Why did he leave?

TBH no one really cares. As long as the network basically appears to work, most of the time, the vibe enforcers respond to any outage with reminders of "boolish!"

The tech is truly not good. Maybe it's good enough for now, and will get better, but today:

• Hub data goes out of sync often and regularly. The unreliability gets papered over by Neynar's AWS Hubs, a de facto source of truth that is heroically kept in sync with the actual source of truth, Warpcast's private Postgres instances. Pinata discovered within ~1 month that they couldn't keep it up. OTOH, Airstack's founder is Adm. David Farragut enough to tell his team, "Damn the torpedoes, full speed ahead!"

• It actually happened. The network can be brought down for $150k~200k. Spin up ~1000 nodes, let run for a few days. One might be charitable to the team and teach me about Waldian survivorship bias, and point to the Honda method: run an engine millions of hours until it breaks, then direct engineering talent towards fixing weak spots revealed under stress. But I challenge every programmer to review the actual fix they made to production code in the public repo.

• Auth is leaky (in practice, not in theory!), yet the team has no impetus to fix it after multiple cases, because it's not a priority. Thought experiment: how technologically easy would it be for an anon indie client dev to use my signer to spoof me? Can anybody protect $100k of value behind a frame secured only by a Farcaster login?

Case in point: channels are the same people over and over, simply with different vibes (kia.eth's screenshot essay). Maybe this is a feature, not a bug. I think no, it's actually an emergent symptom of being very very smol, combined with the team being very good at getting users to drink the kool-aid.

Furthermore, the core community is close-minded, enjoys cult-like echo chambers (a feature, not a bug! of early adopters), accepts wild levels of favoritism while being just welcoming enough towards newcomers to dump on them. Study $DEGEN and now $MOXIE's adoption, distribution and patterns of capital flow.

The smaller, non-core but still tightly-knit outsider communities (organized farmers, paidgroups, artists, non-English pioneers+tourists) mostly just want the same benefits that they have seen accruing to the earlier insider community. Study Proxy Studio's singularly successful (and non-repeatable) infiltration of the in-crowd, mainly via Proxy's own correct identification of Hypersub and Jonny Mack as important partners, the spark recognition by Ted, and ruthless execution. Compare and contrast with the growth methods of other post-permissionless accounts that have collected large followings.

Meanwhile, some accounts have indeed gained popularity by public poking at @dwr.eth. They appear to be mostly theater, mixed with a bit of twisted hopium and/or nihilism. Their bark produces hoots and hollers, but what change has been wrought by any of these critics, over 1, 2, or 3 years? Do they ever say something that risks losing popularity with their existing audience? If not, it's just social performance, without proposing concrete steps to bring about the change they want to see.

This community is not bound by the core values of earlier phases of crypto adoption. Decentralization, transparency and censorship-resistance got rebranded with the practical-minded "sufficient" prefix. The community attracts opportunists and frontier types who enjoy FAFO with unspoken rules (!!), and accept that "being early" is synonymous with dumping on the exit liquidity. :astronauts-always-has-been:

Stated: grow by becoming an appealing alternative to CT

Revealed: protect the algo feed like a rabid dog, maintaining the ability to push individuals important to the Warpcast founders' future, such as Fred Wilson and Jake Chervinsky

Stated: allow the community to contribute to protocol development via the Farcaster Improvement Protocol process

Revealed: Vitalik wants longcasts. If we have to run over the dev community, most ppl will cheer the shipping speed while wondering "where did they go?" about the open-source fanbois

The current growth path appears to appeal to the worst of non-Farcaster:

• "wake up babes, they're giving out cars and getting paid"

• encourage moar content with an ML algo trained to discover the fat part of the distribution

• wait for new users to self-select for those who are best at getting engagement of the influential in-group

Can Farcaster attract the high notes of non-FC? It's not obvious to me that any feature currently cooking in the oven is even directionally correct:

• the wild west of TG/Discord/CT/Solana-land

• deep and evergreen knowledge share on Reddit/Xiaohongshu

• high quality fleeting entertainment a la TikTok/IG/YT creators

• upstream discussion on Substacks and in GCs

• breaking news on X

• dissident expression on fully-encrypted platforms beyond the reach of state actors

• a town square for product launches a la peak HackerNews

Any FC business is merely playing for leftovers that Merkle hasn't placed a clear stake in the ground for themselves, or for their friends.

Study Ponder vis-a-vis Vocdoni, as part of the Composer Actions rollout.

Study Daimo. Merkle saw a clear threat to a space they consider a key part of their future, and therefore placed a stake in the ground: this is ours, y'all can play outside of our part, and eventually once the tech is ready (and perhaps after y'all have proven value and fit via user behavior), we will claim this value chokepoint.

Buoy won't have a good exit b/c Warpcast can crush them w/ ~1 month of work; will have to accept a discount for what they have proven to be able to deliver.

Study Supercast in the short- to medium-term. After the pivot, will they be more "playing a different game" than today, or not? See how quickly Woj decided he needed to hide casts (he buried them even more deeply than Warpcast!) after his audience of people-who-pay-in-order-to-not-be-hidden abandoned the platform, leaving him with his much smaller original core of people-who-pay-and-do-not-wish-to-see-low-value-accounts.