The Future of Email

Original Article ↗ Hacker News Discussion ↗

Overall Reaction to the Article

  • Many see the piece as lightweight marketing rather than substantive vision: big title, little concrete “future.”
  • Some were expecting announcements about JMAP or real protocol evolution; instead they saw a rehash of SPF/DKIM/DMARC.
  • Others appreciated the clarification that Fastmail is not silently running AI on user mail, only exposing an optional MCP endpoint.

“Secure Message Centers” vs Email

  • Strong frustration with bank/healthcare/government portals that send a “you have a secure message” email and force login/MFA to read simple notices.
  • People dislike that these systems break the searchable, long‑term archive nature of email and are often inaccessible or hard to back up.
  • Defenders point out regulatory and HIPAA/BAA constraints, plus ticketing/workflow reasons; critics note that paper mail and fax remain allowed, which feels inconsistent.

Authentication: SPF, DKIM, DMARC, BIMI, ARC

  • Many argue DMARC enforcement by Google/Yahoo was painful but ultimately good for the ecosystem and forced institutions to fix broken setups.
  • Others worry about power concentration: when big providers “require” something, it effectively becomes global policy and can squeeze small/self‑hosted servers.
  • Discussion of BIMI (logo display) and ARC (forwarding) appears, but seen as marginal or even failed experiments; BIMI certificates are criticized as expensive.

Encryption, PGP, and Security Model

  • Multiple comments: yes, end‑to‑end encryption is technically possible (PGP, age, Proton, Delta Chat), but key distribution, UX, and ecosystem integration are major blockers.
  • Some argue widespread encrypted email is overrated: TLS already protects transit, headers stay exposed, and encryption breaks spam filtering and server‑side features unless servers hold keys, which undercuts the benefit.
  • Others want a future where organizations sign/encrypt all mail, possibly with in‑person key exchange; however, getting users to adopt this remains difficult.

Spam, AI, and Inbox Control

  • Concerns that AI agents acting on email could be easier to fool than humans, while simultaneously being used to generate more convincing phishing.
  • Suggested mitigations: whitelisting/“friend request” style systems, HEY‑style screening, masked/alias addresses, and even proof‑of‑work or “proof of human effort.”
  • Some think HTML removal and invite‑only inboxes would help; others say HTML is manageable and the real issue is consent to receive.

Fastmail, Proton, and Hosting Options

  • Many praise Fastmail’s reliability, JMAP support, aliases/masked email, and lack of enshittification; others report weak spam filtering and clunky desktop app UX.
  • Proton trade‑offs cited: mandatory apps, limited integration (no CalDAV/contacts sync), poor search, alias limits.
  • Self‑hosting is viewed as still possible (with modern stacks and proper SPF/DKIM/DMARC) but increasingly constrained by IP reputation and big‑provider policies, reinforcing a drift toward email as a semi‑walled garden.