Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Future of Everything Is Lies, I Guess: Safety

aphyr.com

Post reply on HN
Print view View original

181–190 of 195 posts

Re: The Future of Everything Is Lies, I Guess: Safety

#181

Post by simianwords »

intended wrote: post #172
simianwords wrote: post #67

Earlier quoted context omitted.

Its usually the educated and elite PMC types who have grievance with technology. They secured their status and have lucrative jobs mostly with the help of technology and they are too scared to have anything threaten their position in society. It is highly hypocritical to behave this way but they don't seem to have the self awareness to observe it objectively. Ask any poor person in India what their sentiment is with…

I’m in India, and I sure as shit haven’t seen what you are talking about. In 2025, we lost 22931 crores to cyber fraud - about 2.7 billion USD. People are now saying that they are relieved if the losses were only single digit crores lost. India invented digital house arrests. There’s entire districts/cities where the primary revenue stream is from scams. Cops don’t want to involve themselves with cyber crimes because…

It’s easy to fall into the trap of overindexing on local issues. On a holistic level internet brings people to the same level by democratising knowledge.

I’ll ask you this: would India be better off without internet? If your ultimate goal were democracy, would you end internet to promote democracy in India?

Re: The Future of Everything Is Lies, I Guess: Safety

#182

Post by everyday7732 »

Bombthecat wrote: post #180
philipkglass wrote: post #36

In short, the ML industry is creating the conditions under which anyone with sufficient funds can train an unaligned model. Rather than raise the bar against malicious AI, ML companies have lowered it. This is true, and I believe that the "sufficient funds" threshold will keep dropping too. It's a relief more than a concern, because I don't trust that big models from American or Chinese labs will always be aligned wi…

Uhhh you know that the paperclip problem stems from ai just following a task, not understanding what it is doing? Not from being misalignment. I would go out a limb and say that current a could create a paperclip problem, given powerful enough tools.

The argument is that it's misaligned because it only values one thing: more paperclips, while human values are much more varied and complex.

Debatable whether it truly understands what it's doing or not, but the argument usually assumes that it does know what it's doing at least in that it's able to imagine outcomes and create plans to reach its singular goal, making it a very simple toy example of a misaligned system.

Re: The Future of Everything Is Lies, I Guess: Safety

#183

Post by munksbeer »

aphyr wrote: post #133
cs02rm0 wrote: post #111

Earlier quoted context omitted.

Although Ofcom doesn't think geo blocking is sufficient to absolve them of that liability. Crazy as that is.

I actually wound up geoblocking the UK based on Ofcom's February 2025 presentation for small services providers--they said that they intended to target "one-man bands" who (e.g.) failed to perform a child risk assessment or age verification, but that a geoblock would be considered compliant. I don't like doing this, but as someone who visits the UK regularly (and has been regularly pushing Ofcom on this matter) I fig…

I'm glad you have done this and I wish more would follow the same course. The more content that becomes unavailable in the UK, the more people might start to pay attention to the stupidity of the law.

I doubt it, but even from an irrational anger perspective, I hate that these idiots can do idiotic (and worse, counter productive) stuff, and get no comeback on themselves.

Re: The Future of Everything Is Lies, I Guess: Safety

#185

Post by dns_snek »

_DeadFred_ wrote: post #122
dns_snek wrote: post #100

Earlier quoted context omitted.

Broad alignment =/= Wealth maximization.

The market aligned us with children working in sweat shops after we outlawed it by convincing us it was OK if it was foreign kids and we got to share in pocketing the savings not just the evil factory owner.

Yes I'm well aware. Of course that's not how things are advertised to people, and they absolutely hate it when this is pointed out to them. This tells me that deep down they don't actually agree with how the system operates.

Re: The Future of Everything Is Lies, I Guess: Safety

#186

Post by jkman »

schnitzelstoat wrote: post #184

It's a tool, some people use the tool to do bad things. But they already did bad things before. Virtually all of the arguments here could also be applied against the Internet itself.

