The wrap-up era is over. We read every HOGE Weekly so you don’t have to.
Daily and weekly crypto wrap-ups became commodity slop the moment AI could write them. Here is what we learned from reading every HOGE Weekly we ever published, and what is replacing them.
We have published 84 HOGE Weeklys since the format launched in October 2024. We have also, over the same period, watched the crypto-newsletter ecosystem absorb something on the order of 2,000 new “this week in crypto” products from competitors, AI farms, marketing automation tools that disguise themselves as Substacks, and one notable instance of a person who wrote their entire Monday digest by feeding the previous week’s HOGE Weekly into an LLM and asking for it back in a slightly different tone of voice. Last month we sat down with the editorial team and read every HOGE Weekly we ever produced, in order, and we made a decision: the wrap-up era is over. This is the piece explaining why, what we learned, and what we are going to do instead.
What is at stake, for us and for any publication that takes its readers seriously, is the difference between content that exists because someone had something to say and content that exists because the calendar demanded a publication slot. The weekly digest, as a format, started life as a useful service in an era when readers could not realistically keep up with a fast-moving industry and needed a trusted curator to bring them the important developments. It has become, almost universally, a low-value summarisation product that competes with thirty similar products in every inbox. Banking had a similar evolution forty years ago when the morning “market open” call shifted from a value-add by the trading desk to a commodity output by an automated terminal. Crypto wrap-ups are at the equivalent moment.
Why we read every one of them
The exercise was not nostalgic. We were trying to answer a specific question: if we picked the 84 HOGE Weeklys in chronological order, how many contained anything that we would not have published in a more durable format anyway? The honest answer is twelve. Twelve weeks — out of 84 — produced original reporting, distinctive analysis, or framing that we would defend on its own terms. The other 72 were competent summaries of things the reader could have found on Block, The Defiant, or the relevant project’s own Discord. Many of them were better-written than the originals. None of them did anything that justified asking a reader to spend twelve minutes of their Sunday morning reading a 1,800-word newsletter.
| Category | HOGE Weeklys | Share | Notes |
|---|---|---|---|
| Original reporting / framing | 12 | 14% | Pieces we would defend on their own terms |
| Competent summary, no original angle | 43 | 51% | Restated coverage from major outlets |
| Heavy on price commentary | 18 | 21% | Mostly post-hoc explanations of weekly candles |
| “Look at this funny tweet” digest | 9 | 11% | Memetic content with thin analysis |
| Pure project-news roundup | 2 | 3% | Mainnet launches, governance votes |
The 51% in the middle row is the painful number. Slightly more than half of what we asked our writers to produce, and slightly more than half of what we asked our readers to consume, was material that any moderately literate crypto reader could have assembled themselves from a CoinDesk push notification and a glance at the relevant project’s Twitter. Our writers are talented; the system was asking them to do work that did not use their talent. That is the structural problem we are responding to, and it is not unique to us.
The AI farm problem
Sometime in early 2025 it became trivially possible to produce a credible-looking weekly crypto digest with a single prompt to a frontier language model. The model would read the week’s headlines from a couple of RSS feeds, identify the dominant themes, write a 1,500-word summary in whatever voice the operator wanted, and ship it to a Substack with a generated header image. The cost per issue, including infrastructure, was roughly forty cents. The format collapsed into a commodity almost overnight. Search “weekly crypto wrap-up” on any platform that ranks newsletters and a substantial fraction of the results are now AI-generated; SEMrush data suggests that the long-tail search volume for crypto-newsletter terms is now dominated by domains that did not exist twelve months earlier.
This is not a moral complaint. The AI-farm operators are doing what the format invited them to do. The original weekly-digest model was, even when humans wrote it, mostly a transformation operation: read sources, compress, restate in house voice, ship. That operation is exactly the thing language models do well. The honest response is not to wring our hands about authenticity; it is to acknowledge that the format has lost its defensible moat and to move our editorial energy somewhere a model cannot follow. The banking comparison: when ATMs replaced teller cash withdrawals, the banks that survived did not double down on producing better tellers. They redeployed the people to advice work that machines could not do.
What replaces the wrap-up
The replacement, for us, has three components. The first is what we are calling Field Notes — pieces that report on something a writer actually saw or did. Marcus’s six-months-on-a-Lightning-node piece is a recent example; the on-the-ground report from Token2049 in Singapore was another. These are 1,500–2,500 word pieces that read more like long-form magazine journalism than newsletter content. They take longer to write and they cannot be summarised in a tweet, which is exactly the point. The cost of producing one is roughly the cost of producing four weekly wrap-ups; the marginal value to a reader is, we believe, substantially higher than that ratio.
