Hacker News new | ask | show | jobs
by api 2368 days ago
You're forgetting spam. Spam destroyed all those first generation federated systems. IRC survived because it was too niche for spammers to target much but spam is the primary thing that killed Usenet and email as a truly open system.

The closed systems were better able to fight spam because they could easily ban people and IPs.

On a deeper level spam, "brattish" commercial sites, etc. all come from when money got involved.

The old Internet was mostly noncommercial. Money changes everything.

Even on the new sites I saw a massive shift when e.g. it became possible to monetize YouTube videos. All the sudden everything became about engagement and controversy and got big and divisive and dumb and flashy.

Ultimately we must adapt or perish. There is no going back. I think all new systems must be designed with the trial by fire of spam and other profit motivated attacks in mind from the start.

7 comments

IRC has mechanisms that make dealing with spam easy.

Usenet on the other hand required cooperation from all providers. Actually I blame Google for killing Usenet. They used Microsoft's EEE (Embrace, Extend, Extinguish). They acquired DejaNews, renamed it to Google Groups, provided a gateway that allowed everyone to use Usenet. This introduced a lot of spam to the network, but whenever someone reported it, they did nothing. Eventually they introduced their internal groups, and shifted search in a way that it got hard to use Google Groups for searching Usenet posts.

They did similar thing with XMPP (Jabber). When they introduced Google Talk, their service was interconnected with the other XMPP servers. Once it got popular they discontinued it and introduced Hangouts (then later iterations) Hangouts was still connected people could see each other being present people on Hangouts could message anyone, but people on other XMPP couldn't message people on Hangouts. It didn't even show an error. This made many users switch to Hangouts to continue taking with their friends.

They attempted to do the same thing with email, but were less successful (since many big companies are also providing the service), this was done through introducing various anti spam measures. You now have to jump through various hoops (SPF, DKIM, RBAC) to have your service still reach Google uses. It didn't matter that I used the same IP and domain for 15 years never had spam sent from it, but suddenly my emails started being silently classified as spam without any warning.

> IRC has mechanisms that make dealing with spam easy.

> Usenet on the other hand required cooperation from all providers. Actually I blame Google for killing Usenet. They used Microsoft's EEE (Embrace, Extend, Extinguish). They acquired DejaNews, renamed it to Google Groups, provided a gateway that allowed everyone to use Usenet. This introduced a lot of spam to the network, but whenever someone reported it, they did nothing. Eventually they introduced their internal groups, and shifted search in a way that it got hard to use Google Groups for searching Usenet posts.

> They did similar thing with XMPP (Jabber). When they introduced Google Talk, their service was interconnected with the other XMPP servers. Once it got popular they discontinued it and introduced Hangouts (then later iterations) Hangouts was still connected people could see each other being present people on Hangouts could message anyone, but people on other XMPP couldn't message people on Hangouts. It didn't even show an error. This made many users switch to Hangouts to continue taking with their friends.

The Jabber coopting by Google always felt like something straight out of the old "Embrace, extend, extinguish" playbook of yore.

Unfortunately, the "old" internet involved a lot of trust. Once sufficient numbers of untrustworthy players enter the system, you have to figure out how to protect yourself from spam and malware, and for most people that means using locked down systems such as iOS and services that verify identity via phone numbers and whatnot.
> The closed systems were better able to fight spam because they could easily ban people and IPs.

Unless the server allowed one to send email or post to usenet without having to log in first, then there's no reason why the provider couldn't simply disable the account or block the originating IP from connecting to the server. From what I can tell, the providers weren't interested in blocking spam by blocking IPs or disabling accounts. This is very similar to the robocall problem and phone companies not really trying to fix it.

> IRC survived because it was too niche for spammers to target much but spam is the primary thing that killed Usenet and email as a truly open system.

A year or so freenode was hit by spam, and now everyone needs to verify, so spam still exists, as does IRC.

Even as late as 2000, usenet survived spam, conversations continued, and spam in email was far worse. Spam in email went the way of the dodo around mid-to-late 00s, with the centralisation of the providers (gmail, yahoo, hotmail)

I run my own email server, and use a separate address per correspondent. This avoids spam entirely. I have sometimes got some spam from some addresses but simply disable that address and then the problem is gone.
That's too much time and work for 99% of users, even tech-savvy ones.

Think of it this way: lets say you value your time at $50/hour (very conservative for a tech-savvy person). If it takes an hour a month to admin that box, that's a $50/month e-mail service you have not including VPS/VM cost.

Yes, although I am not asking everyone to do it. I am only saying I do it. It is the same protocol as everyone else's email, I just set it up my own way. Other people who like to do can try this too, but I am not trying to ask everyone to do who does not want to do.
I’ve found FastMail to be my happy medium: I can give out arbitrary aliases over a couple of domains, and only have to do the config once per domain, but they take care of the actual email server part. I can then later add mailbox rules for aliases I give important senders.

Granted, the domain config would be overwhelming for my friends and family not involved in IT - even my mechanical engineer husband doesn’t quite understand what I’m doing.

Do you accept wildcards, or create a new alias in advance every time? I used to accept wildcards around turn of millenium but spam was overwhelming.

How do you stop "mainstream" providers from sending your emails to spam?

I do not accept (and never have accepted) wildcards. I create a new alias in advance every time.

I use the ISP's server for sending, and use my own server for receiving (the menu for install Exim has a "smart host" option which does this).

> Money changes everything.

And almost always for the worse.

Greed is as normal and natural as sex. If there is no legitimate, productive outlet for it, it finds illegitimate and unproductive outlets. This is IMHO who communist states turn into mafia states: the mafia becomes the outlet. Part of adapting to the change is going to be building legitimate profit avenues into systems so that the profit motive can find productive channels.

All the old systems had no channel at all for profit making, so it made it's own in the form of spam and similar.

> If there is no legitimate, productive outlet for it, it finds illegitimate and unproductive outlets. This is IMHO who communist states turn into mafia states: the mafia becomes the outlet.

The reason that former communist countries like Russia became oligarchies (what I think you might mean by mafia state) is because people with significant power in the previous communist system seized control of huge state assets during the disorder that accompanied the system's collapse, and in doing so took ownership of large pieces of the economy. Even under communism, power was closely held among a in-group, and that stayed the same afterward.

The oligarchy didn't emerge because average people were bored and provided for, didn't know what to do with themselves, and therefore decided to set up illegitimate enterprises.

Also, plenty of people in capitalist countries set up illegitimate enterprises like financial frauds and consumer scams of all sorts (remember 2008?), and capitalist societies have their own oligarchs - though we call them plutocrats instead.

It doesn't feel that natural. I think we should stop resorting to human nature when something fits our mood. Violence is natural but we try to stifle it because it haa few benefits. Greed should be the same. I don't think our current worship of money is healthy for the future - if you believe what they say about climate change atleast.
Yes, I agree. But none of that changes the fact that when something shifts to being for-profit, that thing usually gets worse.
> On a deeper level spam, "brattish" commercial sites, etc. all come from when money got involved.

Recall too that this is about when Scientologist decided to do something about people saying bad things about them on Usenet. Though maybe you can call that money, too.

So are you saying that you believe YouTube will also perish because it's now mostly spam, albeit sanctioned spam? Not a facetious comment, I genuinely wonder. It'd be a wonderful twist, in a way.
YouTube's deplatforming and demonetizing is about improving content quality for a broad audience, so they have been taking steps to fight spammy content. They will probably have to do more.

The systems that died were systems that were structurally unable to fight spam or where doing so was prohibitively expensive in time or money.