What's the deal with all those weird wrong-number texts? • Hacker News Board

Re: What's the deal with all those weird wrong-number texts?

#101

Post by drekipus » Fri, Jul 01, 2022, 8:41 PM UTC

spbaar wrote: post #80

I've been a bit unemployed and bored so I've responded to a few of these to see what the angle is. There are also a bunch on telegram. Strangely, most do not push hard at all and drop off. They may mention crypto in passing but thats it. One time a scammer said hey can you help me an make an account for this crpyto scam website. When i said i did they said great and nothing else.

There's a bit of an art to stringing them along. You gotta act like the damsel in distress some times.

I got added to some whatsapp group investment scam once, when there was multiple scammers targeting me it was easier to do because they all didn't want to give up on me. Then they seemed to invite "big brother" so I guess was some leader who had much better English, and he was trying to prove the legitimacy of the platform. Went for about a week and a bit from memory. I sent them pictures of my daily walks and asked how much better it would be if was so rich I didn't have to walk at all!

Re: What's the deal with all those weird wrong-number texts?

#102

Post by wpietri » Fri, Jul 01, 2022, 8:43 PM UTC

stepanhruda wrote: post #81

jerf wrote: post #28

Earlier quoted context omitted.

"Train the scambot to perfectly push your buttons by having it practice against your own style of thought as embodied by your social corpus." Fortunately, that's not really what these language models can do. They can easily be trained to mimic you. They can be trained to mimic what normal people reply to you with. But there's no way to train the transformer-based high-probability-next-word AIs to be superhumanly good…

You don’t need to automate the whole process, just use language models to establish rapport for a few weeks and have humans pick up the gullible ones at the bottom of the funnel.

For sure. Or even for a few days to start. It's basically the same playbook as Waymo: get computers to do more and more of the boring parts, having human operators take over when necessary, and using the additional data generated to improve the system.

Re: What's the deal with all those weird wrong-number texts?

#103

Post by mcculley » Fri, Jul 01, 2022, 8:45 PM UTC

wpietri wrote: post #68

Animats wrote: post #7

In this case, the victim deposited the money into a fake crypto platform that told him his investments were performing well, presumably to entice him to deposit even more. Of course, once he tried to withdraw the money, he found he was unable to. That sounds just like the "binary option" business which used to be run out of Israel. The Times of Israel blew that apart with "The Wolves of Tel Aviv" investigation series…

> Scamming people outside Israel was legal in Israel at the time. My head just exploded here. I think I need to go lie down. Please nobody tell the Florida legislature about this.

Florida has been selling swampland to gullible retirees from other states for a long time.

Re: What's the deal with all those weird wrong-number texts?

#104

Post by Transfinity » Fri, Jul 01, 2022, 8:45 PM UTC

jerf wrote: post #28

ctoth wrote: post #23

Terrified to consider what happens when these scammers get hold of large language models here in a year or so. Rather than fading into the background as this article posits, I expect people to have models finetuned on convincing them to make purchases/send money. Probably trained by being pitted against other models which have been trained on the mark's social media feeds. Train the scambot to perfectly push your but…

"Train the scambot to perfectly push your buttons by having it practice against your own style of thought as embodied by your social corpus." Fortunately, that's not really what these language models can do. They can easily be trained to mimic you. They can be trained to mimic what normal people reply to you with. But there's no way to train the transformer-based high-probability-next-word AIs to be superhumanly good…

> But there's no way to train the transformer-based high-probability-next-word AIs to be superhumanly good at fooling you into doing something, on the grounds of lack of training data

The conversations of all those human scammers would be prefect training data for this. You even know exactly what conversations led to payouts. Assuming you can get all your data in one place, of course.

Re: What's the deal with all those weird wrong-number texts?

#105

Post by notahacker » Fri, Jul 01, 2022, 8:49 PM UTC

bigodbiel wrote: post #94

kristopolous wrote: post #42

Earlier quoted context omitted.

I've long assumed they do the exact opposite - try to filter out people who likely see through the game so they don't waste their time mining a hill with no gold. And they do this by intentionally making basic mistakes or other easy to spot errors so the clever people will just see themselves out and by the time their funnel gets to an actual human scammer, they have a highly probable sucker.

Exactly! Even Microsoft had a paper on this ‘Why do Nigerian Scammers Say They are from Nigeria?’ [1]. 'By sending an email that repels all but the most gullible the scammer gets the most promising marks to self-select, and tilts the true to false positive ratio in his favor.’ [1] https://www.microsoft.com/en-us/research/wp-content/uploads/...

