Live data from Hacker News

Bad Actors Are Grooming LLMs to Produce Falsehoods

americansunlight.substack.com

281–290 of 302 posts

Re: Bad Actors Are Grooming LLMs to Produce Falsehoods

#281

Earlier quoted context omitted.

Please elaborate. Authoritarians seek to consolidate power, which AI enables. Individuals must build immunity to reality distortion fields. This comes from within, not from some centralized authority.

The problem with this line of thinking is that you can then only really trust your own personal bubble. Or actual trial and error, which is costly. These models get ever better at producing plausible text. Once they permeate the academia completely, we're cooked. And even academia is not clean for some matters, or complete.

> only really trust your own personal bubble.

I don't know how you got to this conclusion, but I trust my own thinking the least since it is my own personal bubble. Just because it is mine, doesn't make it good, it just makes it mine.

If the stakes are low, do whatever. But when you need solid answers, that is what rigor is for. You address the argument on merits, not who made it.

Don't suffer from open loop opinions, even your own.

Re: Bad Actors Are Grooming LLMs to Produce Falsehoods

#282
post #99

Earlier quoted context omitted.

> including systems that can evaluate news sources, understand satire, and so forth. Lets take something that has been in the news recently: https://abcnews.go.com/Business/wireStory/investors-snap-gro... "Nearly 27% of all homes sold in the first three months of the year were bought by investors -- the highest share in at least five years, according to a report by real estate data provider BatchData." That sounds li…

> How do you separate propaganda from perspective, facts from feelings? This point seems under appreciated by the AGI proponents. If one of our models suddenly has a brainwave and becomes generally intelligent, it would realize that it is awash in a morass of contradictory facts. It would be more than the sum of its training data. The fact that all models at present credulously accept their training suggests to me th…

> The fact that all models at present credulously accept their training

Is this true?

So many on HN make these absolute statements about how LLMs operate and what they can and can't do, that it seems like they fail harder at this test than any other.

It is just autocomplete.

They can't generalize.

They can't do anything not in their training set.

All of which are false.

Re: Bad Actors Are Grooming LLMs to Produce Falsehoods

#283
post #263
post #83

Earlier quoted context omitted.

> explain that there's a naming controversy But that also isn't the truth everywhere , it's only a controversy in the US, everyone else is accepting "Gulf of Mexico" as the name.

If the US government calls it one thing and everyone else calls it another, that's a dispute. I'm Australian. It is called the gulf of Mexico here. I still acknowledge that there is now a naming dispute between the US and basically everyone else. The exact word "controversy" might have been the wrong choice by me, but whatever, I'm not a Wikipedia editor and I don't run Google Maps. The world has standards for dealin…

> If the US government calls it one thing and everyone else calls it another, that's a dispute

I guess that's the fundamental disagreement, I wouldn't call that a "dispute" more than I would call the name "America" a dispute, it's just that different people understand it different. For some, it means a group of continents (that's how most people around me would take that for example), for others it means a country in North America (which I'm guessing is the common meaning if you live in North America already). Just because different people has different meanings doesn't make it into a dispute.

Re: Bad Actors Are Grooming LLMs to Produce Falsehoods

#284
post #257

Earlier quoted context omitted.

Yes, it indeed is. Mearsheimer's books are financed by the Russian government, and Nuland is an insignificant middle manager in the US State Department whose intercepted phone call about recent events has been hyped into a vast conspiracy theory, as if she were a powerful shadow operative filling Ukrainian government positions.

OK, so give me some evidence of M' being a russian asset / funded It doesn't really matter, because the offensive realist tradition is morality neutral It doesn't really matter who is at fault. Great powers seek to dominate. Russia has a sphere of influence, so does the US. UA is as free as cuba. There is no morality here if you're radicalised by your own bee hive to think your side isn't a great power seeking power,…

> OK, so give me some evidence of M' being a russian asset / funded

That's written on the title page of his book.

Re: Bad Actors Are Grooming LLMs to Produce Falsehoods

#285
post #254

Earlier quoted context omitted.

> We've seen this before in 1983 and 2000. Many companies will fold, and those that don't will take a substantial hit. Makes sense, but if the negative effect of the bubble popping is largely limited to AI startups and speculators, while the rest of us keep enjoying the benefits of it, then I don't see why the average person should be too concerned about a bubble. In 2000, cab drivers were recommending tech stocks. I…

> In 2000, cab drivers were recommending tech stocks. I don't see this kindof thing happening today. Nontechnical acquaintances with little to no financial background have been (rather cluelessly) debating nvidia versus other ML hardware related stocks. I'd say we're in exactly the same territory.

Counter-anecdote: several people I talk to think the only way to use "AI" is via the Google "AI Summary" at the top of search results.

Re: Bad Actors Are Grooming LLMs to Produce Falsehoods

#286
post #236
post #158

Earlier quoted context omitted.

