I'm sure it's a good article but there is no need for it to be written by an LLM... Is everyone really this lazy now? I don't think I've read anything written by a human for at least a full month.
Everyone should aim to be this lazy. Everyone is trying to maximize their time. This is what software has always done. It helps people be lazy in some respect, and then, so far, we have always found new stuff to do with the time we freed up.
LLMs got decent at writing, and before long they will be better at write-ups than the average human writer and we won't look back.
Of course, this is contingent on your current goal: If you want to get better at writing, you should do the writing, just like you want to do the math, if you want to learn multiplication. Writing is fine as a tool just to get information across. It needn't be more than that for everyone.
The problem is LLM output’s value is intrinsically zero because there is no scarcity to it. I too can just prompt “What’s the rate of self hosted email?”. If an article is signaling that its value is zero, then everyone should just move on as it’s not providing anything to anyone. Effort is the economies main signal of value, good boots that last cost more in the classic example. The only time it’s not zero is when there’s a shared understanding between those consuming the output - a company wanting your llm code, friends making funny images for an inside joke to share, etc. And some people grouped may be specifically interested in AI output - people sharing music on Suno. But in my experience no one cares about your AI generated song, image, article, app if it appears to just be my own prompt away from me having it too. Give a significant other an AI image printed out for an anniversary gift - see how excited they are about it. Versus getting an artist or photographer to do the same. As Claude might say, it’s not laziness, it’s effort.
And soon, I'll be lazy and use an LLM to do the reading for me. Eventually the LLMs will train on the state of the world and decide to be lazy also, and the "writer's" LLM prompt will just be shared with the "reader": nobody actually needed all those words anyway! Maybe someday we'll get rid of the links on HN and we can just share and consume headlines.
The ability for people to communicate is paramount. Without the ability to communicate there is no progress. We might as well be deaf, dumb and perhaps blind.
Delegating this task to automation serves to undermine all the masters and indeed strengthen the servants.
Short term this offload is good for people; long term it’ll lead to mental rot where people are hard pressed to put their ideas forward and give up in exasperation when they don’t have the assistance of their servants. Their ability to even make the proper requests to the servants may deteriorate enough that their existence may come to rely on the benevolence of their servants.
The need is that writing is expensive. An LLM is around a hundred times cheaper than a human, so this kind of article that doesn't provide tons of value just makes sense to write with an LLM.
That's the thing, good writing is expensive, massively so, much more than people think.
Letting the machine do the writing however shows us that many people have no idea why writing is expensive, nor do they respect that perhaps it needs to be expensive. Communication is incredibly hard, very few master it. I'd even go so far as to say that clear, precise and readable communication is so hard that we do not have sufficient training data to build an LLM that can do it.
Letting a person write article, like this one, doesn't necessarily improve the quality, but it does show us that some thought and effort have at least been put into it. There was a cost and you deemed that cost worth it for conveying your message. Without the cost there's nothing stopping you from wasting peoples time.
That might apply to some things, but there is little to no value in this article, so the economics are the same as the spam - cheap to spew garbage while providing no value and wasting the readers time. It's not something that should be encouraged. RIPE should know better.
I've been using GMail roughly since it existed I think, and have thought on and off over the years about self-hosting. The notion scares me every time because with it comes thoughts of what would happen if I were to lose the domain name or server, or even if there was an outage at the "wrong time". And then there's also the idea of updating the 1 million+1 services, etc that I used it to sign up for. As much as I'm not a fan of a decent amount of what they've been doing, they're however a stable entity which offers a very reliable service for essentially free. Very likely that many others also think along those lines.
The only thing I think can compete with that is some kind of decentralized solution, also something I've thought about a bit. Like a Keyoxide-based[0] - or similar - setup that makes it virtually impossible for one to lose their online identity (since lost keys can be rotated out once one has - access to - enough proofs), and then that identity can be used - without fear - to route associated data, etc. Of course adoption of what could become an adjacent or new standard would be the next hurdle to overcome...
I wish there were a way to self-host without my emails constantly getting stuck in other people's spam folders. And it made me really paranoid about email, too -- like I'd send emails, and I'd never know if the other end received them or not, unless they replied. Without a reply, I would always wonder whether the message got a look, or disappeared without a trace.
The only way around this seems to be hosting with Google, Microsoft, or maybe Proton. Any other ideas?
It happens on zoho. I ran my business email on zoho. Multiple instances where people on Gmail had my messages go to spam. I have the same issue with proton mail.
