How many years until General Direct Dialed telephony dies?


My premise is that the status quo is untenable, and that defections (not having a directly reachable line) will increase, especially among higher SES groups, killing network value.

All whilst service and maintenance costs are increasing.

Paradoxically, low-cost calls are killing the system due to the disutility of junk calls ( ). In areas, particularly the US, robocalls are a plague.

And I suspect the carriers are encouraging this.

Question is: what follows?

As I'd posted this poll a week or so back, general direct-dialed telephony, a/k/a PSTN (public switched telephone networks) seem to be in trouble. I think we could be within five years of their total collapse. And no, not just land lines (already about 25% of their peak in much of the US), but _all_ direct-dialed phones: mobile and VOIP included.

The problem, as I'd written before, is paradoxically the _low_ cost of calls. This is inducing tremendous volumes of junk calls, with one ...


... robocaller admitting to 57 million phone calls in an FTC report.

Online discussions of robocalls are filled with comments telling of how the writer no longer answers unknown numbers, or checks voice mail, if they have service at all.

And businesses are getting hit with both robocalls, to the point of putting lives at risk, and scams, costing dollars. Thousands per day.


When the telephone was first introduced, and as it expanded through the 1950s and 60s, it was, as Facebook in the oughts, an aspirational instrument. Truman beat Dewey, but pollsters, relying on wealthy- and Republican-skewed polling, got the race wrong.

High-toll long distance made long-range romance expensive (and drove a younger me to the campus Unix server for 'talk' sessions with a certain someone), but also made low-level scams and fraud unprofitable. "Voicemail" used to be ...


... called "answering machines", and when I was young, you didn't have either. Either you reached someone and left a message, or the number rang (never more than six times unless you were an insensitive clod), or returned a busy signal (ask an old). Fancy-pants people had secretaries who took messages on slips and routed them to you at the office. And all of this increased the costs of phone-based annoyances.

Another side of this is that wired landlines are dying, but costs remain.


Serving a subdivision of 1,000 homes with 95% phone service costs about the same as one with 5% service -- the trunk lines and switchboxes still need to be there. But the revenues are 19x smaller.

And it's likely that the remaining customers are _not_ higher than average on the socioeconomic status. Though it seems people _will_ drop landline service rather than mobile, when push comes to shove. Landline subscriptions plummeted during the 2007-8 financial crisis, mobile rose.


Which makes me suspect that telco providers, possibly even mobile, are inviting this. They want out of expensive landline obligations.

This, incidentally, mirrors some earlier patterns in networked industries, especially comms and transport. Transit systems especially were frequently private affairs in the 19th and early 20th century, but became government-owned, typically municipalised, especially in the 2nd half of the 20th century. The pattern I see is that _early_ network...

6/, particularly where highly-sought customers can be served in isolation (see the 1980s computer magazines @natecull posted earlier today, featuring ... zarrow people of colour) means early entrepreneurs can achieve high returns on investment. For both electrical and telephone grids, it wasn't until the US government established rural cooperatives for each that rural services were established in much of the country. So much for the free market.

Over a longer span, postal ...


... services followed a similar route. Courier services were a private offering until an upstart maverick came up with the insane notion of a *government run* messaging service. Must've been some a-rab: bin Yamin something or other. (

And of course, high-value carve outs for overnight messaging became A Thing as jet air transport took off in the 1960s (after the USPS helped subsidise both rail and air transport in the 19th and 20th centuries): DHL, FedEx, UPS.


What we're seeing now is the flipside of this transition. The phone network is becoming anti-aspirational, people don't like making, or receiving, phone calls, in general (though specific exceptions still exist). And the business side is getting increasingly unattractive.

One possibility is that government steps in in some way or another. For numerous reasons, including both old and recent history in the US, this is problematic. Putting the NSA in the driver's seat rankles.


Another possibility is that a new set of bespoke messaging service providers emerges. We may already know several of these, or at least the contenders: Google, Facebook, Twitter, Microsoft, Apple, Samsung.

A set of protocol- and api-based multimedia messaging protocols with centralised directories and abuse controls ... might ... arguably be better than what we have today. Though it's unlikely the contenders will play nice with each other, or spare the plebes.

This happened before.


It's pretty much the history of the early telephone industry in the US, where one scrappy upstart, first known as the Bell Patent Association, or later and more popularly, as the American Telephone and Telegraph Company, struck on the bulletproof business plan of buying up long-distance connections and denying them to any competitors.