The second is Backstops — technical reference pieces that explain something the industry routinely refers to but rarely explains correctly. The recent piece on Solana’s optimistic-vs-rooted confirmation paths is an example; an upcoming piece on the difference between intent-based and order-based DEX architectures is another. These are reference works that we expect readers to come back to, not consume once. They reward depth and they punish shallow summarisation, which is the structural property we want our format to have.
The third is HOGE Wire, our event-driven publishing tier. Wire pieces ship when something has actually happened that warrants reader attention — a major protocol upgrade, a significant regulatory action, a market structure event. We will publish three or four Wire pieces in some weeks and zero in others. The cadence will follow the actual news rather than the calendar. This is the part of the transition that most newsletter operators find hardest, because the predictability of a weekly cadence is what funds subscriber-acquisition economics. We are accepting the trade-off because predictability has become indistinguishable from filler.
- Field Notes: reported pieces from real situations, 1,500–2,500 words, on irregular cadence.
- Backstops: technical reference pieces that explain industry terms or mechanisms in defensible depth.
- HOGE Wire: event-driven news coverage published when warranted, not on a calendar.
- The Weekly Digest, as a recurring format, is retired effective with this piece.
What we got right in the wrap-up era
It is worth saying clearly: the wrap-up format was not a waste. The twelve issues we would defend on their own terms included the November 2024 piece that called the EigenLayer airdrop economics correctly before the broader market understood them; the March 2025 issue that explained the Hyperliquid airdrop’s structural difference from the airdrop norm and which our subscribers credit with shaping how they evaluated the next cycle of token launches; and the December 2025 piece that walked readers through the Berachain mainnet launch’s three confusing token mechanics in a way that nobody else did at the time. Those pieces were good. They were also smuggled inside a format that produced 72 other pieces of varying quality, and the smuggling was the problem.
What we also learned, reading the archive, is that the best wrap-ups were the ones where a writer broke the format. The November 2024 EigenLayer piece was technically that week’s wrap-up but devoted 1,200 of its 1,800 words to a single argument and only 600 to the rest of the week’s news. The format had become, at its best moments, a vehicle for the writer to say one substantial thing and then perform the ritual of summarising the rest. We are taking the substantial thing and dropping the ritual.
What this means for your inbox
Sundays will get quieter. Your existing HOGE subscription will continue to deliver everything we publish, but the rhythm will change. In a quiet week you may receive nothing at all; in a busy week you may receive four or five pieces, each shorter than a wrap-up and each tied to a specific development. The trade-off we are asking you to accept is that you will sometimes receive nothing from us and that this is intentional rather than a sign that something has gone wrong with the publishing schedule. We will publish a monthly editorial note — not a wrap-up, an actual letter from the editorial team — that explains what we have been working on and what is coming.
For readers who genuinely valued the wrap-up format and would like a substitute, we will quietly recommend three: the daily “What’s Hot” thread on Stacker News for Bitcoin-flavoured curation; the Milk Road newsletter for entry-level macro and altcoin coverage; and the all-core-devs meeting notes on github.com/ethereum/pm for anyone who wants to know what Ethereum protocol developers are actually arguing about. None of these replace HOGE; they replace the part of HOGE that we no longer want to be. Our market dashboard covers the price and flow data that wrap-ups used to summarise, and our calculators page covers the practical tools that readers most often asked us to embed inside the weekly format.
The honest version of why we are doing this
The honest version is that we are doing this because we became bored of our own product. The HOGE Weekly took roughly fourteen hours of editorial time per issue to produce. We spent fourteen hours producing 84 issues. The vast majority of those hours went into work that did not, in retrospect, justify itself. We could have spent the same hours producing fewer, better, more durable pieces, and we want to find out what happens if we do. The risk is that our subscriber numbers fall in the short term; the bet is that our readers’ actual loyalty rises in the medium term. We are willing to take the bet.
If the wrap-up era is genuinely over, every publication that built its brand on the format is going to have to make some version of this decision in the next eighteen months. Some will retreat further into commodity content and trade on volume; some will pivot to the more durable formats we are pivoting to; some will fold. The thing we wanted to do, in publishing this piece, was to be honest with our readers about the calculation rather than quietly drift into a new editorial rhythm without naming the change. The wrap-up era is over. We read every HOGE Weekly so you don’t have to. Going forward, we will be publishing things that we believe you actually do. For what is coming up on the editorial calendar, see our events page, and for the practical tooling readers used to find buried in the weekly digest, our calculators now live in one place.