Ask HN: resources on AI uncomfortable topics and practical solutions | Hacker News Reader

Ask HN: resources on AI uncomfortable topics and practical solutions

17 pointsluigi238 comments

Hey HN folks, I'm trying to find articles, books, or any other resources that dive into the messy side of AI, like job losses, inequality, and effects on poorer countries. I'm really interested in stuff that offers down-to-earth, actionable ideas without getting all preachy or just throwing regulations at the problem. Don't get me wrong, I know regulations are needed, but I'm tired of reading about obvious points that might never actually happen. I'm open to solutions that may also never happen or might be uncomfortable and politically incorrect. I want resources that talk about real steps that people, organizations, or communities can take, even if they're controversial. If you've got any good recommendations, i'd love to hear em. cheers!

View on HN

> Don't get me wrong, I know regulations are needed

Open calls for regulation are not productive. Regulation should be a last resort, not a knee-jerk response. What do you think needs regulation and more importantly, how would you effectively enforce these regulations. A law is policy + enforcement after all, and a poorly crafted AI regulation will achieve nothing except shifting the balance of power to entrenched groups.

Open calls for deregulation are not productive. Deregulation should be a last resort, not a knee-jerk reaction. What do you think needs deregulation and how would you effectively make sure we ensure a decent quality of life for everyone in society? A lack of regulation is policy after all, and the lack of regulations will achieve nothing but shifting the balance of power to entrenched groups.

This is a poor attempt at trolling because the default state of nature is deregulation. Calling deregulation “reactionary” is just icing on the cake.

This is a poor attempt at getting the point because that can be a problem with disruptive technology that is immediately deployed to a mass market without any care or concern for lasting consequences as long as someone gets rich

You haven’t attempted to address that

> Deregulation is the default

> There is little to no consensus on possible modes of regulation at nearly all levels of the AI pipeline

Your suggestion is reactionary at best and the F word at worst, allowing the govt to dictate who can try new enterprises. Govt isn’t your savior and the attitude that they will solve your problems is misguided.

To be fair to you, I am being snarky and not explicit about what bothered me about your initial comment. You’re right, I haven’t addressed it because it’s not my point, and yeah I am kind of trolling you.

I actually don’t care about AI regulation specifically and this is definitely not about ‘me’ or ‘you’.

I just generally think the technology industry has proved time and time it doesn’t take any responsibility for bad outcomes. AI putting millions out of jobs is a GIGANTIC problem for everyone. So, your comment that seemed blindly in favor of the status quo bothered me. I think leaders in tech need to engage more with elected officials to understand how to make society better for everyone. If you study history, you know that real revolutions (not the type of ‘revolutions’ tech billionaires claim to lead) are about material conditions.

It disturbs me that so many people in tech don’t care about what happens to working people, it seems like most want to just treat the world as a playground, invent ‘cool’ things, get rich, screw everything else.

That’s my point - the draw of regulation in the abstract is about being able to ensure we build a fairer society. I obviously respect your agency and opinions. I just think we have way too many decison makers who think like you, and it’s creating a certain debt that a generation someday will have to pay, and it won’t be pretty.

I'd like more info as well and am actively seeking. So far I've started an AI company and it's looking like the business will focus on custom AI development for companies with the primary focus being ethical implementation of the AI. Creating new jobs and optimizing the current ones rather than outright replacing people with AI which could be comically disastrous anyway and rightfully so. People need to work and I don't necessarily mean just for money or to survive but because work is deeply ingrained into our very being whether we want to work or not. I'm not sure how much of a mark my company will make but I'm hoping that it helps push this tech into a better direction than a tech dystopia ran by either corrupt governments, corrupt corporations, or (most likely) a little of both.

> effects on poorer countries

For what it's worth, I would predict these effects to be dramatically positive: because LLMs basically immediately counter the effects of being a non-native speaker of your business language, and because the relative change in economic productivity is far greater than the relative change an already-trained-and-fluent worker gets.

All of which is to say, it is reasonable to be concerned about political and economic change, but if your starting position is one of caring more about global solidarity than job loss in the richest countries, I think it's more likely that your concern should be in the direction of "how do I make sure this stays legal and cheaply available everywhere?", rather than the direction of regulation.

Reply on news.ycombinator.com