> DMARC: adopted everywhere, enforced nowhere in particular
Unfortunately this only talks about what DMARC policies are published by domains. I was hoping they will go into some analysis of how many mailboxes actually hard-enforce DMARC rules on ingress (as in, respect the DMARC policy as published by the sender's domain, p=reject or whatever else it might be)
I have been hesitant to enforce DMARC rules on my server. Partly because maintaining it was a pain (see opendmarc - maybe better with rspamd now), but also because on my test runs I saw plenty of legitimate incoming mail that simply would get rejected if I were to enforce DMARC on ingress.
I suspect big players like gmail have some kind of soft enforcement/reputation/machine learning approach to DMARC, but this is way too much work for a small time admin like me.
I've been running my own email server with exim and dovecot for 20ish years, I like the freedom it gives me but I have to say it's a significant investment. Not so much keeping things up to date but instead making sure that my emails are actually delivered. I've eventually switched to using fastmail just for sending emails and even with that, I don't get perfect delivery.
Last year I started self hosting my email. I found that once you setup the “usual suspects” (DMARC, DKIM, SPF…) outgoing mail works without issues. In fact it works now better than the (small) provider I was using before.
And setting them up properly is not even that difficult nowadays
If someone can solve the spam problem in email without just manually whitelisting specific providers that heavily monitor their users, then they can completely reverse the direction of this problem. Unfortunately it doesn't seem likely someone could do that.
That would work if most spam wasn't from the cartel. I receive single digit percentage of "spam" from randoms, 95%+ is from Gmail/Google, o365 or mailgun/whatever. Not one of them can either filter their users or actually action abuse reports.
I enforce strict spec and behaviour compliance, as well as policies (SPF, dkim, etc). Aside from a very small number of well known spam/abuse sources, it's largely from legitimate mail accounts on Gmail/outlook etc. or idiot msps.
I wish that was true for all cases. My uni email is 99.9% spam and everything comes from official accounts sending irrelevant stuff to everyone and their grandma
I agree Spam is the hardest. Setting up a mailserver is pretty easy today, e.g. with NixOS-simple-mailserver you can do it with a few lines and a few minutes of config in your DNS provider, and maintenance is minimal.
But I self-host email for ~20 years, and I have not managed to get rid of Spam without not also getting false positives.
I first used SpamAssistant, in later years Rspamd, but I feel like they are just not good enough. Also I find Rspamd config pretty incomprehensible.
Hosted email like also does not solve this, e.g. GMail filters way much (e.g. important company correspondence leading to orders almost being lost because they landed in our GMail spam, so I had to turn off the spam filter entirely).
It's not in the current providers' interest to help people avoid the current providers' advertisement or avoid paying the providers to host the customers' email.
You mean filtering? That could get extremely expensive computationally, though. Tools like Rspamd support LLM-based filtering, but traditional Bayesian filtering is vastly cheaper and more computationally efficient. Hm, maybe LLMs can be used as an additional signal for ambiguous messages rather than as the primary filter.
If you are thinking using AI to classify messages, the scale of email spam is so massive that using conventional techniques won't work. It won't make sense energetically speaking IMO.
I think techniques like HashCash by Adam Back should be revisited.
Mail self hosting is not an easy process to maintain.. If you do manage to get whitelisted you need to ensure you're on top of maintaining it, let alone some larger provider like Google just deciding they don't like you. I do empathise with wanting to self host, I just always found the cost of effort to do so really didn't offer enough benefits versus a good provider like proton for example. It sucks but maybe a fight that's too far gone to win.
It’s not that bad, but it does require some effort. I self-host. I spend far more time using and enjoying email than I do maintaining the infrastructure.
However you do have to live with the occasional blacklist when you host on a VPS like I do. Some idiot in your IP range may get blacklisted for spam and then you get caught in it. It’s generally temporary.
And there are some corporate policies that will flat out reject emails from you which is fun.
I’ve probably spent less than 10 hours over the last 5 or 6 years on maintenance.
20 years of running the classic postfix + dovecot stack for about 10 business domains. I’ve recently been testing out stalwart, with just one company domain, which is nice but seems quite fragile, 4 times now it’s killed itself doing some self updates. If it wasn’t for Claude assistance I probably would have given up on it by now. It is quite feature rich though, but its spam detection is not great, too much ham flagged as spam, not enough spam flagged as spam. 2FA is a nice bonus but for some unknown reason it breaks itself every few weeks and I gotta turn it back off again. If I could get stalwart stable I would migrate all my domains to it.
