at the moment, but eventually we can probably use them for eugenics or hopefully something even worse!
|
|
#
?
Dec 16, 2022 13:30
|
|
|
a strange fowl posted:add to that particle pollution in the city, pesticide pollution in the countryside, lead everywhere, microplastics, phthalates, pfas, ultraviolet radiation, chronic stress, wars, famines, natural disasters, lol lmao it's enough to drive someone to religion. gently caress the 21st century. Fuvk the computer we used to get our strange and exotic diseases for which we had no names and whose impact we couldn’t fathom the honorable way, by working for pennies under pitiless robber barons and living in polluted cities blackened with soot. now you can get them anywhere on earth, without even noticing. you don’t even have to try. kids these days are too spoiled i tell you.
|
|
#
?
Dec 17, 2022 08:40
|
|
|
TehSaurus posted:Thanks fellow poster! It’s been a while since I’ve looked at literature about medical applications. I did interview someone who worked on a system for mammogram classification but that was years ago and I was much dumber than I am now, if you can believe it. The medical system is deeply racist of course so training a model on any data produced by that system is pretty lol worthy. There are a number of techniques that could theoretically be applied to minimize the racism but capitalism, etc. Study shows AI can identify self-reported race from medical images that contain no indications of race detectable by human experts.
|
|
#
?
Dec 20, 2022 05:25
|
|
|
i think the internet is pretty cool OP. my phone has little gay people on it who talk to me sometimes
|
|
#
?
Dec 20, 2022 05:43
|
|
|
a strange fowl posted:i am deeply afraid of what is going to happen when they decide it's more cost effective to just not moderate socials at all and just let everybody have the unfiltered content all the time for clicks theres always going to be some moderation on big socials for two reasons: payment processors and advertisers
|
|
#
?
Dec 20, 2022 06:41
|
|
|
Weka posted:Study shows AI can identify self-reported race from medical images that contain no indications of race detectable by human experts. The Right Brothers - Phrenology Was Right
|
|
#
?
Dec 20, 2022 06:43
|
|
|
thanks for the paywalled article
|
|
#
?
Dec 20, 2022 06:43
|
|
|
It's not paywalled for me, idk why the difference but here. In general just use archive.ph to break paywalls. https://archive.ph/62aPj
|
|
#
?
Dec 20, 2022 06:58
|
|
|
a strange fowl posted:add to that particle pollution in the city, pesticide pollution in the countryside, lead everywhere, microplastics, phthalates, pfas, ultraviolet radiation, chronic stress, wars, famines, natural disasters, lol lmao it's enough to drive someone to religion. gently caress the 21st century. Fuvk the computer I’m going orthodox ☦️
|
|
#
?
Dec 20, 2022 07:18
|
|
|
Weka posted:Study shows AI can identify self-reported race from medical images that contain no indications of race detectable by human experts. Ah, yeah. People have this thought that if they remove sensitive information like race, gender, whatever from their dataset that the resulting model will be unbiased. This is obviously totally wrong because data is generated by a process which has inbuilt biases. I have lost track of the current thinking on fairness in ml, but the last time I checked there was a push to include this information in the dataset, while using counterfactuals to validate the results. For example if you change Joe’s race from white to black, the model shouldn’t suddenly start denying him parole or whatever. It doesn’t seem to me that the math would work out on that - I’d think you would basically get the same model as the race blind one. But I’m not very smart either. There are many other approaches, but they all seemed hard to implement, and didn’t necessarily guarantee an improved result, so I don’t expect them to ever really catch on. Also the cruelty is the point, so why would we try to reduce it?
|
|
#
?
Dec 20, 2022 19:24
|
|
|
have they tried programming the computer for affirmative action?
|
|
#
?
Dec 20, 2022 23:30
|
|
|
a strange fowl posted:add to that particle pollution in the city, pesticide pollution in the countryside, lead everywhere, microplastics, phthalates, pfas, ultraviolet radiation, chronic stress, wars, famines, natural disasters, lol lmao it's enough to drive someone to religion. gently caress the 21st century. Fuvk the computer 
|
|
#
?
Dec 21, 2022 04:40
|
|
|
|
| # ? May 2, 2024 15:25 |
|
|
Who cares you live for like 50 years don't worry about it
|
|
#
?
Dec 21, 2022 05:03
|
|
