Hacker News Reader

Comment by yamtaddle | Hacker News Reader

The ones who will benefit the most from AI, in terms of being able to scale the most with it, will be the ones who care very little about reputational risk and have relatively high tolerance for mistakes and misses—that is, those who can operate these kinds of models without needing to carefully check & monitor all of its output.

... so, scammers, spammers, and astroturf-heavy propaganda operations. This tech may 5x or 10x a lot of jobs' output over the next few years, but it can 100x or more those, and it can do it sooner.

View on HN

Your criteria also all match for authoritarian regimes. Many of which love AI based surveillance solutions.

Speaking of which, George Soros had this to say about China in early 2022:

https://www.project-syndicate.org/onpoint/china-and-xi-jinpi...

"China has turned its tech platforms into national champions; the US is more hesitant to do so, because it worries about their effect on the freedom of the individual. These different attitudes shed new light on the conflict between the two systems of governance that the US and China represent.

In theory, AI is morally and ethically neutral; it can be used for good or bad. But in practice, its effect is asymmetric.

AI is particularly good at producing instruments that help repressive regimes and endanger open societies. Interestingly, the coronavirus reinforced the advantage repressive regimes enjoy by legitimizing the use of personal data for purposes of public control.

With these advantages, one might think that Xi, who collects personal data for the surveillance of his citizens more aggressively than any other ruler in history, is bound to be successful. He certainly thinks so, and many people believe him. I do not."

I agree the effects are asymmetric. an important consideration behind all of this is that there is very little that government or power can do to extend the lives of humans, and very much they can do to shorten them. When we thing of "good vs bad" it is important to consider how permanently malevolent many choices can be and how limited is our ability to recover. we can never undo an unjust loss of life. Technology as an amplifier for human behavior always more _potential_ for harm in the wrong hands and applications.

As people, we are very sanguine about the way technology can improve things, and we have see a lot of real improvement. But we're not out of the woods yet.

I find it funny how people are concerned about government surveillance while completely ignoring the one which is being used in a very totalitarian way already for many: Surveillance at the workplace in all its many forms. Companies have a very strict hierarchy, meaning from a political aspect they are, by and large, authoritarian, and worse, there is often some unspoken "class" level thinking in it that is used to reason such surveillance into place (among other things). Even worse, there are many places where private entities and government ones flow into each other, muddling the already dirty waters even more. So it is not just governments to be worried about, it is every, especially larger, entity where the people in command have some screws loose (which covers a large percentage of them).

Company is paying us, vs we are paying government. I can resign and get out of the relationship with my company. No such luck with government lol. So the dynamic is not quite the same.

We are technically free to forfeit our citizenship and move to another country, but the majority of us are not even afforded the luxury of quitting a poor job.

Oh, yes, that's definitely occurred to me. :-(

That case is a bit different because authoritarian regimes also don't want to lose control over their surveillance infrastructure. So I would expect increased use of AI to process data, but much less so of anything that involves the AI proactively running those systems.

The militaries, OTOH, are going to hook that shit into fully autonomous drones and such.

I think authoritarian regimes hooking AI up to process data is already bad enough, I don't think having the AI "run" those systems is even being discussed in the article, the opposite really, instead of sci fi scenarios of an AI a la HALL 9000 making decisions, its really just allowing for a more powerful black box to "filter out the undesirables du jour".

And I think we could extend the definition of authoritarian regime to include any and all private interests/companies that can leverage this

My whole point is that losing control (the Terminator scenario) isn't what we should realistically fear. AI completely under control but following the orders of a ruthless authoritarian regime is really, really bad. We don't need to have faith in some hypothetical godlike paperclip maximizer AI to see that.

Losing control is not necessarily a "Terminator scenario", and it's unfair to look solely at the most extreme case and its probability to dismiss it. A far more likely case that I can see is subtle bias from the AI injected into decision making processes - it wouldn't even need to be directly wired into anything for that, although that helps. This doesn't require the AI to have agency, even.

USSR collapsed in part because the system was entirely planning-centric, and so many people just wrote straight-up bullshit in reports, for reasons ranging from trying to avoid punishment for not meeting goals to trying to cover up personal embezzlement and such. An AI doesn't even need a reason to do that - it can just hallucinate.

"Alignment is unsolved for humans, too."

Indeed, but we have literally tens of thousands of years of experience with behaviorally modern humans, so we can usually spot the signs much better.

I would add startups to the list in addition to the more explicitly negative groups that you mention. There's a tremendous amount of both reputational risk and technical/organizational inertia to overcome in Amazon overhauling Alexa to use LLMs for responses, but much less for a new startup because they have far less to lose. Same for Apple with Siri, or Google with Google Assistant. I expect that we'll see exciting changes in these products before too long, but they're inherently going to be more conservative about existing products than new competitors will be.

Yes and no. That same inertia of the larger companies allows them to stay in business even after large missteps. Remember Lenovo built-in Superfish [0]? Amazon/Apple/Google are entrenched enough that can take risks, apologize, rollback, and wait for short term memory to pass and release the v2 under a different product name.

Across all startups, yes they can be risky and there will be innovative outliers that take off. But for a single no-name startup a misstep can be a death knell.

[0] https://en.wikipedia.org/wiki/Superfish

I would say that intelligence agencies, militaries, and authoritarian regimes could also meet those criteria because they operate outside of public scrutiny.

Increasing numbers of scammers hit diminishing returns though. There is only so much money in the hands of the vulnerable/stupid.

Reply on news.ycombinator.com