Conversation
Sooo since I'm desperately trying to make an LLM export some methods via a shared library from a crate*, I'm basically making these companies loose more money! I'm the hero!**

* I'll post later about how Rust sucks at encapsulation
** Shit, it's a local LLM...

RE: https://kolektiva.social/@pivic/116935929646970100
0
0
2

@pivic @pluralistic You all realize that when it pops, the economic recession that comes with it will hit households at a devastating rate… right? I mean, I’m pro AI, from a scientific perspective, not at how it’s used now against humanity. But damn, this is going to evaporate so much freaking money… Can we plan the pop before the end of the year, and get it over with?

0
1
0

@pivic @pluralistic "passing the same $10B IOU around really fast so they can all pretend they have it" is the best quote ever.

2
0
0

@alan @pivic @pluralistic
Maybe not the best quote ever but in this context it's pretty darn good.

1
1
0

@henriksundstrom Come on, grant me a little hyperbole here ;)

1
0
0

@alan
Okey, go ahead then 😀 ☮️

0
1
0

@pivic @Viss @pluralistic This just makes me want to use AI to help it pop faster. But I don’t want to contribute to the climate crisis or accidentally make them “successful”.

2
0
0

@Taffer @pivic @pluralistic this is sorta the same argument as saying you meticulously going through your recycling bin is going to have some global impact. i promise it wont. individuals are such a tiny part of it all it wont really move the needle. its stuff like amazon doing all their front line support as llms, or government agencies doing massive llm things etc.

1
0
0

@Taffer @pivic @pluralistic yeah thats the only thing there. it'll help your feelings, and your consience, but thats basically it

1
0
0

@Taffer @pivic @Viss @pluralistic people could just decide to not use it anymore, today... but, ofc they will not do so, far to reasonable bernd

1
0
0

@usako @Taffer @pivic @pluralistic its hard to "decide not to use it" when 3/4ths of the vendors you use for nearly everything now have llm chatbots that you flat out cannot avoid using unless you stop using the service. and some of those services are like "the only isp that exists in your area", or "your pharmacy"

0
0
0

@alan
> passing the same $10B IOU around really fast so they can all pretend they have it"

It's quantum money. As long as nobody is checking which account it's in, it's in all of them : P

@pivic @pluralistic

0
0
0

@pivic @pluralistic "can't go forever" can be too long though. There is still value to extract from humans and Earth. Squeeze it till it pops. And then use the collapse to tighten the grip.

Better be sooner than later.

1
0
0

@Julianoe
> use the collapse to tighten the grip

Disaster capitalism FTW!

@pivic @pluralistic

0
0
0

@pivic @pluralistic

wait? i can make them loose money when i pay them? i... I like that?

1
0
0

@utf_7
> i can make them loose money when i pay them?

You can make them lose even more money by playing with the free toys, and not paying them a red cent. It's one of the few good arguments for using a Trained ; to find creative new ways to waste its available compute, and cost its trainer money ; )

@pivic @pluralistic

1
0
0

@Viss @Taffer @pivic @pluralistic I understand the logic and I agree individual actions are insignificant compared to large institutions, but I still believe there's value in me and other individuals rejecting AI.

As a teacher, if I casually use AI in front of my students, I give them implicit permission to use it and to become dependent on it. However, by rejecting it, loudly and vocally, I am showing my students that there is a choice not to use it. My intentional disuse of it could encourage others to reject it and walk away from companies that glorify it. Yes, that might be idealistic, but I also feel it would be naive of me to believe my choices never have an impact on anyone or anything.

1
0
0

@AlexCorby @Taffer @pivic @pluralistic i think showing it to students in a classroom setting is a materially different context than the one we're talking about

0
0
0

@strypey @utf_7 @pivic @pluralistic
If you just "use the toys" you also help them train then, which is one of the reasons why they are free.
So no

1
0
0

(1/?)

I've got a lot to say about this, but first, let's remember we are allies, discussing strategy. If we keep that in mind, this can be both more productive, and more pleasant for both of us.

@MaryMarasKittenBakery
> If you just "use the toys" you also help them train then, which is one of the reasons why they are free

They're free so people will use them, and talk about them, creating the impression of demand. This horse has well and truly bolted.

@utf_7 @pivic @pluralistic

1
0
0

(2/?)

The only thing that can reduce the harm they cause now, is bursting the "AI" bubble as quickly as possible. Costing them money is an achievable way for anyone to contribute to this.

Sure, they'll be harvesting data about our usage, and yes, that data will be used for more Training. But without comprehensive privacy law reform, the same is true for all manner of data, much of which is now harvested in ways that can't be avoided.

So why not burn their money playing with their toys?

2
0
0

(3/?)

Obviously I'm not advocating for using Trained MOLEs for serious work. The more creatively frivolous and nonsensical the use is, the better. Not only because it's more fun, but because it adds weird noise to the data being gathered. Which gets blended into any useful signal they can gather about what people actually want an AI to do for them.

1
0
0

(4/4)

Your argument also presumes that we can all continue to studiously avoid Trained MOLEs, in all their forms. At this point, they're being stuffed into everything from OS, to internet banking, to government services. Like the net itself, they're no longer an opt-in for curious geeks, and we're well past the point where waldenponding strategies are effective.

Again, the best way I can see to avoid further harm, is to try anything that might speed up the bursting of the bubble.

1
0
0

@strypey

further, using the tools doesnt necessarily mean my data getting harvested "caughts in selenium"

1
0
0

@utf_7
> "caughts in selenium"

Huh?

0
0
0

@strypey Not burn their money playing with their toys, because I still want to breathe some fresh air and do fancy the idea of a healthy planet, if that’s not already too late.

1
0
0

@gerritvanaaken
> I still want to breathe some fresh air and do fancy the idea of a healthy planet, if that’s not already too late

Me too. Which is why we need to bankrupt the Training industry ASAP. Which anyone can contribute to as an amusing hobby, by finding nonsense uses for them and *not* paying. Thus the thread you're replying to. Are we in the same conversation now?

0
0
0