While AT&T's 1980s anti-trust conviction is fairly well-known, the fact that the company operated under anti-trust restrictions (and, to be sure, a ...


... de facto government-granted monopoly) since *1913*, originally the Kingsbury Committment ( In the 1950s, additional restrictions, including on the sale of computer systems, meant that when a couple of upstart AT&T programmers wrote a game-playing system the company realised it couldn't sell it, and so gave it away for free. That was called Unix. (

My experience is that similar conditions lead to similar behaviours, and that we'll ...


... likely see a reprise of business practices and behaviours during the monopoly-forming period of the early phone system. That will include gamesmanship over protocols, patents, backplanes, and last-mile connections. It's going to be really ugly, probably for a decade or two.

(Assuming there's anyone to connect and any technical way to do so.)

Another possibility is some sort of distributed mesh federated system. I'd like to see that, I have my doubts it will scale particularly well.


I'm more than happy to be proved wrong. But past systems seem to have first shown issues at about 100k users, and hit substantial problems as _actives_ climbed into the 1m - 10m range.

The fact that Facebook _has_ scaled to _billions_ is actually quite an accomplishment. The problems have finally appeared, but they were kept in check for quite some time.

For open and federated systems, there seem to be a number of failure paths, though I'll save those for another discussion.


What's possible as well is that a new comms system will grow out of some institution that finds it has a pressing communications issue and has the means within itself to come up with a solution. This is in fact how virtually all previous communications technologies have emerged.

Clay tablets and papayrus from grainkeepers and priests (often one and the same), the telegraph was a natural adjunct of railroads, for both right-of-way and signaling and messaging needs.

The Internet ...


... was co-developed by the US Military, who needed a highly-resilient communications network, and US academics, who wanted a cheap one. Features including file transfer, remote access, email, group communications via Usenet, and the Web grew from this and similar (Cern/WWW) requirements.

The careful reader will note that I've not mentioned who lead in early adoption of the telephone, and truth is I don't know -- information isn't clearly available at least on brief search. Though ...


... I suspect that banking, government, military, and news applications were among the early and prominent ones, possibly with the first world war playing a significant role.

And the adoption curve does suggest wartime acceleration (and a Depression depression). The 50% adoption mark wasn't reached until over 50 years after the telephone's invention.

(Source: HBR:

But somewhere in that history should be clues as to what might lead next-gen comms development.


And I've got a hunch that it's going to be some sector, one that's highly information and communications dependent, with a mix of _both_ uniformity _and_ multi-agency, which might lead the path out.

That is: It's got to be cohesive enough to be able to make decisions and have an agreed common set of needs. But it's also got to be diverse enough that independent agents are working collaboratively and cooperatively toward a solution. An entirely in-house system for in-house ...


... consumption won't work.

A consortium of multiple governments -- say, the leading industrial US states (California, New York, Texas, Florida, Illinois, Pennsylvania, modulo politics), or a consortium of national states (say, within the EU) might pull this off. Academia again is a possibility. The financial sector. A set of major shipping or trading companies, possibly.

The problem with the IT industry, as it stands today, particularly the advertising and retail-driven FAANG ...


... companies, is that their incentives are _not_ strongly aligned with their users (though they've got strong incentives to general usability). That should count as a handicap for Google/Alphabet, Facebook, and Amazon. Looking at other major tech companies, IBM, Accenture, Oracle, SAP, and Salesforce are other contenders, though these lack a strong public element. And each tends somewhat too strongly to NIH solutions.

The possibility is that as with Unix and AT&T, the solution will...


... appear out-of-sector, though I don't have any strong leads on where that might be.

Another possibility is that the extant PSTN might get its collective shit together. As has been noted (including in comments by @szbalint to this thread) the robocall and other issues plaguing the US aren't evident elsewhere, notably Europe.

Whilst I _do_ think that regulatory reforms might address the immediate problem (my vote: chargeback to the connecting telco for spam calls, with those then ...


free to pass those costs on to their upstream(s) until the problem is resolved, effectively de-profiting the trade) I suspect the underlying structural problems in PSTN are too severe, even in its mobile and VOIP variants.

And, more importantly, that the fixes won't come in time to prevent major defections from the network, which is what its death entails. As I posted to HN: within the next five years, at least one firm in the Fortune 100 will publicly cease use of PSTN. And for the ...


... rest of us, the idea that you can count on the fact that a business, organisation, or individual can make or receive PSTN calls will no longer be 90%, or even 50%, assured. (I suspect the breakdown point is somewhere close to 71%, at which point the chance that any _two_ entities have phone access reaches 50%.)

