Recently, I've found my preferred way to interact with lots of products is getting Claude to use a CLI tool, and then just talking to it out loud about what I want to happen.

So I made a CLI tool for Ghost, to see what it would be like in our own product - it's called ghst and it essentially makes everything in the UI available by CLI (and MCP).

Mostly this was an exercise in doing it just to see if I could - but it turned out to be a lot more interesting than I expected.

We've spent 10+ years focusing on having a clean, well designed interface for Ghost. It's something we care a lot about, and spend a lot of time on.

But within about ~1hr of using Ghost via Claude/CLI, it was hard to imagine going back to caveman-clicking around a browser to get something done. Particularly for complex or compound tasks that might require visiting several different areas of the app.

The experience is so different when you're just describing what you want to happen, out loud:

• "I saw this theme called [whatever], can you install that for me?"
• "give a complimentary subscription to [member name]"
• "how much traffic did my post get last week?"

Which is both obvious, and at the same time kind of jarring.

I know Ghost's UI extremely well, and know exactly where to go and what to click to do the thing I want – and even for me, using Claude is significantly faster/easier than clicking myself.

So how big would the delta be for a regular user who doesn't already know the UI inside out?

My initial thought was "huh, I wonder if UI even matters anymore?" - maybe everything just becomes a CLI / voice interface for a database, as various people have been suggesting about CRM tools.

But I don't think that's quite right. I notice when I interact with the product via Claude - I usually still keep the UI open, but my relationship to it is different. I use UI to see what happened, verify things look "right", and get an overview of what's going on.

Which is kind of familiar, because "agent does the actions for me / I review the results" has obvious parallels to AI coding workflows.

Before I'd be in VS Code all day doing the thing myself, but now I use Codex Desktop which is a UI designed entirely around optimising for: agent does the actions for me... I review the results.

Anyway, I don't know what my conclusion is here other than to say that AI+CLI is a really cool pattern, and I think it's likely to meaningfully influence what "UI" means over the next few years.

There are still tons of rough edges and reasons for why this is not yet a fully-formed paradigm (regular humans do not, and should not, ever need to know what "CLI" or "MCP" even means), but I like where it's going.

If you'd like to try out ghst - the beta announcement is here:

Developer Beta: ghst cli

Hi everyone, Wanted to share a new ghst CLI tool we’ve been working on - as a developer beta - which allows you to interact with Ghost publications from the command line. In short, this allows you (or an LLM like Claude, or Codex) to automate tasks within Ghost using a set of pre-built tools. Pretty much everything you can do in Ghost Admin, you can do using this CLI. For example: Create/edit posts and publish them Import or export members Download/upload/activate themes Find out which post…

Ghost ForumJohn

Why am I making an RSS reader? I don't know. There are a thousand RSS readers out there already, and they all read RSS feeds, but none of them work the way I'd like them to. So I'm making my own.

It's obviously a terrible idea. So why do it?

Mostly for fun.

There are all sorts of interesting things being published on the web every day, but they're kind of all over the place.

I enjoy sitting down on a Sunday morning with a cup of coffee and an hour to myself, but I don't enjoy opening my email inbox to read a newsletter, only to get sidetracked by 5 new emails that I should probably respond to. I don't enjoy wrestling my RSS app, which hasn't been updated in about 10 years and has a horrible reading experience. Reading links from Twitter gets me lost down an algorithm hole. YouTube doesn't even show me new videos from people I'm subscribed to.

I wish it were just all in one place. Without all the noise and engagement farming. Just a quiet little spot where I could catch up with things I care about.

That's what I'm making.

I'm calling it Alcove.

If you want to follow along with this questionable side project of undefined scope, I'm sharing live updates of progress on Twitter, here - or on Social Web platforms (Ghost, Mastodon, Flipboard, Threads etc) on @alcove@blog.alcove.news - or on Bluesky here.

There's a more than half-decent chance of a disastrous outcome.

We've now generated over $136M from free, open source software with Ghost

$36M in revenue from managed hosting
$100M in platform revenue for indie publishers

The ecosystem is growing really quickly at this point, and really excited about where things are heading next.

