Outbound SIP calling is now available to all organizations | Hacker News Reader

Outbound SIP calling is now available to all organizations

developers.openai.com

76 pointsarbayi65 comments

Small reminder from Germany: I’ve been using my mobile number since early 2000s, provide it as contact info in lots of places and sign up forms, and have received maybe 10 spam SMS overall, and less than 1 spam/callcenter call per year.

I reported an obvious fraud call to the authorities once per web form, and they followed up about two weeks later thanking me for my submission and that they banned the number block and company behind it.

Strict regulation and enforcement works.

More than a few places won't even recognize VoIP numbers as legitimate primary numbers, and this will only make it worse. I'm not surprised an organization tied to Microsoft is going to ruin another thing I like.

VoIP numbers are not really a thing. I mean there are pools usually assigned from that other companies are aware of, but it's not enforced. You can buy a custom number and use it if you want. For example I had a residential landline number from my home town assigned and routed to a mobile in another country for better billing when my family wants to call.

Yeah but you can easily use an api for HLR lookup and see where it is ported to

What's HLR? Can you explain?

The Home Location Register (HLR) is the telephone-provider-specific database that routes calls coming from other providers; while originally a 2G term (it's HSS in LTE and UDM in 5G), people still just call it a HLR lookup. The actual HLR lookup itself goes to a separate country-wide database that maps phone numbers to the cellular or SIP provider. So, when you place a call, your provider can use that database to then lookup the HLR database and reach out to it to further route your call to that person's specific location.

It's not fully reliable. For example you can have local residential lines that in practice go over VoIP and the provider allows you to use the same account to call out. They'll be classified as just normal landlines. It will vary wildly from place to place, but I think soon the type field in HLR will mostly lose its meaning.

(Just checked my company number, running on a hosted PBX, and it's marked as landline with no extra porting info that would contradict it)

I'm well aware. I've been using VoIP for a long time, but as I said, some services refuse to recognize VoIP numbers as they can query and check. It doesn't matter if the number was ported and originated from a mobile service or landline telephony provider.

Netflix is an example of a company that takes a hardline stance not allowing a VoIP number on an account, even while banks don't seem to care since it's the customer's choice.

The comparison with / deferral to banks is not really relevant. Apples and oranges.

I'd be curious to hear your justification for why it's so much more important for Netflix to ensure there's a "real" phone number behind the account compared to banks.

(Note: I don't agree with any equivalency or even correlation between phone numbers and legal persons)

Netflix could have their access to video content cut off if they don't comply with all the ridiculous power trips of content "rights" owners. Banks only have to comply with actual laws.

Banks are required by law to perform KYC. They would otherwise face much greater consequences than Netflix ever would.

Phone numbers are irrelevant for that though, so the only reason banks would think they need it is a false correlation between phone numbers and identities.

Netflix also seems to think phone numbers have something to do with identities. But you'd expect banks to be more serious about verification.

Banks in the developed parts of the world only really need to know the national identifier of their customer to report their assets for tax purposes, and be capable of verifying passports/government id tied to the identifier when they show up physically to claim what's theirs. Asking about anything else is only needed because the Usians bullied everyone into wasting humanities limited resources on it because they're administratively stuck several centuries behind when it comes to governance.

See, I don't believe that IDs are only used for verifying that someone in-person is the owner of an online account, because if that were the case, I wouldn't need an ID for banks that don't have physical locations. Rather it seems ID verification is used to correlate a legal person's accounts with each other, and make it more difficult to perform confidence scams. (Confidence scams typically include things like using a different identity to disconnect your actions, and their consequences, from your own legal identity.)

They are. And the government sets an ID standard, everyone implements it and it's done. There's no ever escalating paranoia.

But the ID is of course tied to a nationally centralized auth/login system for the digital variant of the services, as it has been for decades. The notion of being capable of using someone else's ID in the first place is just absurd in this day and age.

I've never had to authenticate to any sort of federal or state system to create or operate a bank account. I just provide ID and SSN and that's it. I have to do it to access my taxes, but that's because the tax agency itself is already on the federal or state level.

Also, glad to hear that identity theft has stopped existing, it's about time. Hmm, why are identity fraud alerts still a thing? I'm sure they just haven't gotten around to removing it yet.

> It doesn't matter if the number was ported and originated from a mobile service or landline telephony provider.

It still depends on public databases with hopefully-accurate data. I'm registered on Netflix on a number which can make and receive VoIP calls. It's just not a usual VoIP-specific provider. https://news.ycombinator.com/item?id=50401117

> For example I had a residential landline number from my home town assigned and routed to a mobile in another country for better billing when my family wants to call.

Why not just use some calling system that isn't legacy phone based? My preference is Signal because it's easy enough to use (including for non technical people), works well enough, and is private.

I think you guys havent seen how large corps use/run call centers/BPOs. Indian (legal) call centers rake in billions of $$$ a year. This kind of feature is to try to get a leg into that market. 100% sure enterprise stacks asked for it.

Is this why I've gotten 7+ spam robocalls in the past 2 days?

That's kind of cool. I hope more companies do this. If I understand it correctly, I can provide my own SIP server, so ChatGPT can join conferences or answer incoming calls.

totally that's the kinda easy, but overall OpenAI is building ecosystem owning as much as much part as possible. i think telephony soon could be token bill possibly

This is how we get AI regulations.