And once that point is reached, things will change very, very quickly. The flipside of network growth dynamics, with their positive-feedback cycles, is network _decline_.


@dredmorbius Sounds like the development of social media: here in Switzerland people have mobile phones that aren’t listed in phone directories unlike landlines. Landlines are full of spam calls. Basically mobile phones are now like chat systems: one-on-one whitelists.

@dredmorbius The problem is the alternatives are all balkanized and shitty.

SMS: Insecure.

Skype: Doesn't work half the time.

FaceTime, iMessage: Works if you and the other person have iPhones (common in the US, only rich people elsewhere).

Discord, WhatsApp, Telegram, Signal, Line, etc.: Guess which service any person you want to contact is on. No switching board. Most use phone numbers as their ID.

@mdhughes Pretty much exactly this.

The question is: how do we break out of this local non-optimum?

I see either some community-with-a-comms-need (as with universities and the early Internet) or (maybe) a working federated system emerging (but working, non-fussy, scalable, cheap, not abusable). I don't see either yet.

@dredmorbius The more I think about how to make a switchboard for these, the more it looks like identity services like Google or Facebook, LinkedIn, or even old, and those are horribly abusable, and die faster than the services they connect to.

Half those services allow direct links in, the other half are walled prisons.

The idea of a government-run system working and not being monopolized by one operator (just like phones were by AT&T) is just impossible.

@mdhughes The Directory Problem is a huge one, and it's pretty clear that numerous entrants into online / Web 2.0 space seem to have had this prize in mind. LinkedIn, Plaxo, AoL, Google, and of course, Facebook.

But a single huge directory has problems.

Both postal and telco systems, as well as email, offer potential models for working around this, and the idea of some sort of locality, whether via geography or network topology, would solve a lot of problems.

FB has one locale effectively.

@dredmorbius We are in the transition to phone calls mediated by chat applications.


Junk calls are mostly a FCC-related problem.

They don't really happen in Europe. The FCC messed up in enforcing existing laws and regulations to go after shady companies and telcos for allowing the junk call situation to develop.

@szbalint There appear to be plenty of call centre calls in Switzerland on landlines, as far as I can tell. I hear that that insurances will call, there are telemarketing polls, and some fraudsters, too. I should ask them about incident rates, though. How many calls per week would you deem to be OK?

@szbalint Fair point (and I addressed that late in the tootstorm). I think the structural problems with PSTN are too great, and that the defections will hit too soon for either extant US telcos to recover or continue.

I also suspect that the EU and elsewhere may simply be lagging the US experience by a few years, though we should find out on that point fairly shortly. The same underlying switching and signalling limitations as apply to the US also apply to Europe, and bad actors will bad act.

@dredmorbius Learning about robocalls was such a weird experience. That's not a thing here (possibly in Europe in general?) so their existence seems absurd.

@jonne @dredmorbius

Here in New Zealand we've been getting hit with complete fraud robocalls - mostly scary recorded voices saying THIS IS [$actual Internet company who they aren't] WE ARE CUTTING OFF YOUR INTERNET and if you pick up you get routed to some ridiculous pay-per-minute thing in some foreign country, and of course the number's faked. Either that or HI WE ARE MICROSOFT (they aren't) PLEASE TO INSTALL OUR MALWARE.

It's only become a massive plague in the last.. three years, I think?

or you got 'tu-du-doot', in which case you fat-fingered the wrong number 🤣

@dredmorbius It's the #@*&#@$ topology - any can reach any in one step is a nightmare. An absolute nightmare. That's what's going to take us down. We need something like web of trust to put a distance metric back in; anything that doesn't have one is going to implode.

@feonixrift Or one-on-one whitelisting. Basically the UI that mobile phones already have, but with different defaults: Silence all incoming calls by strangers! @dredmorbius

@kensanata @dredmorbius
This is the ball of radius one and all other radii are infinite topology. It works great, if people are willing and familiar that they need to get introductions from other humans in order to connect. Given the success some networks have had with FOAF type layers, I think this can be leveraged and improved upon.