For a while it felt like the entire web was moving toward big centralized platforms — but as subscription businesses grow larger we consistently see that people want to run their own platform and have full control.

I'm extremely bullish on publishers using Ghost, but also the future of open, shared infrastructure more broadly.

It's taken 12 years to get to this point as a bootstrapped non-profit org with no outside funding, but the position we're in now is pretty remarkable:

Other platforms have hundreds of millions of dollars in VC debt to repay, and shareholders to enrich.

We don't.

We can just keep reinvesting 100% of what we make into improving the product.

That's going to create some really interesting asymmetric market dynamics over the next few years.

My biggest point of uncertainty about Ghost 6.0 was whether people were going to "get" the social web integration.

The technology is wonderful, but complex. For many people, terms like ActivityPub, Fediverse, bridge, protocol, server, toot, boost, and Webfinger are alienating and confusing. They subtly imply that unless you understand what all these words mean, this might not be the place for you; in the same way crypto terms—blockchain, web3, wallet, keypair, nonce—are a wall of jargon that scream "you don't belong here" to normal people.

The work of a product team, when working with new technology, is to abstract away as much of this complexity as possible, so that it feels friendly and approachable to new people.

To send an email, you don't need to know what SMTP, IMAP, POP, DKIM, SPF, or DMARC are. To browse the web, there's no requirement to understand HTTP, DNS, servers, SSL, TTL, load balancing, or caches. The most significant impact these protocols have is perhaps that users never have to think about them.

So while building the social web integration for Ghost, we weren't just reasoning about how to make it work and what it should do—we were thinking deeply about how to frame it. What words to use. What to compare it to. How to explain it. How to make it not need explaining at all.

Will people "get it"?

This question consumed more of my mental energy than anything else, right up until the moment we finally hit launch this past Monday.

My personal nightmare would have been if the response to the launch was another chorus of "I don't understand what the point of this is"—"this is too complicated"—"what does [x] even mean?" I've seen it happen so many times before when people try to figure out this tech and how it relates to their lives. The graveyard of technically superior but user-hostile products is vast.

But, I'm thrilled to see—at least so far—that hasn't been the case.

To be sure, there are still points of functional confusion. Chief among them: Why doesn't post X from platform Y show up on platform Z right away? But for the most part, I've been really encouraged by how many people have just jumped right in and started using it, without getting stuck and needing more explanation. They're just... publishing. And connecting. And it's working.

My strongest belief about the social web is that if we want it to succeed, we have to keep lowering the barrier to entry.

We have to keep minimizing the need for arcane language. We have to keep solving the things that people expect to work, but don't, rather than endlessly explaining how the underlying technology works. We have to create more familiarity with concepts people already know.

Let's not forget that email, as a technology, was based on the humble letter. To/from, subject, inbox, outbox—these were all words based on sending physical memos. The metaphor made the transition accessible.

The interface and format of a new technology can often be the single biggest factor in determining its adoption.

After all, for over a decade, we've had artificial intelligence capable of performing some pretty incredible tasks. The moment it really caught fire, though, was the moment it became a chatbox. Not when it got smarter. Not when it got more powerful. When it got simpler.

I think we've taken a big step in the right direction with the social web in Ghost 6.0.

And now we need to keep going.

Last week marked 12 years since the launch of Ghost on Kickstarter. I was 24 when I first started working on the idea, and back then I had no idea what I was doing.

Fast forward to 2025 and I still have no idea what I'm doing, but we've come a long way nevertheless. What started as little more than an idealistic open source pipedream has blossomed into a business with $8M/year in annual recurring revenue and a full-time team of 34.

The core idea behind Ghost, in 2013, was a simple one. I've always believed that independent publishing on the web is essential, and that the technology behind it plays a significant role in its ultimate success or failure.

Back then, Vox and Buzzfeed were at the forefront of a major shift in the publishing industry. Being technology startups as much as media companies, their proprietary software was a distribution lever to which the rest of the industry had no answer. They were miles ahead of traditional media orgs, many of whom had only recently been convinced of the need for a website.

