Hacker News new | ask | show | jobs
by tsaixingwei 1377 days ago
I tried to operate my own email server as well on a VPS, and I have been thinking that the way to solve these problems is to solve the problem of spam itself. Detecting spam puts the costs on the email providers... when the costs should be born by spammers.

Perhaps some sort of digital stamp (digital signatures similar to stamps on physical envelopes) for each email sent paid for with micropayments in a cryptocurrency like nano (note: I don't own any crypto). Small cost per email like 0.01 cents that is trivial for legitimate senders but not for bulk-sending spammers. SMTP servers should put all incoming unsigned emails into spam folders. This will disincentivize spamming (probably not eliminate it) enough that self-hosting emails might be possible again without having to swim against the tide.

3 comments

Paper mail, phone calls, and text messages all cost money, and yet you get massive amounts of spam from each of those. No, putting a price on those does NOT help. What will happen is that one source of low-quality spam that will be priced out of existence will subside, and now companies with more money will be priced into the game, because they will decide that sending spam is competitive again, as there's so much less of it nowadays. So then those companies are sending you spam and there's no escape from that. What's the next step? How do you fix that? The "digital stamp" thing hasn't been thought out too far into the future.

Those are the equivalent of your banner ads and pre-roll ads. The next step is ad blocking. Since there's not a lot of companies that actually want to pay for this stuff now, there's actually not so many of them, and you can enumerate them. They'll try to randomize the text using auto-gen and eventually things like GPT-3. This might be successful. How do you protect against that?

Let's say you protected against that. The next step is ad integration. The kind of stuff SponsorBlock removes from YouTube videos. Small mentions of ads and sponsors, integrated into the content. Interaction reminders. Donation begging. SponsorBlock works well because all that stuff is public, but that might not work well for email. Are you willing to let an equivalent of SponsorBlock read your email? Would you trust it? It would require a completely new paradigm for such blocking addons, where we're sure - by means of technological assurances - that such a blocking program cannot spy on us by leaking emails back to the mothership. That's a tough one, and I have a feeling in the current browser-centric environment the effort is just so large, and the required approaches are so far from what's being done nowadays, and the payoff of satisfying a fringe mail-midroll-ad-blocking need is so small, that it simply won't be done.

What's next after that? SponsorBlock hasn't caught up widely enough to cause people to find workarounds. I don't know, but I'm sure it'll come.

> Small cost per email like 0.01 cents that is trivial for legitimate senders but not for bulk-sending spammers

Given how much is spent on ads, I’d say that an email is still worth more than that. So, what you get is spammers paying minuscole amount of money to skip the spam filter and killing the idea at the root.

Or they can simply steal these credentials from one of the millions of hacked sites and cause additional trouble to them.

Within the confines of "a digital stamp is a good idea" [1], what you're saying isn't quite true: instead of being treated as a whitelist, a digital stamp can just be seen as an finite increase in credibility of some quantum that's decided over time by adaptive filters.

[1] It's not quite. See sibling post to yours.

Cryptocurrency is a pyramid scheme and even if it weren't, the technology crushes itself under its own weight.