Bullpen Thesis I: X is the Endgame
The pitch is short: every crypto market, one app. A single surface where you can trade anything that matters. But that is only the floor. The real bet is that the intersection of social and trading is where the next wave of alpha discovery already lives, and that X is the endgame for that intersection. Bullpen — bullpen fi — is built to supercharge it. The goal is simple to picture and hard to build: find alpha while you scroll, act on it with one tap, and get paid for paying attention to the right people.
This is the first piece in a thesis series. It sets the frame — why we build on top of X, and the design constraints that fall out of that decision. Every later choice traces back to the reasoning here, and to the scar tissue of features we have shipped, measured, and retired against real market feedback.
What is a bull pen?
A bull pen, in the old Wall Street sense, is an open trading floor: rows of desks, no walls, analysts and traders and managers packed in together. Information moves at the speed of someone turning their head. I have worked in that setup my whole career — intern desks on Wall Street, the ShapeShift office in Denver, the Galaxy research desk. Bullpens compress learning. They maximize transparency and the speed at which an idea becomes a position.
That is the DNA we want the product to carry, from the backend down to the pixels. Bullpen is not just an app; it is a digital bullpen, built so that ideas flow as fast as trades.
The stall: ten years since a new social app broke out
From 2004 through 2016, social media was an explosion: Facebook, Twitter, Instagram, Snapchat, TikTok, Discord. Each reshaped how billions of people connect and consume. Then, sometime in the late 2010s, the breakout social app stopped appearing. There were flashes — Clubhouse, BeReal — but none endured. Meta could not even keep Threads relevant; a hundred million signups in 2023 bled out by eighty percent within months.
Why has nothing sustained since TikTok and Discord? Three forces, stacked together, explain the stall.
Algorithmic takeover: from following to For You
The shift from follower-based feeds to algorithmically ranked For You feeds, pioneered by TikTok, changed the unit of competition. Recommender systems do a few things orders of magnitude better than social graphs: a creator with a tiny following can blow up, a user can discover interesting content without curating anything, and the content well never runs dry. The cold-start problem — the traditional kryptonite of any new social network — mostly dissolves. A new user no longer clicks through a list of interests during onboarding; the feed learns them in roughly twenty swipes. New users join faster and retain harder.
Incumbent dominance: walled gardens win
Instagram Reels and YouTube Shorts watched the recommender advantage with, apparently, zero downside, and copied the model into products that already had hundreds of millions of users. They prioritize discovery over the accounts you followed, which creates a flywheel: creators distribute where discovery is highest, which keeps supply and demand locked inside the incumbents. For any emergent app this is not a headwind, it is a death spiral. Why join a new app when TikTok, Instagram, and YouTube already know what you will love?
Private graphs: the group-chat era
The least-discussed force is the rise of the group chat. From the mid-2010s on, people stopped posting publicly unless the post was performative, and moved the real conversation into private spaces — iMessage, Telegram, WhatsApp, Discord. Discord crossing a hundred million monthly users marked the moment. Intimate, entrenched networks with established usage patterns are extraordinarily hard to pull users out of. A new social app is not just competing with another feed; it is competing with a group of friends.
Why video is the wrong shape for trading content
Every platform's algorithm rewards a distinct kind of post. Instagram rewards polished, visual-heavy output. YouTube rewards long watch time. TikTok rewards short, trend-tied clips that maximize session length. X rewards concise, data-dense takes that amplify a sharp idea.
These differences are not cosmetic. They decide what kind of knowledge survives on each platform. The best traders tend to be thesis-driven: they work from data and first principles, and they reason slowly. That method sits awkwardly with the video bias of TikTok and Reels.
- Trading is cognitive and data-dense; it rewards pausing, rereading, and verifying. Text does that in minutes; video does not.
- Many of the best traders on social are pseudonymous. Video rewards faces and reaction thumbnails; text rewards the idea.
- Charts and dense arguments compress poorly into fifteen seconds. The supply of great trading content on X is, by our estimate, an order of magnitude larger than on any short-form video venue.
This is not a claim that no good trading content exists on video. It is a claim about the distribution: text is where the depth lives, and we expect that to hold for a long time even as AI tools — bullpen ai and the like — lower the cost of producing video. Bullpen is anchored to that bet.
Where Bullpen fits
Bullpen is the connective tissue between elite content creators on X and the output that actually matters: PnL. As a crypto trading platform — bullpen finance built for this audience — we are building a transparent place where a valuable insight leads to a measurable result that anyone can learn from and trade on. Trading is becoming a spectator sport. End users want to watch pros navigate volatile markets and complex onchain products with precision, and they want to act on what they see without getting exploited in the process.
The thesis series continues from here. The next installments get into the specific design constraints that follow from building on X, and how each one shows up in the product. The short version: text-based alpha is here to stay, and we are building the place where it becomes executable.