Vision for Pigeon
The first new email client for Apple platforms in over 200 years
The email client selection for Apple platforms is pretty abysmal. Long gone are the days of Sparrow. Now, we are left with Apple Mail (slow and awkward), Spark (AI-laden design wasteland) and Thunderbird (works mostly for techies). I won't even mention Outlook, because we all have three versions of it, and none of them work.
This is where Pigeon comes in, and you can learn about how it will work.
The Problem
Where are all the email clients?
Ask any Apple user on the internet about which email client they use. The vast majority will say (Apple) Mail. It is the default, after all, but its reputation is severely tainted by the myriad of bugs that go unfixed, half-baked backports, and glacial updates.
Those that came from Linux will say Thunderbird. For its use case, it's great… if you follow the Linux philosophy. It's a behemoth that does everything, and none of it as well as it could.
Then there's Spark, a juggernaut straight from the depths of Electron, that fails to realize that JavaScript is not the best backend for shoving useless AI down our throats.
The Solution
Pigeon will have a simple mission, as Cork had, and accomplished it, in the Homebrew space before it: to bring a modern, usable e-mail client to the masses. And there's a lot of potential.
The Usual Email Things
Of course, Pigeon will support all that we expect from an e-mail client.
The Less Usual
In addition, there will be things like `placeholders`, e-mail chain consolidation, and more useful, yet not often explored features.
Modern Tooling
Much like Cork, Pigeon will be made in SwiftUI, and no bloated frameworks like Electron will be used.
No AI?
We are all sick of the constant AI slop. Pigeon will be a completely human-made app, and no AI will be used in its development.
Will LLMs be used at all? Like I said, not in its development. But there's one tiny use case where an extremely limited LLM might be useful, if certain strict conditions are met, and it's not used to generate anything, only answer a simple question.
One idea that made me think of starting developing Pigeon were the constant job rejections we all face in the current climate. What if there was a very local, very small, very limited LLM, like the one built into macOS, that could privately scan emails coming in from a specific address (for example, a recruiter), and answer a simple YES / NO question: was I rejected? And if so, send that email to the trash and don't even bother me with reading the auto-generated rejection.
I'll see where I can go with this idea. But I can say even now, nothing in Pigeon will be generated.
Closing Statement
Pigeon is still in its infancy. For now, it is a rudimentary demo of SwiftUI navigation. But it will grow and develop, and no doubt mature into a great product.
