gregR ☯Inference between the keyboard and the chair.
https://gts.chapek9.com/@bofh_excuses/statuses/01KKVXV79JX4QJ8EWN2S3V5H96
#ai #aiBullshit
Post by BOFH Excuses, @[email protected]
#BOFH excuse #289: Interference between the keyboard and the chair.
gts.chapek9.com
chrisestablished companies are all-in on AI slop agents because it's an easy way to pretend there's an infinite number of potential customers for their existing services
"everyone will be creating thousands of agents, each of those agents will use our service, therefore our potential market is actually 1000x our previous estimates" #makelinegoup #aibullshit
"everyone will be creating thousands of agents, each of those agents will use our service, therefore our potential market is actually 1000x our previous estimates" #makelinegoup #aibullshit
foxyoreos (they/them, commissions open)
So much of these conversations in the tech world boil down to reducing any kind of debate about morality, practicality, or even long-term utility all down to "well we want that to go well of course, but only in the context of us doing everything we were going to do anyway."
He'll concede that all of that is bad, but none of that is as important to him as whether or not he *likes using the tool*.
If he likes the tool, then not using it is off the table entirely.
And I'm supposed to pretend that's adding nuance to the debate. It's not!
To a programmer like Paul, if he likes using agents - whether they're helpful to the economy, to the average programmer, whether we know if they're de-skilling him, whether we know the effects on energy, society, power, or their role in fascism and surveillance..
And even their reframing towards ethics is designed to flatten debate, to flatten discourse about the effect of these models.
They say that they want better discussion, but they don't. They're fucking obsessed with reducing all of it down to "can it build a React website?"
It's impossible to fully disengage from unethical systems. But somepony actively going out of their way to argue that the system *should be used*, and actively arguing against the critters who are saying they want to disengage from it - I think it's fair to say that suggests some things.
I think that a lot of programmers have convinced themselves that if they *say* they have concerns about the moral implications of using these models, then that means they're good. They don't have to do anything about those concerns, they don't have to stop using Claude. Just be unhappy about it.
I view a lot of the "I share concerns" talk as purely performative. None of the programmers I see saying that use AI any differently at all from any of the open boosters and evangelists. They're identical - they just occasionally write comments and blog posts.
Like whenever I hear somepony say that local models will solve the problems with power concentration into big tech, I want to ask them, "if you're wrong, and if they don't, are you going to stop using the hosted models?"
"Do you actually believe that will happen? Enough that you'd wait for it?"