AI isn’t going to “take over the Internet”
-
AI isn’t going to “take over the Internet”
The reason you’re hearing all this shit now isn’t capabilities.
Unless they “pause”, they have no route to profitability. They don’t anyway but there’s zero chance “hyperscaling” makes them any money long term. Uptake is “not what they hoped” to say the least; the big juicy corporate clients don’t want to spend the cost of employees to replace employees as it turns out. The capabilities have been wildly overstated to impress investors. The AI hack things (exploiting common vulnerabilities) are not “AI went rogue” imo, that’s capabilities signalling to the military (their only really viable big juicy customer with tons of money left). And they got another year or so of cash left.
But they can’t say that out loud. That would be suicidal. That would mean instant calls on their debt. “Give us our fucking money back this isn’t what you promised” basically.
So they try to claim it’s a “danger”
It’s no more intrinsically dangerous than my cat on an empty stomach
@Elizafox from what I can see the real danger is the unintended consequences of it's incompetence, that seems to be the greatest danger...
-
True. AI doesn't 'take over' the Internet like hostile aliens would; it's more like they do so by enabling the darkest sides of human nature.
@martinvermeer @NicolasRinaudo @Elizafox I think that the worst thing that "AI" has done is to dilute and pollute the Internet's usefulness. I really didn't want to have to second guess everything I read. Is this real? Or is it a collection of plausible-sounding phrases diverting eyes from authoritative source material?
-
@martinvermeer @NicolasRinaudo @Elizafox I think that the worst thing that "AI" has done is to dilute and pollute the Internet's usefulness. I really didn't want to have to second guess everything I read. Is this real? Or is it a collection of plausible-sounding phrases diverting eyes from authoritative source material?
@AbramKedge @NicolasRinaudo @Elizafox
Yep. Of course disinfo existed before AI, but AI is a force multiplier making everything so much harder. If only we had made media literacy part of the curriculum everywhere...
-
@AbramKedge @NicolasRinaudo @Elizafox
Yep. Of course disinfo existed before AI, but AI is a force multiplier making everything so much harder. If only we had made media literacy part of the curriculum everywhere...
@martinvermeer @NicolasRinaudo @Elizafox malicious misinformation is relatively easy to detect - it has an agenda. Slop is more problematic, it just gets things wrong, it is unbiased incompetence.
-
J jwcph@helvede.net shared this topic
-
@Elizafox from what I can see the real danger is the unintended consequences of it's incompetence, that seems to be the greatest danger...
@Vonskinnback Same as it's ever been. Every industrial accident can be pinned down to incompetence or laziness or greed.
-
@martinvermeer @NicolasRinaudo @Elizafox malicious misinformation is relatively easy to detect - it has an agenda. Slop is more problematic, it just gets things wrong, it is unbiased incompetence.
@AbramKedge @martinvermeer @NicolasRinaudo The spiciest take I have is that AI has basically automated the SEO fucks who were already making the Internet terrible.
SEO should just categorically not be a career.
-
@AbramKedge @martinvermeer @NicolasRinaudo The spiciest take I have is that AI has basically automated the SEO fucks who were already making the Internet terrible.
SEO should just categorically not be a career.
@AbramKedge @martinvermeer @NicolasRinaudo Oh and content mills, which go hand-in-hand with the SEO bastards. I almost worked for a content mill once until I was like "fuck this, what am I even doing". There's stories online of people who worked for them. They're like the proto-slop generators.
-
@AbramKedge @martinvermeer @NicolasRinaudo The spiciest take I have is that AI has basically automated the SEO fucks who were already making the Internet terrible.
SEO should just categorically not be a career.
@Elizafox @AbramKedge @martinvermeer @NicolasRinaudo +5 Insightful.
-
Honestly this is one of those “wait you actually believe them” kind of things but people take them at face value.
@Elizafox Cory Doctorow wrote about it yesterday, explaining it is "criti-hype" and its scope is to made us "think about AI as a powerful, transformative technology, not as a better autocomplete" to mantein "those sky-high valuations" intact "until the investors can cash out."
https://pluralistic.net/2023/06/04/ayyyyyy-eyeeeee/ -
AI isn’t going to “take over the Internet”
The reason you’re hearing all this shit now isn’t capabilities.
Unless they “pause”, they have no route to profitability. They don’t anyway but there’s zero chance “hyperscaling” makes them any money long term. Uptake is “not what they hoped” to say the least; the big juicy corporate clients don’t want to spend the cost of employees to replace employees as it turns out. The capabilities have been wildly overstated to impress investors. The AI hack things (exploiting common vulnerabilities) are not “AI went rogue” imo, that’s capabilities signalling to the military (their only really viable big juicy customer with tons of money left). And they got another year or so of cash left.
But they can’t say that out loud. That would be suicidal. That would mean instant calls on their debt. “Give us our fucking money back this isn’t what you promised” basically.
So they try to claim it’s a “danger”
It’s no more intrinsically dangerous than my cat on an empty stomach
@Elizafox Not only the military, but also the civilian surveilance state. The only seemingly viable business model for data centers is to process the ever growing stream of surveillance data (video, sound recordings, internet/banking/"smart" device usage patterns etc), and sell it to governments and law enforcement as "surveillance as a service". They keep selling the AGI fantasy, but they are processing every bit of data so Palantir can eventually have every person on Earth on the kill chain.
-
S suneauken@mastodon.world shared this topic
