Fastmail

Fastmail
Source: The Email Provider Google Doesn't Want You to Know About (Fastmail Interview)

Here is a summary of Henry's interview with Ricardo Cignas, Chief Engineer at FastMail, highlighting significant and surprising points.

Summary

Ricardo Cignas has spent 20 years in email—10 at Pobox.com, which FastMail acquired a decade ago. He remains driven by FastMail's four core values: being stewards of your data, good internet citizens, and prioritizing user empowerment. These values, he emphasizes, are lived daily, not just marketing slogans.

FastMail vs. Gmail and Proton

Against Gmail, Ricardo argues FastMail wins on both ethics and experience. Gmail's business model—large-scale data analytics—means "if you're not the customer, you're the product." More personally, he finds Gmail creates "friction" in daily use, while FastMail offers "no friction" productivity 1.

Against Proton/Tutanota, FastMail deliberately avoids native end-to-end encryption. Ricardo explains this is a principled technical stance, not a limitation:

  • Email standards (RFC 822) transmit messages in the clear
  • Key distribution for strangers remains unsolved
  • "Walled garden" encryption breaks when emailing outside the ecosystem, creating "weird stuff"—signed-but-unencrypted messages, or emails that say "you have an email" and redirect to web portals 1

> "I don't ever want to see [email] go away. I'd like to see it continue to evolve."

Surprising Technical Revelations

Point Detail
Phone number resistance Ricardo: "I don't want to know your phone number. I don't want to know anybody's phone number who's using FastMail." Required only for fraud prevention, not identity 1
VOIP workaround was intentional Mobile signup bypasses phone verification—"at least somewhat on purpose" 1
JMAP origins FastMail created JMAP ~2014 to replace IMAP's inefficiencies; now an IETF standard (RFC 8620/8621) 1
Offline architecture Browser runs a "tiny JMAP server inside of the tab" for caching—enabling offline without native apps 1
Support staff see "lorem ipsum" Customer service sees structural placeholders, not actual content, unless users explicitly opt in 1

Open Source Philosophy

FastMail maintains Cyrus (92% of commits in last 10 years), the leading open-source IMAP/JMAP/CalDAV/CardDAV server. However, Ricardo defends not open-sourcing everything:

  • "Every extra button is a problem"
  • Open-sourcing creates "enormous" maintenance overhead
  • "How do we produce the most benefit for the most people?" sometimes means standards work over code dumps 1

JMAP: The Hidden Speed Engine

FastMail's name predates JMAP by over a decade, but modern speed comes from layered engineering:

  1. Ajax UI (2013-14): Shifted to browser-based application with local caching
  2. JMAP: HTTP/JSON protocol replacing IMAP's "Shakespearean" complexity
  3. Optimistic UI: Delete appears instant; server catches up after 1

Security benefit: JMAP uses standard web push (RFC 8030) instead of persistent connections, eliminating the need for cloud services to store OAuth tokens for push notifications—a significant improvement over how IMAP clients handle Gmail on iOS 1.

The "Good Internet Citizen" Ethic

Ricardo's most passionate argument: the internet's federated, decentralized nature is fragile. Being a "good citizen" means:

  • Building tools that let users separate life contexts ("you might as well be two strangers communicating with yourself")
  • Supporting open standards first, proprietary APIs only when unavoidable
  • "Don't be a dick"—explicitly stated as corporate and personal responsibility 1

Future Directions

  • Teams/small business expansion—carefully adding features without "fruit salad" complexity
  • JMAP for calendars/contacts—standards finalized, implementation polishing
  • Automatic OAuth configuration—eliminating "IMAP server name" setup friction entirely 1

Key Quote on Privacy Trade-offs

> "There's always that slider between... convenience and secrecy. And people make their choice on that front."

FastMail's position: they provide privacy (data stewardship, no analytics) and security (infrastructure hardening, zero-trust goals), but not secrecy (mathematical unavailability of data). For that, they explicitly recommend Signal 1.