The hardest thing in setting up your emails server is to keep your messages out of spam folders. Even if you are completely legit company that sends out ONLY human-written personal emails, you will get tagged as spam first.
I've self-hosted for about 2 months, and promptly switched to iCloud. Just because I've been spending too much time calling people and telling them to check their spam folders.
Recently our business moved to another domain. We had a guy who has spent over a month "warming up" traffic, so our new domain won't be tagged as spam by default.
And on the other side, projects like stalwart email ( http://stalw.art/ ) and mox ( http://xmox.nl/ ) makes it so much easier to build an e-mailsystem on your server.
The problem lies with the big e-mail providers marking your e-mail as spam, even if you have everything configured right (MX, SPF, DKIM, DMARC, etc.) and you're not on any blacklists. They use their mysterious opaque algo to determine that your e-mail is spam and you have no idea what you could do to not get in peoples junk boxes.
> They use their mysterious opaque algo to determine that your e-mail is spam and you have no idea what you could do to not get in peoples junk boxes.
And that’s not hard?
Running a modern mail server, a spam filter, a self-hosted webmail, and learning just enough about DNS and networking to get mail through: this is what’s not that hard.
Figuring out how to stay whitelisted, or accept ending up in people’s filters, asking them to whitelist you while your public IP gradually gains trust, is a job.
That's pretty much what I mean. Implementing the authenticity measures is very easy, making sure the e-mails get delivered to peoples inboxes without getting marked as spam is hard.
I just wish these big parties would give you a more formal way to make clear you're not a spammer.
Nothing will ever ruin email. A private email server, or a federation of mutually-whitelisting ones, is the ultimate social media, and not at all a bad personal information management system too.
My self-hosted email filters too little spam, while GMail etc filter too much (e.g. important company correspondence leading to orders almost being lost because they landed in our GMail spam).
I self-hosted for a while in the early 2000s. It was a hassle. My box got hacked -- I'm sure I missed a security update or something -- and while I didn't lose any data because of my obsessive backup paranoia, it was enough of a Thing that I just punted.
I used Gmail for Domains for a hot minute, but never really liked it very much. Soon I found Fastmail, and I've been using them ever since (so about 13 years now).
I switched to M365 as my provider a few years back as my emails kept being flagged as spam.
I've done it for decades, but I don't know much beyond the basics of email hosting tbh, however my understanding is that these days email coming from self-hosted servers carries a much higher spam/junk risk and so often automatically marked as spam.
Maybe there are ways to avoid this happening, but when a few clients all reported that they were not receiving my emails I didn't have much of a choice but to switch.
Hosting my own server with mailinabox. low traffic, but keeping their random subscriptions, newsletters and occasional single purpose address. Haven't had any issues so far.
I run my own mail server, and I am very concerned about its reputation and deliverability. I've run my own Linux server full time since at least 2017, obviously well before LLM's, but only recently put up a mail server. I don't usually run any agents on the server, such as Claude Code, ChatGPT Codex, or Grok Build, all of which I have subscriptions for but don't run on this server. Instead, I manually administer the server using ssh, while getting interactive help from ChatGPT in a chat window.
It was complicated to get security, DNSSEC, DKIM, DMARC right this way. I did it just under a year ago, around September 15, 2025.
I tried to follow and understand what I was doing, and keep good records on it. It was a complicated process, but everything was up and working at the end and I received email my server sent me. It felt good to be able use this simple script to send myself email:
#!/usr/bin/env python3
import smtplib
from email.message import EmailMessage
msg = EmailMessage()
msg["From"] = "(user's from address)"
msg["To"] = "(my gmail address)"
msg["Subject"] = "(a test subject)"
msg.set_content("(a test body)")
with smtplib.SMTP("127.0.0.1", 25, timeout=30) as smtp:
smtp.send_message(msg)
print("Message submitted successfully to local Postfix.")
and have it show up instantly in my gmail within seconds. This is powerful. (I've removed the details from this comment.)
I understand that as my server's reputation grows, it becomes very powerful to be able to email anyone without any intermediary except potential spam filters. I am very careful about what I put on my server.
Nevertheless, I had a recent experience that was a little sad for me. I wanted to forward some mail to my gmail address automatically, and used ChatGPT Sol 5.6 on medium thinking time to help me with the administration to do this. (I find extra high and pro thinking times are excessively long.)
I was very disappointed with the results, which included a very badly written regex for authentication. I'll write up my experience in an experience report. It was a big disappointment for me.
The burden on administrating one's own email is already very large, and it is disappointing that we can't really rely on frontier models to help with it.
