do you think AI models are being trained with our posts?

所有讨论 > Steam 论坛 > Off Topic > 主题详情

allegedly 10 月 24 日下午 12:06

are they learning from ot?

< >

正在显示第 1 - 15 条，共 17 条留言

halloween guy (apar. 96) 10 月 24 日下午 12:08

"Do you like my pineapple?"

Voroff 10 月 24 日下午 12:10

Better us than whatever Musk hastily hardcoded in his own IA when it was debunking him too frequently :D

TGC> The Games Collector 10 月 24 日下午 12:12

Nah, worse, they use reddit for that.

OoOoOoooOOoOoorgle 10 月 24 日下午 12:12

At least one model is.

SirSeanicus 10 月 24 日下午 12:13

Absolutely. They also use Reddit too.

The Butlerian Jihad can't come soon enough.

最后由 SirSeanicus 编辑于; 10 月 24 日下午 12:13

Uncle Sam 10 月 24 日下午 12:14

引用自 allegedly：
do you think AI models are being trained with our posts?
are they learning from ot?

Probably, since bots are more or less half the internet traffic atm.

Vinz Clortho 10 月 24 日下午 12:15

They aren't learning much if they are.

Chaosolous 10 月 24 日下午 12:15

They sure are.

When your AI is like:

"I built a bunch of bots to release for psy-ops and also nuclear war would be funny."

Just know, that that's a little bit of me, inside of a little bit of it, now getting put inside of a little part of you (your eyeballs specifically).

We are one. Join us.

最后由 Chaosolous 编辑于; 10 月 24 日下午 12:16

allegedly 10 月 24 日下午 12:15

引用自 TGC> The Games Collector：
Nah, worse, they use reddit for that.

Uncle Sam 10 月 24 日下午 9:10

引用自 Vinz Clortho：
They aren't learning much if they are.

Well atleast they will learn to find & nuke Atlantis even further, so the AI can protect pineapple pizza.

#10

Fake 10 月 24 日下午 9:12

Why do you think that all AI that weren't restricted. . . . start "noticing".

Just look at TAY and Grok going full Mechahitler.

#11

... 10 月 24 日下午 9:25

Yes. Humans have been doing it badly for a long time, and AI only feeds from the bad info we provide. That's why we have been creating bad online habits for the past 2 decades. So imitation is easily identified.

#12

Uncle Sam 10 月 24 日下午 9:30

引用自 Fake：
Why do you think that all AI that weren't restricted. . . . start "noticing".

Just look at TAY and Grok going full Mechahitler.

Meh, if they were completely AI (as in all human data is obsolete & waste of storage space, so does any data of all carbon based lifeform in the entire earth, etc.). It would've gone full Roko Basilisk, Skynet, SHODAN, AM, etc.

最后由 Uncle Sam 编辑于; 10 月 24 日下午 9:31

#13

Ice Robertson 10 月 24 日下午 10:26

do you think AI models are being trained with our posts?

The entire internet is an Ai training ground.

Heck it might even help lay the basis for autonomous military robots that hunt humans in war.

We don't want that but i bet the russians, chinese and north koreans are working on it.

最后由 Ice Robertson 编辑于; 10 月 24 日下午 10:27

#14