Substack is no longer just a bit annoying; it’s a live demo of how platforms rot when growth metrics outrun product sense. What started as a clean publishing layer is turning into another over-engineered, over-algorithmic UX mess.

The problem, and a better path
The good news: everything Substack is currently mangling, identity, delivery, discovery, and reading, can be done better with tools that don’t treat users like data sources first and readers second.
Alternatives like Ghost, self-hosted stacks, or even lightweight static-site-plus-email setups offer more control, better performance, and far less dark-pattern noise.
If Substack wanted to reverse the slide, it would need to:
Simplify authentication
One coherent flow: either true magic links or proper passwords with modern options like passkeys, not a Frankenstein combo of email followed by two factor verification. Clean state handling, minimal redirects, and predictable sessions would turn “security theater” back into actual usability.
Stop aggressive data harvesting
Kill default contact scraping and design discovery features that don’t require raiding someone’s address book. A transparent graph built on follows, referrals, and explicit sharing would still power recommendations without optioning user trust.
De‑clutter the interface
Respect “no” when users decline notifications, and reduce modal spam to rare, high-signal events. A writing and reading tool should feel like a document editor with distribution, not a Vegas lobby of pop-ups.
Tame the algorithm
Make the “For You” feed strictly opt-in, with visible controls, and keep default views subscription-first and chronological. Algorithms should be assistive filters, not gatekeepers between writers and readers.
Fix the front end
Reduce unnecessary re-renders, avoid over-coupled client-side routing, and optimize scripts so reading doesn’t trigger repetitive refresh storms that make the platform useless for subscribers.
A publishing platform that can’t reliably hold text on screen has a fundamental engineering priority problem.
Saner alternatives for builders
For anyone tired of wrestling with enshittified platforms, there’s a healthier stack:
Ghost + email provider
Ghost gives you subscriptions, memberships, and posts on your own domain, while services like Mailgun or Postmark handle the email legwork. You keep the data and the relationships.
Static site + newsletter tool
Pair a static site generator (Hugo, Astro, Next static export) with a mailing service like Buttondown or a simple SMTP-backed solution. Fewer moving parts, almost no JavaScript bloat, and excellent performance.
Custom, minimal auth layer
Use a focused auth service (or a well-implemented passwordless/passkey flow) instead of bolting on increasingly complex, brittle login journeys. Fast, predictable authentication is the foundation of any serious publishing UX.
Substack is choosing complexity and control where simplicity and user agency would win. The tragedy isn’t that these problems are hard; it’s that better answers already exist, just usually outside platforms that feel entitled to own the whole stack.