This very much depends you been U.S. citizen or not been U.S. citizen.
"PRISM is a code name for a program under which the United States National Security Agency (NSA) collects internet communications from various U.S. internet companies. The program is also known by the SIGAD US-984XN. PRISM collects stored internet communications based on demands made to internet companies such as Google and Apple under Section 702 of the FISA Amendments Act of 2008 to turn over any data that match court-approved search terms."
"Section 702 permits the Attorney General and the Director of National Intelligence to jointly authorize targeting of non-US persons reasonably believed to be located outside the United States.
By targeting, it is meant that US persons or persons located in the United States may not be the intended targets of the collection. The targeting must have as its object the acquisition of foreign intelligence information, as this is defined by FISA."
I remember when Gmail came around and everyone was more than happy to hand off their email to Google. No spam and ability to send and receive huge (for the time) attachments I think were the main things people wanted.
Not sure if that was before or after everyone was happy to hand off their web browser to Google.
The alternatives to Gmail, at the time, were either Hotmail, or email provided by your ISP who charged money for additional email accounts. Those were the options.
You like using Google and Microsoft products as an end user (even ignoring all the ethical issues with these companies)? That seems surprising to me. I find their products so horrible compared to e.g. Fastmail.
We have to figure out a way to get out of FB social, iOS/Android phone, Google/Microsoft email, AWS/GOOG/MSFT cloud oligopolies. There is no competition and there will be massive rentier capitalism extraction from all these players for the rest of civilization.
TBH, I blame IT leadership. The easy choice is Google or Microsoft. "No one ever got fired for buying IBM" is now "No one ever got fired for subbing to Office." Someone needs to bite the bullet and stitch together an alternative array of services. They've cultivated their own monoculture by hitting the easy-button and forking over for Office365, etc.
Which if we're being honest, Microsoft and Google products are rarely best-in-class at any one thing that they do. Teams, SharePoint, Google Docs, someone somewhere is doing those things much better. Those competitors just don't necessarily have a whole ecosystem that makes it a big Easy Button for corporate IT to smash. And they sometimes charge as much by themselves as Office365 costs in whole. But I know one problem my own employers IT team struggles with a bit is putting out the constant fire of different teams deciding the Microsoft tool we pay for sucks and going rogue and trying to subscribe to something else. Dropbox was a big one, but Teams and Planner alternatives come up a lot too.
Not pretending to have all the answers but it doesn't seem to be a completely intractable problem.
It starts by just avoiding them as much as possible. E.g. self-host email. If you encounter a system that refuses to receive your email, I give you permission to use your old Gmail account with that system, while using your self hosted account for everything you can.
If your email is Google or Microsoft, there are other people with self-hosted email who feel pressured to use their Google or Microsoft accounts to send email to you.
I'm sure it's a good article but there is no need for it to be written by an LLM... Is everyone really this lazy now? I don't think I've read anything written by a human for at least a full month.
This seems like a load bearing issue. Would you like me to put together a plan to put human-written text in your daily routine?
Oh you.
> Is everyone really this lazy now?
Everyone should aim to be this lazy. Everyone is trying to maximize their time. This is what software has always done. It helps people be lazy in some respect, and then, so far, we have always found new stuff to do with the time we freed up.
LLMs got decent at writing, and before long they will be better at write-ups than the average human writer and we won't look back.
Of course, this is contingent on your current goal: If you want to get better at writing, you should do the writing, just like you want to do the math, if you want to learn multiplication. Writing is fine as a tool just to get information across. It needn't be more than that for everyone.
The problem is LLM output’s value is intrinsically zero because there is no scarcity to it. I too can just prompt “What’s the rate of self hosted email?”. If an article is signaling that its value is zero, then everyone should just move on as it’s not providing anything to anyone. Effort is the economies main signal of value, good boots that last cost more in the classic example. The only time it’s not zero is when there’s a shared understanding between those consuming the output - a company wanting your llm code, friends making funny images for an inside joke to share, etc. And some people grouped may be specifically interested in AI output - people sharing music on Suno. But in my experience no one cares about your AI generated song, image, article, app if it appears to just be my own prompt away from me having it too. Give a significant other an AI image printed out for an anniversary gift - see how excited they are about it. Versus getting an artist or photographer to do the same. As Claude might say, it’s not laziness, it’s effort.
And soon, I'll be lazy and use an LLM to do the reading for me. Eventually the LLMs will train on the state of the world and decide to be lazy also, and the "writer's" LLM prompt will just be shared with the "reader": nobody actually needed all those words anyway! Maybe someday we'll get rid of the links on HN and we can just share and consume headlines.