@feonixrift My experience with friend-of-a-friend like sharing has been dismal. Sadly, enough of my "friends" have terrible taste and I'm better off not knowing. @dredmorbius

@kensanata @feonixrift @dredmorbius Well, I'd distinguish FOAF as a network strategy from FOAF as a communications strategy. FOAF for networking is, I hope, more likely to be a scale-tolerant and resilient system. But I've also had abysmal experiences of FOAF comms.. Retroshare, for example: once you hit FOAF connectivity sufficient for global routing.. there's this one guy who spams all the fora with antisemitist screeds. And anonymous, censorship-resistant FOAF networks can't deal with that. 🤷

@cathal @kensanata @dredmorbius

This is why I think we need a toy network - a sandbox of completely simulated 'nodes' and 'links' and 'posts', purely statistical, not involving real servers just simulations - on which to try these things out! Then we would be able to work from at least simulated data when hashing out which methods might work, rather than pure imagination.

@cathal Reputational accountability for introductions and recommendations _is_ something that I think should scale _somewhat_ better than raw content-recommendations cases. Though given the disasters (and susceptibility to manipulation) of the latter, we may want to dial up the skepticism fully.

@kensanata @feonixrift

@kensanata Some friends are better recommenders than others.

In cases, getting agreement from a set, or having a local authority (in the sense of "has good taste and/or sense") might work.


@feonixrift The whole protocol of introductions and references which society and business used to rely on may have some lessons.

The ancient methods were slow, and probably _too_ slow, for modern needs. But a reboot of the general idea might work, and could be based around existing social, community, business, and/or financial institutions based on trust.

Without becoming oppressively Big Brotherish _or_ petty-tyrannical.


@dredmorbius @feonixrift @kensanata

Unfortunately introduction-based screening is *literally* the definition of "prejudice" and is about the worst possible "solution" to this problem.

it's going to lead to massive, casual, race, class and sex-based discrimination, just because *that's how introductions work*. You don't even get a foot in the door unless you Know The Right People. And of course you don't know the Right People if you're Not Our Kind Of People.

@natecull Depends. If it's based on social introductions yes, and very much so.

If it's based on some sort of blinded introductions, merits-based, some sort of entitlement (in the sense of, "by virtue of some vetted property X, party is entitled to participation", where X could be any of various diverse qualifications), or even fully-blinded / sortition based vetting, which largely goes to say "we know _who_ this is and can vouch for identity (mostly), but ...

@feonixrift @kensanata

@natecull ... the rest of it is up to you to sort out.

Dorm-room assignments are something like this (though freighted with all that's involved in college admissions). You know that your roommate meets entrance requirements, but otherwise, they're an unknown.

That's a pretty significant issue all told, you're right.


@feonixrift @kensanata

@dredmorbius @feonixrift @kensanata

Yeah, hmm. Some kind of verified vetting might work. I wonder. Some kind of token?

The corporate obsession with outsourced call centres I think is what's driving this robocall nastiness. The corporates WANT random offshore businesses to be able to call people on their behalf, but without any verification.

It's funny to me that we hit this problem with email decades before we hit it with voice... I guess it took that long for voice-over-TCP/IP to catch up?

@dredmorbius @feonixrift @kensanata

I know I've taken to screening calls with my answering machine and only picking up if it's someone who's already called me. Even then, Completely Fraudulent Robocalls (tm) really get my blood boiling, because you're never sure that it's NOT a legitimate call from a real company saying they're doing something stupid and horrid.

And then there's just the endless political/commercial pollsters and ads.

@dredmorbius @feonixrift @kensanata

But the biggest trend I see is just so many people whose only phone is a mobile. No landline because why pay for two phone numbers? And even the phone, they mostly use for Internet messaging. Whatsapp and such.

@natecull s/phone/computer/

And by that, I mean both telco and compute are consolidated to a single device.

@feonixrift @kensanata

@dredmorbius @feonixrift @kensanata

I suppose it's *possible* that some kind of absolutely unbiased 'introduction authorities' might emerge from (lol) the 'Free Market' or (more lol) the universities, but, uh.

I don't really think so.

I do think though that either a telephone or an email system that lets you casually fake a 'from' address is *utterly* broken and unfit for the most basic of purposes.

@kensanata Whitelisting, with a possible distribution of (preferably locally/limited) vetted general numbers, would be preferred. Blacklisting does not scale, present systems prove this.

Vetted numbers would include local or regional service and government contacts, though those would then become attack / spoof targets for spammers. A well-known vetted number can reach many others. 1:1 whitelist personal contacts, not so much.


@feonixrift Friction can warm the heart, and distance has its place!

WoT has been in the news of late, and not in a good way.

There's a particularly interesting relationship between trust and comms technologies, and it tends to bode poorly as well. I think comms tech improvements _erode_ trust, on net. Though I may be wrong.

Sign in to participate in the conversation

Generalistic and moderated instance.