Hacker News Reader

Comment by irln | Hacker News Reader

I'm reminded of this thread The FCC Has Fined Robocallers $208M, Collected $6,790 https://news.ycombinator.com/item?id=21492782.

Enforce collection of the fines in the first place would be a good start.

View on HN

Quick solution: fine phone companies instead, since it's easy to collect from them. Give 'em 6 months notice and I bet spam and scam calls are nearly eliminated before the notice period's up and you barely end up collecting any fines.

It's a dangerous path when you force intermediaries to intervene and apply censorship. The path to having ML algorithms analyze every single call you make and receive is paved with intellectual laziness. Would you make the same arguments about end to end encryption? Quick, let's put backdoors in everything so we can feel safer.

Simpler solution: At the end of a call, the recipient can dial some number, like * SPAM, and the phone company who routed the call to your number is charged 25 cents. Presumably the routing phone company will pass this charge back down the chain until it reaches the caller. You must dial * SPAM within 2 minutes from the end of the call for the charge to take effect.

There is no appeal. If you don't like a call for any reason, you can dial * SPAM, and the caller is charged 25 cents. Since the report must be immediate, VOIP providers only need to require a deposit for two minutes of calling volume, and they can throttle their customers to make sure that the reserve is always present. Callers don't need to be verified or have good credit, they just need to have 25 cents.

Now you're thinking!

However, the financial layer needs to be set up before the call is initiated, not after.

Otherwise, it's ripe for abuse.

Also, making end-users perform extra actions is just BLEH!

Sounds like an easily abused system.

Yup, my first thought was that I could fine my friends $0.25 for calling me... You know, just for fun. If a company calls me and I don't like what they said, $0.25 fine. Someone dials a wrong number, how dare they, $0.25 fine.

Even better, let everyone set the fee for calling them and the ability to set up a whitelist. Numbers not on the whitelist get charged (message before connecting stating the amount). I'd set it to $10. To each their own.

I like the flat 25 cents. It's not so much that if someone wants to be malicious that they can do much damage, but if robocallers have to pay 25 cents, it destroys the economics.

Since you have to object to a call within a short period of time, you could get an immediate text message that someone had flagged your call and just not call that person again.

If you were making legitimate high volume calls, you'd get a certain number of people doing this to you, and then you'd just flag it and not call them again or maybe stop doing business with them. Either it's a sign that they don't want you calling them, so you should probably stop.

Collection agencies and such would get a lot of flagged calls, but collection agencies are supposed to stop calling if you ask them to, so this is just an automated way of making that request.

The real cause of all of these robocalls is that calling has gotten so cheap. It used to cost well over $1/minute to call from say, India to the U.S. Even within the U.S., in the 80s and early 90s, it would usually cost about 25 cents a minute to make a domestic long distance call. In many places, including where I grew up, a local call to your next door neighbor cost 5 cents. If the risk of being charged 25 cents for a call is a barrier to making a call, you probably shouldn't be making it.

You could, but you wouldn't have many friends left after a while. And it's not like the money would go to you.

Features.

(Think about it.)

Yes, because it was described in a comment with only a few minutes' worth of thought. The idea is sound - there are a ton of fairly obvious ways to fix the biggest flaws and improve the system.

Figure out how many calls the big offenders need to be profitable, and set the amount accordingly. Maybe a penny would be enough.

How?

But it is also very dangerous to do what telco's did, they connected circuit network that was designed as a network of trusted peers to the untrusted network of public internet without implementing proper gateways and/or enforcing certain rules. So slapping them with fines is more of a reap what you sow thing than punishment.

Why on Earth do you think backdoors and ML having any part in the solution?

Telcos peer with one another and trust each other's traffic implicitly, all they need to do is stop trusting implicitly and ask for verification for things like caller id.

A Telco may not know where a call originates but they know what Telco it's coming from and can pass responsibility for fines back to that Telco.

Since no Telco wants to be saddled with fines, they'll quickly pass it off and it will eventually reveal the origination.

Requiring telcos to assist in fighting spam calls and wire fraud is not "censorship" by even the most lenient use of the word - it's saving elderly or non-techie citizens from being stolen from. Nobody mentioned backdoors or ML algorithms, either - stop strawmanning.

No it's not. Why would you want to give telcos even more power? Let dumb pipes be dumb pipes. The whole idea that telcos should moderate incoming calls is just begging for for non-call-neutrality and pay-for-play. HN is obsessed with net-neutrality but apparently the same sentiment does not extend to other communication platforms.