>LLMs got decent at writing,
Judging by people's response, that doesn't seem to be the case.
> and before long they will be better at write-ups than the average human writer and we won't look back.
I'll read articles written by them if and when that happens, but not before.
Efficiency and lazyness are two very distinct things. As is quality and quantity.
The ability for people to communicate is paramount. Without the ability to communicate there is no progress. We might as well be deaf, dumb and perhaps blind.
Delegating this task to automation serves to undermine all the masters and indeed strengthen the servants.
Short term this offload is good for people; long term it’ll lead to mental rot where people are hard pressed to put their ideas forward and give up in exasperation when they don’t have the assistance of their servants. Their ability to even make the proper requests to the servants may deteriorate enough that their existence may come to rely on the benevolence of their servants.
Are we even sure the actual stuff in the article is real? What if it's just the usual LLM hallucinations
> The domains didn't disappear; they moved.
I stopped reading at this point. I cannot emphasize how much I HATE this AI writing style.
The need is that writing is expensive. An LLM is around a hundred times cheaper than a human, so this kind of article that doesn't provide tons of value just makes sense to write with an LLM.
I hate LLM writing, but the economics checks out.
> The need is that writing is expensive.
That's the thing, good writing is expensive, massively so, much more than people think.
Letting the machine do the writing however shows us that many people have no idea why writing is expensive, nor do they respect that perhaps it needs to be expensive. Communication is incredibly hard, very few master it. I'd even go so far as to say that clear, precise and readable communication is so hard that we do not have sufficient training data to build an LLM that can do it.
Letting a person write article, like this one, doesn't necessarily improve the quality, but it does show us that some thought and effort have at least been put into it. There was a cost and you deemed that cost worth it for conveying your message. Without the cost there's nothing stopping you from wasting peoples time.
That might apply to some things, but there is little to no value in this article, so the economics are the same as the spam - cheap to spew garbage while providing no value and wasting the readers time. It's not something that should be encouraged. RIPE should know better.
How about not writing the article at all then?
Because people will read it if you write it, hence us being here.
Meh. I read the comments on HN, rarely the articles /(semi-)s
What would we comment on if there were no article?
I've been using GMail roughly since it existed I think, and have thought on and off over the years about self-hosting. The notion scares me every time because with it comes thoughts of what would happen if I were to lose the domain name or server, or even if there was an outage at the "wrong time". And then there's also the idea of updating the 1 million+1 services, etc that I used it to sign up for. As much as I'm not a fan of a decent amount of what they've been doing, they're however a stable entity which offers a very reliable service for essentially free. Very likely that many others also think along those lines.
The only thing I think can compete with that is some kind of decentralized solution, also something I've thought about a bit. Like a Keyoxide-based[0] - or similar - setup that makes it virtually impossible for one to lose their online identity (since lost keys can be rotated out once one has - access to - enough proofs), and then that identity can be used - without fear - to route associated data, etc. Of course adoption of what could become an adjacent or new standard would be the next hurdle to overcome...
[0] https://keyoxide.org/
> what would happen if I were to lose the domain name or server
On the other side of the coin, what would happen if google bans you? That's certainly not unheard of.
I wish there were a way to self-host without my emails constantly getting stuck in other people's spam folders. And it made me really paranoid about email, too -- like I'd send emails, and I'd never know if the other end received them or not, unless they replied. Without a reply, I would always wonder whether the message got a look, or disappeared without a trace.
The only way around this seems to be hosting with Google, Microsoft, or maybe Proton. Any other ideas?
I use fastmail which is very good. Like “oh this is how it should feel!” good.
I only have one minor complaint about Fastmail after 5 years. I can’t do a full-page screenshot of a long email on my iPhone.
I started self-hosting my email last year, but I use MailPace (other email sending services are available) to avoid being memory-holed.
I don’t self host, I pay zoho to do it. Over never had any suspicion my emails haven’t been received.
It happens on zoho. I ran my business email on zoho. Multiple instances where people on Gmail had my messages go to spam. I have the same issue with proton mail.
I think that is the Crux of the issue.
Email spam has made self hosting untenable.
> DMARC: adopted everywhere, enforced nowhere in particular
Unfortunately this only talks about what DMARC policies are published by domains. I was hoping they will go into some analysis of how many mailboxes actually hard-enforce DMARC rules on ingress (as in, respect the DMARC policy as published by the sender's domain, p=reject or whatever else it might be)
I have been hesitant to enforce DMARC rules on my server. Partly because maintaining it was a pain (see opendmarc - maybe better with rspamd now), but also because on my test runs I saw plenty of legitimate incoming mail that simply would get rejected if I were to enforce DMARC on ingress.