That's a lazy argument. Obviously tools are tools. But if tool A revolutionized human society and has massively advanced technology (and CAN be used for harm), where tool B's positive impact is a drop in the bucket by comparison and has the potential for an outsized amount of harm, obviously tool B is comparatively a bad tool.

Re: The Future of Everything Is Lies, I Guess: Safety

#187

Post by bigfishrunning »

themafia wrote: post #129
bigfishrunning wrote: post #98

Earlier quoted context omitted.

> Meanwhile.. just.. ask an LLM if you can mix certain cleaning chemicals safely. the cost of the wrong answer to this question is so incredibly high that I hope nobody is sincerely asking an LLM for this information. The things people trust to "machine that gives convincing answers that are correct 90% of the time" continue to shock me

> is so incredibly high that I hope nobody is sincerely asking an LLM for this information Google trumps the search results with it's LLM box. There's only one reason to do that. They know their audience is not engaging in discretion. > The things people trust to "machine that gives convincing answers that are correct 90% of the time" continue to shock me People are having intimate relationships with chat bots. There…

The liability of google's search box saying "Ammonia and bleach mix to make a great cleaning agent!" (disclaimer: please don't do that it will kill you) seems really high. I feel like we're all living in crazy world.

Re: The Future of Everything Is Lies, I Guess: Safety

#188

Post by jagged-chisel »

seizethecheese wrote: post #124
jagged-chisel wrote: post #46

"Alignment" In what world would I ever expect a commercial (or governmental) entity to have precise alignment with me personally, or even with my own business? I argue those relationships are necessarily adversarial, and trusting anyone else to align their "AI" tool to my goals, needs, and/or desires is a recipe for having my livelihood completely reassigned into someone else's wallet.

Why would relationships with a commercial entity be "necessarily adversarial"? A commercial relationship depends on the product providing more utility than the cost (for the consumer) and providing more revenue than cost (for the commercial entity). This means that while some components of the relationship may be adversarial in some areas, it cannot really be entirely adversarial.

> Why would relationships with a commercial entity be "necessarily adversarial"?

Because they want to separate me from as much of my money as they can, and I want to keep as much of my money as I can.

Re: The Future of Everything Is Lies, I Guess: Safety

#189

Post by Jtarii »

munksbeer wrote: post #183
aphyr wrote: post #133

Earlier quoted context omitted.

I actually wound up geoblocking the UK based on Ofcom's February 2025 presentation for small services providers--they said that they intended to target "one-man bands" who (e.g.) failed to perform a child risk assessment or age verification, but that a geoblock would be considered compliant. I don't like doing this, but as someone who visits the UK regularly (and has been regularly pushing Ofcom on this matter) I fig…

I'm glad you have done this and I wish more would follow the same course. The more content that becomes unavailable in the UK, the more people might start to pay attention to the stupidity of the law. I doubt it, but even from an irrational anger perspective, I hate that these idiots can do idiotic (and worse, counter productive) stuff, and get no comeback on themselves.

>I'm glad you have done this and I wish more would follow the same course. The more content that becomes unavailable in the UK, the more people might start to pay attention to the stupidity of the law.

The law isn't going to be repealed because a bunch of nerds geoblocked their personal blog.

Re: The Future of Everything Is Lies, I Guess: Safety

#190

Post by reasonableklout »

zozbot234 wrote: post #99
tyrust wrote: post #88

Earlier quoted context omitted.

> precise alignment with me personally, or even with my own business Seems like a strawman, I don't think anyone means this when talking about alignment. More general goals, like avoiding paperclip maximization, are broadly applicable to humanity.

If you've built an agent that can act even vaguely close to a paperclip maximizer, you've already solved 99.999% or more of the alignment problem. The hard part of alignment so far is getting the AI to do something useful in pursuit of the right goal, and not just waste energy. We still have no idea how to do this with any effectiveness: even modern "RL from verified feedback" systems are effectively toys, the equiva…

Huh? Modern RLVR systems are toys that can’t do anything useful in the real world?

We must be living in completely different worlds. Claude and other agents have completely upended work for me and every single other software engineer I know.

Post reply on HN