Elections are how we get that, but imagine the nonsense we’re all going to have to endure between now and January.

It's time to just stop using the telephone, it's only for scams and mother's day

My phone use dwindled to almost nothing after my father died. I unplugged my landline then because he was the only real person who dialled it.

I very rarely get calls on my mobile; only about one in twenty aren’t scams.

Nah, this is just how people will get screwed. Proper regulations will only be introduced if those in power care.

Oh, Jesus Christ.

Cancer. You won't talk to a human being anymore, and everyone will block numbers not in contacts.

The FCC and Verizon / ATT / T-Mobile allowed the phone system to be ruined in this country by continuing to route calls for spammers and scammers for decades. It's handfuls of people and organized crime causing the vast majority of the problems. Instead of actually investigating this stuff and shutting it down systemically and throwing these people in prison, they let it get so bad that we're just deciding as a society that we'd rather not have real-time voice communications because the few companies that run them haven't been good stewards of the technology.

isn't there some application for hardware attestation for calling? Like if I get an iMessage text I know its a real person with an iphone, but no similar system has been applied for phone calls. It can't be that hard for carriers to provide some kind of indication whether a number is an actual phone or not.

There are massive armies of cheap Chinese phone bot farms... I don't think getting spam from a physical phone is any indication there is a real person in the mix.

> and everyone will block numbers not in contacts.

I've had this configured for a while now due to my number being leaked all over the place and getting 4-8 spam calls a day :/

> everyone will block numbers not in contacts.

Already been doing that for a solid 20+ years now, ever since PocketPC/Windows Mobile first introduced the feature. But OpenAI getting SIP support is not some great new existential threat... we've had similar APIs for years with services like Twilio.

Hopefully it’ll be possible to use this to dispatch faxes over SIP as well

Is that why I’m suddenly getting fax spam this year?

Every few weeks, it’s either a fake invoice or investment pitch.

Since I receive them as a PDF in an email, I wonder if prompt injection via fax has been done.

If you’re stateside, anti fax spam laws have more teeth than anti-email spam laws.

That’s honestly more frustrating, given call and SMS spam is more prevalent than fax or email spam for me. Those systems just have better (but not infallible) consumer-side protections.

Get with the times, Congress.

This might be the worst idea in history.

You know we've had SIP via APIs for close to 20 years already, right?

Anyone else eagerly awaiting a deluge of ChatGPT scam calls?

what could possibly go wrong?

This honestly feels like the use case...

Just tell it you're hard of hearing and need it to say "Albuquerque, New Mexico" after every word containing letter "a".

Cringing at myself for referencing Rick and Morty, but it’s too apt.

https://youtu.be/qVpqaac4mZ0

the number of spam calls I get on a well published number has gone way up since 2024. Like dozens per day. Twilio filters don't work, support doesn't "get it". I recently ported in an old-persons number and holy CRAP! Calls starting at 7am their time and lasting through dinner!

Call your representative and bother them about punishing the callers and telcos. It can be regulated. Just needs people to care.

Better yet, prompt ChatGPT to do it.

This. Call their offices a hundred times per day to reinforce the point. Leave voicemails each time. Use different personalities and names and numbers on each. They might not get each literal message, but they're get the figurative message.

I predict by 2030 humans are back to paper and personal messengers just so we can get some peace from slam.

I look forward to iMessage latest revolution from Apple. It takes a revolutionary to bring human connection back in the feel of your hands. Send incredibly beautiful textured linen paper, folded to perfect, tucked in a pocket of starched paper, sealed with the latest in molecular bond technology. Upgrade with iCut Pro, a shimmering blade in silver, gold, rose gold, and matte black that pairs incredibly with iMessage to open each embodiment of human connection with a swift, crisp cut.

Remember they literally had, and cancelled, Steve Jobs's real-postcard-from your-photos passion project.

is THAT what MobileMe was supposed to do???? Holy...

You'll only be able to talk to someone if your Apple watch reports that your heart rate is acceptable for the call, and they won't be able to answer the call if their Apple watch says they're too agitated.

Lowering the barriers to AI executed fraud, 1 SIP trunk at a time!

This screams for regulation. I don't see how OpenAI can claim to avoid liability using regulations originally meant for telecommunications carriers and now used to cover ISPs and Big Tech while also claiming their services are intelligent.

Don't worry, they'll self-regulate. I'm sure they'll successfully ban anyone trying to use this for legit business purposes or for handling personal matters while allowing spammers who are even remotely clever to operate thousands of simultaneous accounts and place billions of calls without getting caught.

Then a month later they'll reinstate your account with zero explanation. Any attempt to contact customer service will be met with AI responses saying they got your message but are upholding your ban.

How is this different from Twilio's Voice API that's been around since 2008?

They 100% added this because they want to be regulated, sued, blamed.

They really went to a lot of trouble to do this when none of us answer the phone to unknown numbers anymore.

They can combine it with the info they have on ChatGPT users and sell readymade call scripts for ..."marketing" purposes to the highest bidder.

It could be very valuable product!

This feature was sponsored by Indian fraud call centres

Just like the telcos, if they're willing to pay, they're open for business.

And just like that, a thousand startup valuations went poof.

Great. More spam

So now The Lawnmower Man can be real jk

Reply on news.ycombinator.com