A theory which would be more convincing except that [i] saying they're from Nigeria also filters out all the gullible people with spam filters, and yet despite spam filters now preventing the majority of gullible people from responding, the scripts haven't changed [ii] the more straightforward explanation is that they say they're from Nigeria because their ultimate objective is getting you to send money to Nigeria...

Ultimately if you're in the business of spamming people on the other side of the world in the hope that 0.001% of them will ultimately send a money transfer worth a month's wages in local currency, your time probably isn't so valuable you can't afford to deal with everyone that replies

Re: What's the deal with all those weird wrong-number texts?

#106

Post by techsupporter » Fri, Jul 01, 2022, 8:50 PM UTC

muststopmyths wrote: post #98

deepspace wrote: post #43

Earlier quoted context omitted.

What would the motivation be for anyone to reply to a text from an unknown person? I have also been getting these for months, in it was 100% obvious from the get-go that it must be a scam. I have never felt the slightest urge to reply.

I guess some of us are old enough that wrong numbers were a thing and it was just polite to let them know. Especially when it sounds legit

Considering how almost every form of social trust has been abused to either spam or scam me, I am old enough to remember that but have also had that politeness completely scrubbed from me.

I wish we still lived in a time when being scammed or spammed wasn't so prevalent, but we don't and I'm not doing anything that might tip off either of those groups that my number is ripe for the picking.

Re: What's the deal with all those weird wrong-number texts?

#107

Post by RajT88 » Fri, Jul 01, 2022, 8:51 PM UTC

I had one of these exchanges recently which I had some fun with, because it was so weird, and they made a point to use a photo of a beautiful woman:

Lady: Doctor Lucy? My puppy is very slow and does not eat dog food, can you make an appointment for me?

Me: Unfortunately I do not treat puppies. Only adult dogs, adolescent foxes, and elderly coyotes.

Lady: Sorry, I added the wrong person, I just checked the number and I saved the wrong number for Dr. Lucy.

I left it at that. I wonder if I should try and bait the scammer. I do love messing with scammers. After the last message I was thinking maybe it was legitimately a wrong number.

Re: What's the deal with all those weird wrong-number texts?

#108

Post by RajT88 » Fri, Jul 01, 2022, 8:52 PM UTC

cmg wrote: post #74

avalys wrote: post #29

Wow, how interesting. I started getting these a few months ago, right after starting a new job in a somewhat high-profile area. I have been getting 1-2 per month but found them quite unusual because none of them ever progressed to an actual scam - just a few messages back and forth about the fabricated "wrong number" situation and then each thread would stop, seemingly earnestly. Since there was no obvious scam, I've…

I've gotten a few in the last couple months, with a US cellular number that I've had for over 20 years. At least two are "I missed you at the gym today!" via text and another was from a Vermont area code via WhatsApp to a "Dr. Jack": > Hi Dr. Jack! My cat is very slow and does not eat cat food. Can you make an appointment for me? > Sorry, you have the wrong number > Sorry, I just checked, I entered a wrong number, pl…

I had almost the exact same exchange, with someone with a photo just like that! 747 area code, was it?

Re: What's the deal with all those weird wrong-number texts?

#109

Post by themaninthedark » Fri, Jul 01, 2022, 8:53 PM UTC

muststopmyths wrote: post #97

I got one of these, almost convincing enough except she said she lived in SF. I asked where in SF and she said Alcatraz. I almost wanted to keep the conversation going just to hear more about her life on the island

Almost sounds like a cry for help. O_o Probably not though.

Re: What's the deal with all those weird wrong-number texts?

#110

Post by notahacker » Fri, Jul 01, 2022, 8:54 PM UTC

ctoth wrote: post #69

cool_dude85 wrote: post #51

Earlier quoted context omitted.

They have, per the article, buildings full of slaves to do the selling. I doubt GPT-4 will be able to compete on price with that.

Per the article the scammers pay a minimum of $8,000/person, plus cost to feed, imprison, etc. Pretty sure that a model that only requires electricity and GPUs to run will work out being less expensive than this especially when you consider that GPT-N (YaLM-1T?) will be able to run as many scams as you have GPUs to run inference on concurrently, increasing your possible take, and won't have to sleep.

I think we can probably rule out OpenAI and equivalent cloud services allowing people to use their APIs to run phone scams. It's even worse PR than bots saying racist things.

And if they need to train their own model, you can get a lot of slaves and poor wannabes for the price of one competent NLP engineer, and the slaves and poor wannabes are less likely to decide they're the brains of the outfit and cut you out of the loop.