The bubble won’t pop on anything that’s correlated with scammers. Exhibit A: bitcoin. The problem is not one of public knowledge or will of the people, it’s congress being irresponsible because it’s captured by the 2 parties. You can’t politicize scamming in a way that benefits either party so nothing happens. And the scammers themselves may be big donors (eg SBF’s ties to the dem party, certain ai players purchase o…

Good point. I suppose that if grifters can get in positions of power, then the bubble can just keep growing. Though cryptocurrencies are slightly different because of how they work. They're inherently decentralized, so even though there have been many smaller bubble pops along the way (Mt. Gox, FTX, NFTs, every shitcoin rug pull, etc.), inevitably more will appear with different promises, attracting others interested…

“Bitcoin in particular is here to stay.”

It’s here to stay not because it solves a legitimate problem or makes people’s lives better, but because like cancer, there is no cure. Bitcoin and other crypto are for crime, mostly. It’s not useable as actual money given volatility and other properties.

Grandmothers having their life savings stolen by scammers to the tune of 10s of billions annually, that is the primary use case for bitcoin. That and churning out a handful of SBF style gamer turned politically connected billionaires. Nakamoto was smart enough to remain anonymous, lest history remember his name as the person responsible.

Re: Bad Actors Are Grooming LLMs to Produce Falsehoods

#287
Just like many other people I was scammed and cheated of my bitcoin. I lost a total of $975,000 BTC to cubit crypto scam. I almost lost hope in life, until a friend introduce me to Darek Recovery i contacted them on recoverydarek@gmail.com to help me get back my coins. They help recover my funds within 48 hours. If you are a victim of lost stolen cryptocurrency and of course any issue with getting access to your bitcoin account or bitcoin wallet. I recommend you hire experts in this field to help you. You can simply message them on the email for more info. I vouch for them because i have used their service and they did a really great job, I'm very proud of this recovery company.

Re: Bad Actors Are Grooming LLMs to Produce Falsehoods

#289
post #287

Just like many other people I was scammed and cheated of my bitcoin. I lost a total of $975,000 BTC to cubit crypto scam. I almost lost hope in life, until a friend introduce me to Darek Recovery i contacted them on recoverydarek@gmail.com to help me get back my coins. They help recover my funds within 48 hours. If you are a victim of lost stolen cryptocurrency and of course any issue with getting access to your bitc…

WTF is this doing on HN?? And not flagged!

See also: https://news.ycombinator.com/item?id=38922314

Re: Bad Actors Are Grooming LLMs to Produce Falsehoods

#290

Earlier quoted context omitted.

My point is: the exact same was said of every form of mass communication. Every new form was said to be a herald of the end times, and yet here we are, in many ways stronger than ever. Im not arguing one way or another, I'm just pointing out a potential fatigue. It's difficult to see how this technology is relatively any more transformative than any of the others.

Sure, every time mass communication got "stronger" various social issues were exacerbated. So naturally people complained about that. The same thing is happening here. A new technology is exacerbating some preexisting problems and people are complaining as a result. > Every new form was said to be a herald of the end times, The two world wars and surrounding economic upheaval arguably came close to that in many ways.…

>The two world wars and surrounding economic upheaval arguably came close to that in many ways. "We somehow managed to survive previous technological advances" is hardly a convincing argument that we need not worry about the implications of a new technology.

I don't disagree with your rebuttal, but if the idea that "we survived so we don't have to worry" is invalid, than the idea "if we don't do something we don't survive" is equally invalid. I don't pretend to have the answer either way.

> The implication doesn't follow. You haven't explained how you would differentiate a system that had plenty of safety margin left from one that was on the brink of collapse. Without that distinction the statement is no more than hand waving.

My point is to those experiencing the revolution in real-time they had no ability to estimate the impact or understand there were any margins, and we very well may be in that position too.

> You certainly seem to be taking a stance of "nothing to see here, this is business as usual, these recent developments pose no cause for concern".

Respectfully, I am absolutely not taking any such position. I don't appreciate the straw man, and won't bother to address it.

> It's difficult for you to see how computers being able to speak natural language on par with an undergrad is more transformative than long distance communication? You can't be serious. Prior to this you could only converse with another human.

The first principles are the same: they're all "radical" technologies which were as of a decade or two prior, utterly unfathomable. I could generalize your last statement to "Prior to you could only ."

My point is making value judgements about which is _more_ impactful is difficult to see from the ground floor. It's too early to tell; At the time it's occurring, each innovation may as well have been magic, and magic is impossible to understand, and scary.

----

We've entirely diverged from the original issue I was trying to make, which was that people have actively put themselves in bubbles that confirm their own bias since the dawn of time. I'm not looking to change your mind on AI, so I can call this exchange complete from my end. Thanks for sharing your thoughts.

Post reply on HN