HN Explorer

Ask HN: What's the end-goal with AI again?

4 points by pkdpic ↗ HN
Maybe people are talking about this more and I'm just missing it but what are we trying do with AI again?

Better search engine? Better Alexa? Better autofill text on emails / sms? Better McDonalds kiosks? Automate web development? Automate art? Automate music? Automate therapy? Automate friendships? Automate all human labor? Star Trek computers? Machines of loving grace? Skynet?

10 comments

[ 4.7 ms ] story [ 34.8 ms ] thread
brucethemoose2
More money.
pkdpic
But for who? Jeff bezos? The Yang Gang UBI?
krapp
Whomever can replace their employers with AIs.
UmYeahNo
Putting as many people as possible out of work, in an ever-accelerating cascade.
CatWChainsaw
Every option you've listed is probably the end goal of somebody, somewhere. Whether or not some/any of them are good/advisable is a separate question.
pkdpic
Yeah good point. Also I forgot "boot stamping on a human face forever"
mindcrime
There is no (unified) "we" in this context, as different people certainly have different goals for AI. I can only speak for myself, but my interest in AI is rooted in the idea that we can use AI to assist humans in solving harder problems than they can solve on their own. I forget who said it, but there was a quote I heard once that can be paraphrased as "Step 1, invent AI. Step 2, use AI to solve (climate change|hunger|poverty|disease|whatever)". That's pretty much the angle I approach this from.
pkdpic
I'm hoping that's the unstated underlying opinion the wider dev community is operating on. That's enough to satisfy me.
turtleyacht
I thought AI had promise when it was "augmented intelligence," like genetic engineering--finding optimizations in things, like a novel circuit design or better aircraft wing.

I hope we use this new AI to model branching narratives based on assumptions to come to useful conclusions. Instead of having to code these data structures by hand, we have a kind of "malleable associative array" that we can query.

But it starts with putting an "AI in every home," with whatever the 21st century equivalent of 64K memory is, and let it store ideas in private.

Given this wonderful thing, we should be able to quickly onboard ourselves all the way to the frontiers of human knowledge, and begin to narrow our study to dedicate on the leverage points for (list|of|hard|problems), in a conversational way.