There's a founder move on X that produces more qualified attention per minute of effort than anything else on the platform, and it has an unflattering nickname: reply-riding. Find a post that's taking off in front of your buyers, write the best reply in the thread, and let the post's momentum carry your words to an audience you didn't build.

The unflattering nickname exists because the move has a degenerate form: the reply-guy account that surfs every viral post with recycled one-liners, farming impressions from strangers who will never buy anything. Run that version and you'll earn reach that converts to nothing and a reputation to match. The founder version is a different discipline wearing the same mechanic, and the differences are exactly three: what you ride, when you catch it, and what you say.

What to ride: the audience test, then the relevance floor

The degenerate form optimizes for the post's size. The founder version optimizes for whose replies section it is: ride conversations happening in front of your buyers, which usually means the creators and operators your buyers follow, plus the threads where your product's problem is being discussed by name.

Build the list deliberately. A dozen-or-two accounts whose audiences overlap yours: the practitioners in your niche who post daily, the bigger voices your customers quote, the adjacent-tool founders whose threads fill with your exact ICP. This list is a real asset; curate it like one. Prune the accounts whose threads never produce a conversation worth having, add the ones that keep showing up in your customers' likes.

Then hold the floor: relevance is binary, not a dial. If you don't have something genuinely on-topic to add, the answer is no ride, whatever the post's numbers. An off-topic reply on a viral post is worse than silence: the readers it reaches learn only that you interrupt, and an off-topic reply is exactly the shape reply-quality filters exist to bury. The test that keeps you honest: would this reply make sense to the post's author as a contribution, or only to you as a distribution play? If the second, close the tab.

When to catch it: early beats big

Post 1 covered the physics: the feed runs on velocity, and reply visibility compounds from the front of a thread. The practical target is a post that's rising, not risen: minutes to a couple of hours old, engagement arriving faster than its author's baseline, reply count still thin enough that a good answer can anchor.

Two refinements operators converge on, both worth stealing:

Measure takeoff against the author, not the world. A post with 200 likes means nothing in isolation: for one account that's a quiet Tuesday, for another it's the best post of their year. The signal is the outlier: this author's post moving multiples faster than this author's normal. Outlier-vs-baseline finds the breakout in a mid-size account's thread (where your reply can still matter) instead of pointing you at celebrities (where it can't).

Decay is steep, so speed is structural. The difference between replying in minute 20 and hour 6 is usually the difference between anchoring the thread and decorating its basement. You cannot check X twice a day and catch rising posts; the math doesn't work. Either you're sampling the feed frequently during your active hours, or something is watching for you and surfacing the risers (post 5 covers the watching machinery; post 8 puts it in the routine). This is one of the few places in founder GTM where infrastructure genuinely beats diligence.

What to say: the anchor reply

The reply that earns the ride has a recognizable shape, and it isn't clever. It's additive: it gives the thread something it didn't have.

Extend, don't applaud. The author made a point; you have the adjacent point their post implies but didn't reach, the data from your corner of the space, the failure mode they didn't mention. That's the reply readers stop on.
Disagree usefully when you disagree. A specific, respectful "this breaks down when..." outperforms agreement everywhere on X, because it starts the exchange the algorithm feeds on and readers actually learn something. Pick the side you genuinely hold; symmetric both-sides mush is invisible.
Write it in your voice, at conversational length. Two to five sentences usually. Not a mini-essay (that's a post, and post 3 covers where it goes), not a one-liner. If your replies read like a chatbot, the ride is wasted; threads full of professionals smell template instantly.
Product mention: almost never. The ride's payoff is profile visits and follows from people who liked how you think, not clicks from a link you dropped. Your bio and pinned post do the converting (same funnel logic as every channel: the reply earns the click, the profile closes it). The exception is the thread that literally asks for tools; even then, disclose and name competitors fairly.

What riding builds (and what it can't)

Run this discipline for a quarter (a handful of well-chosen rides on most days, per the rhythm rules in post 4) and the compounding is real: the creators start recognizing you, their audiences start following you, and your own posts stop landing in a vacuum because the people who've read your replies now populate your reach. Reply-riding is how a small account bootstraps the audience that makes every other X play work.

What it can't do is close the loop alone. Riding borrows other people's rooms; eventually you need your own room, where the conversation happens under your posts and every replier is a hand raised at you. That's the next post, and it's the half of the engine founders skip.

If you'd rather not run the tedious part by hand, that's the part Thread Otter does: it watches the conversations in front of your buyers, surfaces them while they're still fresh, filters the noise before anything reaches you, and drafts the reply for your call. The list of who's worth riding, and the opinion in the reply, stay yours.

This is post 2 of the 8-part X for founders series. Next up: your own comment section, the most valuable surface on X that founders systematically waste.