I suspect big players like gmail have some kind of soft enforcement/reputation/machine learning approach to DMARC, but this is way too much work for a small time admin like me.
I've been running my own email server with exim and dovecot for 20ish years, I like the freedom it gives me but I have to say it's a significant investment. Not so much keeping things up to date but instead making sure that my emails are actually delivered. I've eventually switched to using fastmail just for sending emails and even with that, I don't get perfect delivery.
Last year I started self hosting my email. I found that once you setup the “usual suspects” (DMARC, DKIM, SPF…) outgoing mail works without issues. In fact it works now better than the (small) provider I was using before.
And setting them up properly is not even that difficult nowadays
If someone can solve the spam problem in email without just manually whitelisting specific providers that heavily monitor their users, then they can completely reverse the direction of this problem. Unfortunately it doesn't seem likely someone could do that.
That would work if most spam wasn't from the cartel. I receive single digit percentage of "spam" from randoms, 95%+ is from Gmail/Google, o365 or mailgun/whatever. Not one of them can either filter their users or actually action abuse reports.
The reason for you only seeing spam from the larger providers is that they are filtering out everyone else.
They should also deal with the individual accounts etc, but thats the difficult problem.
I enforce strict spec and behaviour compliance, as well as policies (SPF, dkim, etc). Aside from a very small number of well known spam/abuse sources, it's largely from legitimate mail accounts on Gmail/outlook etc. or idiot msps.
Spam probably isn't as bad as you think it is. It's not the dark old days where 99% of emails are spam.
You should self-host, so you can do the experiment for yourself. You might be surprised. You can always stop self-hosting if you don't like it.
I wish that was true for all cases. My uni email is 99.9% spam and everything comes from official accounts sending irrelevant stuff to everyone and their grandma
Recent estimates show it's down closer to 50% of all emails now. Which is obviously still outrageous.
I agree Spam is the hardest. Setting up a mailserver is pretty easy today, e.g. with NixOS-simple-mailserver you can do it with a few lines and a few minutes of config in your DNS provider, and maintenance is minimal.
But I self-host email for ~20 years, and I have not managed to get rid of Spam without not also getting false positives.
I first used SpamAssistant, in later years Rspamd, but I feel like they are just not good enough. Also I find Rspamd config pretty incomprehensible.
Hosted email like also does not solve this, e.g. GMail filters way much (e.g. important company correspondence leading to orders almost being lost because they landed in our GMail spam, so I had to turn off the spam filter entirely).
I self host email and surprisingly do not receive any Spam (apart from the occasional cold calling support@ contact@ etc)
It's not about receiving spam, it's about legitimate messages you send getting filtered as spam.
If you set up SPF/DKIM, this rarely happens. I self host and my email works just fine. I also rarely see spam.
It's not in the current providers' interest to help people avoid the current providers' advertisement or avoid paying the providers to host the customers' email.
Wasn't bitcoin supposedly invented to do that?
Ironic given the amount of crypto spam that exists.
Seems like AI is in a pretty good position to do that tbh.
You mean filtering? That could get extremely expensive computationally, though. Tools like Rspamd support LLM-based filtering, but traditional Bayesian filtering is vastly cheaper and more computationally efficient. Hm, maybe LLMs can be used as an additional signal for ambiguous messages rather than as the primary filter.
If you are thinking using AI to classify messages, the scale of email spam is so massive that using conventional techniques won't work. It won't make sense energetically speaking IMO.
I think techniques like HashCash by Adam Back should be revisited.
its not really difficult to filter spam off. I spend more time dealing with physical spam in my snailmailbox than my email spamfilter
Mail self hosting is not an easy process to maintain.. If you do manage to get whitelisted you need to ensure you're on top of maintaining it, let alone some larger provider like Google just deciding they don't like you. I do empathise with wanting to self host, I just always found the cost of effort to do so really didn't offer enough benefits versus a good provider like proton for example. It sucks but maybe a fight that's too far gone to win.
It’s not that bad, but it does require some effort. I self-host. I spend far more time using and enjoying email than I do maintaining the infrastructure.
However you do have to live with the occasional blacklist when you host on a VPS like I do. Some idiot in your IP range may get blacklisted for spam and then you get caught in it. It’s generally temporary.
