The post is a personal account of quitting Gmail after years of tolerating creeping annoyances, with the breaking point being AI features pushed directly into core email actions. The author describes auto-generated summaries, suggested replies, writing prompts, and “improve” nudges as disrespectful friction in a product that should stay out of the way. They also note that some of the most intrusive AI can only be disabled by turning off older useful features like Gmail’s automatic inbox categorization, which made the whole thing feel less like optional assistance and more like coercion. The author moved to Fastmail with a custom domain.
For executives, this is a clear signal that users increasingly see forced AI as product debt, not innovation, and that smaller paid tools can win by being faster, simpler, and aligned with user intent.
Strongly negative toward Gmail and forced AI features. People were annoyed by the patronizing UX, skeptical that the features are useful, and increasingly open to paying for simpler email products that respect user choice.
01 Fastmail is winning here less on privacy ideology than on product discipline.
People who moved from Gmail kept coming back to the same things: it is visibly faster, stays out of the way, has barely changed in years, and feels like software built to do one job instead of a distribution channel for adjacent bets. The interesting part is not that a paid niche service exists. It is that users now describe Gmail as the clunky option and Fastmail as the polished one, which flips the old expectation that scale automatically buys better UX.
A focused paid product can now out-execute the default giant on basic usability. That is a dangerous signal for any platform team stuffing extras into a mature workflow.
02 Gmail still owns one feature that competitors have not convincingly matched: automatic inbox categorization that works with no setup.
Users who depend on Primary, Promotions, Social, and Updates said this is not cosmetic. It is how they keep a busy inbox usable without maintaining brittle rule systems. That means Gmail’s moat is no longer “better email overall.” It may be one or two sticky workflow features that keep people despite broad dissatisfaction.
Users will forgive a lot if one workflow feature saves them real daily effort. Replacement products need a credible answer for that last sticky habit, not just a cleaner UI.
03 The strongest defense of AI-written email came from people who struggle with writing for structural reasons, not laziness.
Non-native speakers, dyslexic users, and people in high-stakes bureaucratic or legal contexts said LLMs can be genuinely useful for translating blunt intent into socially acceptable business language. That cuts against the dominant disgust with generated email. The useful framing is that the problem is not assistance itself. It is making that assistive mode the default for everyone and then dressing it up as universal productivity.
AI writing help has real accessibility value. The product failure is forcing an accommodation into the mainline workflow instead of keeping it explicit and user-controlled.
04 Gmail’s reputation for unbeatable spam handling looks a lot shakier than its legacy brand suggests.
Multiple people said legitimate messages, including court-ordered class-action notices, reliably land in spam or vanish, while obvious marketing junk still gets through. One commenter’s more interesting claim was that this is less about technical incapacity than about where Google chooses to invest attention. If true, Gmail is no longer optimizing for “never miss wanted mail.” It is optimizing for a rough equilibrium that is good enough at internet scale.
At large scale, “good enough” filtering can become a liability when the provider stops caring about edge cases that matter to users. Brand trust can lag real product quality for years.
05 The migration advice converged on a pragmatic playbook: buy a domain first, do not try for a clean break, and use the old Gmail account as a forwarder while you update accounts piecemeal.
People who had actually done it stressed that the hardest part is not moving messages. It is unwinding identity, recovery, and login dependencies built over twenty years. That makes email migration less a technical problem than an account-graph problem.
Owning the domain is the strategic move. It turns a painful one-time provider exit into a manageable background process for the next switch.
01 For some users this entire problem is self-inflicted because Gmail’s smart features can be disabled cleanly enough.
Several people said they turned off smart features long ago and never see the AI prompts, summaries, or draft generation at all. That weakens the claim that Gmail is universally unusable. It suggests rollout state, browser differences, and settings history are part of why experiences vary so widely.
The worst Gmail experience is not universal. Some of the outrage reflects a messy rollout and opaque settings model as much as the existence of the features themselves.
02 Not everyone sees AI-assisted writing as fake or insulting.
Some people use it as a social safety layer that catches unintended harshness, rewrites dictated drafts into something presentable, or reduces the cost of necessary but tedious communication. In that frame, the feature is not “Google thinks I’m stupid.” It is “Google assumes many users would rather not spend twenty minutes polishing a routine message.”
There is real demand for tone and drafting assistance. The objection is mostly about coercive presentation, not the existence of the tool.
03 The anti-Gmail mood sometimes overstates the alternatives.
A few people said Gmail still wins on spam filtering, raw ecosystem convenience, and the simple fact that most of their digital life is already tied into Google. For them, minor UI irritation does not outweigh deliverability confidence, integrated services, and the cost of switching. That keeps Gmail sticky even among users who dislike its direction.
Incumbents can absorb a surprising amount of product decay when the surrounding ecosystem is still valuable. Frustration alone does not always produce churn.