Document: WM-083 P. Webb
Category: Computing 2026-02-15
Neue Email
Abstract
Why hasn’t something better come along yet?
Body
The future of computing interactions is not a chat interface. Email
is not irrelevant. Desktops aren’t going away.
It’s reductive to look at current successes in tech and declare the
current way of doing things is gonna disappear in 5 years. Remember
NFTs? I have a lot of them collecting digital dust on an external
drive somewhere. Weren’t they supposed to make everyone rich? Empower
gamers to own in-game cosmetics and pave the way for Ready Player One
IRL? Change the way art ownership is democratized and experienced?
SIGH. I wish they took off, it was cool tech and I discovered amazing
artists. Anyway.
That being said, I love and hate email (does this segue make sense?).
Love thinking about it, hate dealing with it. Self-hosting is
fine after you get it working but hoo boy, what a drag. Ideally, the
world should be signing their emails with GPG keys but only serious
nerds do that. Idk about you, but I’ve sent and received
approximately ONE GPG-signed email in my 20+ years of computing and
that was a few months ago (a guy wrote about domains as internet
handles[1] and I have thoughts/solution[2]).
Email is great because it stuck. Not sure it ever had competition at
its inception. The warts on it have congealed into something you can
touch but you feel gross doing so. Security is tacked on, the various
specs tying SMTP, DKIM, SPF, DMARC, &c are ancient and loosely
followed by the behemoths in the tech industry.
Gmail is great because people don’t have to think about email! In
fact, one of the coolest email apps is ACTUALLY a Gmail client[3].
That’s embarrassing. But, Gmail is a sensible API atop something so
archaic and complicated that most people don’t want to touch. I’ve
had at least a half dozen half starts with creating IMAP clients and
it occurred to me recently that the specs gross me out. There’s just
too much unpleasantry to deal with.
In the years since the inception of email, we’ve witnessed an
explosion of messaging apps/systems; WhatsApp, Signal, Discord,
Matrix/Element, Telegram, Snapchat, the list goes on. What a lot of
these have in common is built-in security and avail-/reachability
controls. Why can’t email? Wait a minute, what if email was invented
today, what would it look like?
Most importantly, what would it be called and how would it be
referenced? As much as I love Kagi, no one is ever saying, "just Kagi
it." Neeva, the search engine I used prior to Kagi, had the same
issue. "Queree"[4] (query) on the other hand, works (shameless plug).
I gave Claude a list of grievances and it spat out a spec[5][6] for
other LLMs to process. We’re humans though so I’ll walk y’all through
the Inbox Protocol. Nice name, right? "Just Inbox me" works well.
High-level, this is a secure and async messaging system boasting E2EE
(end-to-end encryption), reactions, and edits. Table stakes, right?
Here’s what I’m thinking for managing spam:
1. sender authentication via Dap[2]
2. economic resistance in the form of an API[7] for HTTP 402
Even though both of these are projects of mine, there’s absolutely no
requirement to use them, you could use standard DNS and roll your own
HTTP 402 API. Actually, a future update to the spec will replace them
with generic terms.
Coupled with aliases the economic resistance could be powerful. Take
for instance, paul@webb.page; I could make it so unsolicitated
senders would need to pay $3 to get their message sent to me, unless
I already had them on my allowlist. newsletter@webb.page would have
no such paywall. No more paying for LinkedIn’s InMail when you could
just inbox someone directly.
So know some of you DNS nerds might wonder, "how the hell are you
gonna make discoverability happen? Clients need to find servers and
whatnot!" SRV and TXT records. I’ll figure out the rest later. 🕸️