And there are some corporate policies that will flat out reject emails from you which is fun.
I’ve probably spent less than 10 hours over the last 5 or 6 years on maintenance.
20 years of running the classic postfix + dovecot stack for about 10 business domains. I’ve recently been testing out stalwart, with just one company domain, which is nice but seems quite fragile, 4 times now it’s killed itself doing some self updates. If it wasn’t for Claude assistance I probably would have given up on it by now. It is quite feature rich though, but its spam detection is not great, too much ham flagged as spam, not enough spam flagged as spam. 2FA is a nice bonus but for some unknown reason it breaks itself every few weeks and I gotta turn it back off again. If I could get stalwart stable I would migrate all my domains to it.
Why switch then? Need more excitement in your life?
I have been running Postfix for 22 years, from a big corporate installation to my own on a Raspberry Pi Zero. It just does its job.
The hardest thing in setting up your emails server is to keep your messages out of spam folders. Even if you are completely legit company that sends out ONLY human-written personal emails, you will get tagged as spam first.
I've self-hosted for about 2 months, and promptly switched to iCloud. Just because I've been spending too much time calling people and telling them to check their spam folders.
Recently our business moved to another domain. We had a guy who has spent over a month "warming up" traffic, so our new domain won't be tagged as spam by default.
And on the other side, projects like stalwart email ( http://stalw.art/ ) and mox ( http://xmox.nl/ ) makes it so much easier to build an e-mailsystem on your server.
I have deployed stalwart as my replacement for mailu and iv never had such an easy time setting up mail, i love it.
At this point, self-hosting email is as much a practical convenience as it is a political statement. Or maybe I'm just stubborn.
Self hosting is not that hard actually.
The problem lies with the big e-mail providers marking your e-mail as spam, even if you have everything configured right (MX, SPF, DKIM, DMARC, etc.) and you're not on any blacklists. They use their mysterious opaque algo to determine that your e-mail is spam and you have no idea what you could do to not get in peoples junk boxes.
> They use their mysterious opaque algo to determine that your e-mail is spam and you have no idea what you could do to not get in peoples junk boxes.
And that’s not hard?
Running a modern mail server, a spam filter, a self-hosted webmail, and learning just enough about DNS and networking to get mail through: this is what’s not that hard.
Figuring out how to stay whitelisted, or accept ending up in people’s filters, asking them to whitelist you while your public IP gradually gains trust, is a job.
That's pretty much what I mean. Implementing the authenticity measures is very easy, making sure the e-mails get delivered to peoples inboxes without getting marked as spam is hard.
I just wish these big parties would give you a more formal way to make clear you're not a spammer.
Spam ruined email. It's really as simple as that.
I hardly ever see any spam nowadays, happy fastmail user.
Nothing will ever ruin email. A private email server, or a federation of mutually-whitelisting ones, is the ultimate social media, and not at all a bad personal information management system too.
Edit: typo
spam filters are effective at dealing with that
Not in my experience.
My self-hosted email filters too little spam, while GMail etc filter too much (e.g. important company correspondence leading to orders almost being lost because they landed in our GMail spam).
It works both ways right? If you self-host you are seen as the spam.
I self-hosted for a while in the early 2000s. It was a hassle. My box got hacked -- I'm sure I missed a security update or something -- and while I didn't lose any data because of my obsessive backup paranoia, it was enough of a Thing that I just punted.
I used Gmail for Domains for a hot minute, but never really liked it very much. Soon I found Fastmail, and I've been using them ever since (so about 13 years now).
Interesting topic. I wonder if the AI was given notes to base the article on, or if it was given the data and also made the full analysis itself.
I switched to M365 as my provider a few years back as my emails kept being flagged as spam.
I've done it for decades, but I don't know much beyond the basics of email hosting tbh, however my understanding is that these days email coming from self-hosted servers carries a much higher spam/junk risk and so often automatically marked as spam.
Maybe there are ways to avoid this happening, but when a few clients all reported that they were not receiving my emails I didn't have much of a choice but to switch.
I would also be interested in self-hosting but have limited understanding of what would be required to ensure my email does not get flagged as spam.
Stopped reading at "The MX record says where the mailbox lives."
Interestingly, DNS does have such a record (MB,) just never widely used - https://datatracker.ietf.org/doc/html/rfc1035/#section-3.3.3
Hosting my own server with mailinabox. low traffic, but keeping their random subscriptions, newsletters and occasional single purpose address. Haven't had any issues so far.