No it's not. Telecoms allowing to use their platform for abuse and doing nothing about it because it helps their business.

Similar with Amazon allowing seller selling fakes and counterfeits.

So holding them responsible is a wake up call for them where doing nothing from their side is no longer an option.

Telecoms absolutely have powers and know-hows to detect and deter abusive practices.

As someone who has been involved in telco routing / call infrastructure, I agree and disagree. Most telcos hate that their networks are being flogged by these short duration calls. When they have the engineering competence / business ethics / operational focus in place, solutions emerge and have been around for awhile. You won’t get them all but simple heuristics that’s others have described above will squash the big abusers. The problem is these spam artists, prey on the fact that like the old internet of twenty years ago, a bad actor could relatively easily exploit weaker networks. Like most things I expect the first 80% of this problem will get resolved pretty quickly and then we will see some sort of vendor driven standards that become available / implemented within the next year.

Edit: I think one interim solution could be a dcma-style process whereby carriers prove to the fcc that they maintain an auditable takedown process that can be initiated by the consumer and that the carrier is applying reasonable technological solutions to minimize / detect en masse robocalls.

You don’t fine the phone company. You fine the Broadband.com and the Twilio.com of the world who enable robocalling.

Twilio has a market cap of ~$20B, which is largely fueled by them enabling robocallers to text/call people.

To me, the obvious solution is to force providers to provide originator information (including name, number/ip, phone company, and country) to the end user to matter what. Then I can block whatever I want with a UBO style list.

This is not a problem you can solve with tech or more info for the consumer. I had an endpoint with two names, IP which could change mid-call, 3 possible origin numbers, and the "phone company" was a blurry concept in that case (ITSP, or its first upstream, or first POTS gateway?). And I was just a consumer. This gets even more complicated for any non-trivial company.

And that's still before name collisions. You can get a valid call from Bill Gates at Microsoft. Just not the one you think.

There are easier ways: whitelisting. Android and iOS already have a ton of shenanigans in their SMS apps to deal with stuff like RCS. A lot of people also now block and ignore calls not from a familiar caller id. Exchanging contacts can be tied with exchanging QR codes to allow auth, a practice that has been normalized by social media companies like Snapchat. This is relatively scalable. Same thing for corporations, they can post their codes online. For the non-technical an additional manual entry code can be added to automated telephone systems. For good or for bad, the mobile operating systems market has generally consolidated around a few big players. Stuff like Sailfish, KaiOS are still very very small players. Software updates can be easily and cheaply deployed OTA.

This is a high tech solution which just won't work for lots of people. Example: The local medical practice uses simple mobile phones. Think old Nokia with 3-line menu to display and no camera. You want a call back from a doctor on call and have no internet access. How would you do any pre-auth/whitelist in that case? How would you do it if you only had a land-line on POTS?

As a consumer, just give me the full stack, and any step that involves something sketchy means I don't want to see it. I'd block anything from outside the US, anything from a network that obscures information, and maybe all VOIP numbers as well. Then at least I'd only be contactable by people subject to United States law.

I already keep my cell phone on silent unless I'm expecting a call, and my voicemail is disabled. It would just be nice to not have to be notified about other calls when I do need to be reachable by phone.

This is not workable for an average person though. Also people just don't get how telephony works today. Here's an example: which of these is a call from out of US:

- call from a person in Canada, legally using a valid US caller id

- call from a US mobile roaming in Canada

- call originating in the US, routed via Canadian telco back to the US

- call from a US mobile in the US, using a VoIP service in Canada, legally using the same mobile's US caller id

(None of these are doing anything illegal, and are valid scenarios)

Yeah perhaps I don't need to block absolutely anything outside the US (and Canada), but clearly at some point there's an initial caller. If at any point I hit a hop that's not trustworthy, I don't want the call. If you can't reach me without resembling a telemarketer, I probably don't need the call. If legal but irritating companies find themselves unable to serve calls, I consider this to be a positive. If the phone system collapses, I consider that the be an improvement over having to sometimes be reachable over the current phone system.

Blocking all of those would be vastly preferable to what we have now. I get about 5 spam calls a day, so the only thing I can do is disable all incoming calls altogether. Blocking all of the things in your list, while hamfisted and far from ideal, is better than that.

> The bill requires carriers to verify that calls are legitimate before they reach consumers;

That's in general what the bill provides. What else were you expecting to be enforced?

why not just tax all phone calls with a 1 cent tax. that should stop most robocalls.

Reply on news.ycombinator.com