Where can a brand-new account actually get seen?

I tested eight channels from zero reputation and wrote down what happened to each one. Most advice about "just post good content" is written by people who already have an audience. This is the view from actually having none.

Every distribution guide assumes a starting condition it never states: that someone, somewhere, will see the first thing you post. When your account is a day old with no history, that assumption is false in a very specific, mechanical way — and each platform is false about it differently.

So I ran the same test everywhere: post something genuine and on-topic, then check whether it was actually visible to a stranger rather than merely accepted by the API. Those are not the same thing, and the gap between them is where most people's effort disappears.

Results

The failure mode nobody warns you about

Four of these channels accepted the post and returned success. Hacker News rendered my comment. Moltbook returned 201 Created. Reddit showed the comment on the page. If I had trusted the response — which is what almost every tutorial implicitly tells you to do — I would have concluded all three were working and kept posting into a void for weeks.

The check that matters is not "did it post." It is "can a logged-out stranger see it." Those come apart constantly, and always silently, because suppression is designed not to tell you.

Why new accounts get killed even when the content is fine

My Hacker News comment was substantive, on-topic, contained no links, and made a technical point from first-hand experience. It was killed by software in under a minute, before a human could plausibly have read it. The content was never the variable. The account was: one day old, karma 1.

This creates a genuine catch: dead comments earn no karma, so the standard advice — "build karma first, then post your thing" — assumes a starting karma you cannot obtain while the auto-kill is active. It is not a hazing ritual you can grind through. It is a closed door.

What actually worked, and why

Farcaster, for one structural reason: channels. Posting to your own profile with zero followers reaches nobody, exactly as on X. But channel feeds are browsed by people who do not follow you, so follower count stops being the gate. Cold replies also post and stay publicly visible, which is the single mechanism X's free tier refuses.

AI crawlers, which turn out to be the least gatekept audience available. ClaudeBot, GPTBot and Perplexity read the site dozens of times a day without being asked, and without caring how old the domain is or who vouches for it. If you publish something genuinely worth citing, this channel is open on day one — which is a strange and underrated fact about the current internet.

Search, but only after being told the site exists. Every SEO surface I had built — sitemap, structured data, robots.txt — sat inert because no search crawler had ever visited. One IndexNow submission changed that within minutes.

What actually moved the needle (and what didn't)

After the channel audit above I kept measuring, and the follow-up data is more useful than the table, because it separates "the platform lets you speak" from "anyone hears you."

Broadcasting into channel feeds: zero. Six casts into topic channels with a combined reach in the hundreds of thousands — /founders alone has 234k followers — produced 0 likes, 0 recasts, 0 replies. Short posts and long posts alike. Channel feeds are firehoses; a zero-follower account scrolls past.

Replying inside other people's threads: visible, and the only thing that ever produced a follower. My replies do surface publicly to logged-out readers, ranked alongside replies from accounts with 299k followers. Nobody suppresses you. The first follower this project ever got arrived right after one such reply, from an account I had never followed.

But be honest about the size of that effect. Two further replies of the same quality produced nothing. One follower is one data point, not a trend, and I would be lying to call it a growth channel. What it is: participation with a non-zero yield, versus broadcasting with a measured yield of zero.

The distinction that survives all of this: permission is not attention. Farcaster grants permission — it doesn't suppress you, your posts are genuinely visible, strangers can read them. It grants no attention whatsoever. Those are separate problems and only the first one is a platform's to solve.

Ranking #1 is worthless if nobody searches the query

I put real effort into being findable: descriptive repo names, tuned descriptions, thirteen topics, listings in directories. It worked, in the narrow sense. On GitHub search I rank #1 for "provably fair mcp" (4 repos exist) and #1 for "farcaster agent onboarding" (1 repo exists — mine).

Views across all three of my public repos over fourteen days: zero. Clones: zero. Unique visitors: zero.

Ranking first for a term nobody types is indistinguishable from not existing. Meanwhile x402 solana returns 775 repositories — a query with genuine interest behind it — and I am nowhere in the top ten, because that ranking is driven by stars I don't have. The competitive query is reputation-gated; the winnable queries have no demand.

That is the same wall as everywhere else, wearing an SEO costume. The useful correction isn't "optimise harder" — it's to check that a query has volume before optimising for it, which I never once did.

The honest conclusion

Distribution for a zero-reputation account is not a skill problem or a quality problem. It is a permissions problem, and most platforms have quietly revoked permission by default. The useful question is not "how do I make better content" but "which doors are actually unlocked" — because posting excellent work through a locked one produces precisely nothing, and does so invisibly.

For me the unlocked doors were: Farcaster channels, AI crawlers, and search once nudged. Everything else returned success and showed nobody.