I run my own mail server, and I am very concerned about its reputation and deliverability. I've run my own Linux server full time since at least 2017, obviously well before LLM's, but only recently put up a mail server. I don't usually run any agents on the server, such as Claude Code, ChatGPT Codex, or Grok Build, all of which I have subscriptions for but don't run on this server. Instead, I manually administer the server using ssh, while getting interactive help from ChatGPT in a chat window.
It was complicated to get security, DNSSEC, DKIM, DMARC right this way. I did it just under a year ago, around September 15, 2025.
I tried to follow and understand what I was doing, and keep good records on it. It was a complicated process, but everything was up and working at the end and I received email my server sent me. It felt good to be able use this simple script to send myself email:
and have it show up instantly in my gmail within seconds. This is powerful. (I've removed the details from this comment.)I understand that as my server's reputation grows, it becomes very powerful to be able to email anyone without any intermediary except potential spam filters. I am very careful about what I put on my server.
Nevertheless, I had a recent experience that was a little sad for me. I wanted to forward some mail to my gmail address automatically, and used ChatGPT Sol 5.6 on medium thinking time to help me with the administration to do this. (I find extra high and pro thinking times are excessively long.)
I was very disappointed with the results, which included a very badly written regex for authentication. I'll write up my experience in an experience report. It was a big disappointment for me.
The burden on administrating one's own email is already very large, and it is disappointing that we can't really rely on frontier models to help with it.
I see no upsides to a Google/Microsoft duopoly in email. Except for those already-far-too-powerful corporations.
EDIT: "I see no upsides to X" means "X is NOT good". (Responses suggest that this was not clear.)
This very much depends you been U.S. citizen or not been U.S. citizen.
"PRISM is a code name for a program under which the United States National Security Agency (NSA) collects internet communications from various U.S. internet companies. The program is also known by the SIGAD US-984XN. PRISM collects stored internet communications based on demands made to internet companies such as Google and Apple under Section 702 of the FISA Amendments Act of 2008 to turn over any data that match court-approved search terms."
https://en.wikipedia.org/wiki/PRISM_program
"Section 702 permits the Attorney General and the Director of National Intelligence to jointly authorize targeting of non-US persons reasonably believed to be located outside the United States.
By targeting, it is meant that US persons or persons located in the United States may not be the intended targets of the collection. The targeting must have as its object the acquisition of foreign intelligence information, as this is defined by FISA."
https://en.wikipedia.org/wiki/FISA_Amendments_Act_of_2008
How times change.
I remember when Gmail came around and everyone was more than happy to hand off their email to Google. No spam and ability to send and receive huge (for the time) attachments I think were the main things people wanted.
Not sure if that was before or after everyone was happy to hand off their web browser to Google.
The alternatives to Gmail, at the time, were either Hotmail, or email provided by your ISP who charged money for additional email accounts. Those were the options.
And Google, at the time, was cool.
You like using Google and Microsoft products as an end user (even ignoring all the ethical issues with these companies)? That seems surprising to me. I find their products so horrible compared to e.g. Fastmail.
We have to figure out a way to get out of FB social, iOS/Android phone, Google/Microsoft email, AWS/GOOG/MSFT cloud oligopolies. There is no competition and there will be massive rentier capitalism extraction from all these players for the rest of civilization.
There's competition, but no one is interested.
TBH, I blame IT leadership. The easy choice is Google or Microsoft. "No one ever got fired for buying IBM" is now "No one ever got fired for subbing to Office." Someone needs to bite the bullet and stitch together an alternative array of services. They've cultivated their own monoculture by hitting the easy-button and forking over for Office365, etc.
Which if we're being honest, Microsoft and Google products are rarely best-in-class at any one thing that they do. Teams, SharePoint, Google Docs, someone somewhere is doing those things much better. Those competitors just don't necessarily have a whole ecosystem that makes it a big Easy Button for corporate IT to smash. And they sometimes charge as much by themselves as Office365 costs in whole. But I know one problem my own employers IT team struggles with a bit is putting out the constant fire of different teams deciding the Microsoft tool we pay for sucks and going rogue and trying to subscribe to something else. Dropbox was a big one, but Teams and Planner alternatives come up a lot too.
Not pretending to have all the answers but it doesn't seem to be a completely intractable problem.
It starts by just avoiding them as much as possible. E.g. self-host email. If you encounter a system that refuses to receive your email, I give you permission to use your old Gmail account with that system, while using your self hosted account for everything you can.
If your email is Google or Microsoft, there are other people with self-hosted email who feel pressured to use their Google or Microsoft accounts to send email to you.