It was as if a couple of outlets had built the world's first printing press, while everyone else was still sitting around writing out news bulletins by hand.

The ambition for Ghost was to create a dedicated open technology stack for independent publishers, from solo-creators to large newsrooms, to compete in a fast-moving digital media landscape – without becoming beholden to closed networks owned and controlled by private companies like Facebook and Medium.

A small band of loyal early adopters understood our ideas from the very start, but many were less convinced. They saw these large tech platforms as neutral, benevolent infrastructure providers, unlikely to screw anyone over, for it would not be in their interest to do so.

That outlook has shifted substantially after twelve years of algorithm changes, data capture, privacy scandals, and business model pivots.

So, as awareness and interest in independent technology have grown, so has interest in Ghost.

Over the years, we've focused consistently on building the best tools for publishing on the web. In the past 5 years, in particular, we've also focused heavily on building ways for creators, journalists, and publishers to run a sustainable business on the web.

We added paid subscriptions in Ghost a few years ago to allow publishers to charge for their work and, since then, our own growth has primarily been a side-effect of our work to help others grow.

In fact, we're about to pass $100Million in revenue earned by independent media businesses using Ghost.

Along the way, we've seen a fantastic group of publishers adopt the product. From hard-hitting journalism of 404Media and Platformer News, to political coverage from Tangle News and Lever News.

From The Harvard International Review to The Stanford Review, and JSTOR in academia.

From local journalism at Hell Gate, Kyiv Independent, Mill Media, Berkeley Scanner, The Dublin InQuirer, and The 51st, to the latest in music from Hearing Things and Drowned in Sound.

YCombinator runs on Ghost, as do First Round Review and The Diff, along with the media arms of Duolingo, WealthSimple, Kalshi, Cloudflare, DuckDuckGo, Domino's Pizza, and (fittingly) Kickstarter.

OpenAI, originally a research lab, first launched as a Ghost publication before it grew into the $300billion behemoth it is today. At one point, around the launch of ChatGPT, Ghost served more traffic for OpenAI each month than all our other publishers... combined.

I'm incredibly proud of how far we've come as a small, independent non-profit organisation shipping free, open source software.

As ever, we're competing with prominent, VC-funded platforms with tens of millions of dollars and hundreds of employees, playing zero-sum games to try and control the entire media industry.

Now, though, it feels like there's more appetite than ever for something different.

Incidentally, if indie publishing, open source, and JavaScript are your thing – we're hiring! Our team is fully remote, distributed globally, and passionate about building a healthy future for the open web.

Ghost

Ghost is a proud non-profit organisation building open source technology for fiercely independent, professional publishers.

Ghost

We're hard at work, and this summer we'll be shipping Ghost 6.0 – our next significant step forward in publishing. It's going to be a big one.

I'm reading Range by David Epstein at the moment, of which the below is an excerpt.

I can't help but draw parallels between the evolution of AI, 10+ years ago, and today – if you just switch chess:programming, and freestyle:vibe-coding.

It feels remarkably similar.

In a 1997 showdown billed as the final battle for supremacy between natural and artificial intelligence, IBM supercomputer Deep Blue defeated Garry Kasparov. Deep Blue evaluated two hundred million positions per second, That is a tiny fraction of possible chess positions-the number of possible game sequences is more than atomsin the observable universe—but plenty enough to beat the best human. According to Kasparov, "Today the free chess app on your mobile phone is stronger than me." He is not being rhetorical.

"Anything we can do, and we know how to do it, machines will do it better," he said at a recent lecture.

"If we can codify it, and pass it to computers, they will do it better." Still, losing to Deep Blue gave him an idea. In playing computers, he recognized what artificial intelligence scholars call Moravec's paradox: machines and humans frequently have opposite strengths and weaknesses.

There is a saying that "chess is 99 percent tactics." Tactics are short combinations of moves that players use to get an immediate advantage on the board. When players study all those patterns, they are mastering tactics. Bigger-picture planning in chess—-how to manage the little battles to win the war—is called strategy. As, Susan Polgar has written, "you can get a lot further by being very good in tactics"–that is, knowing a lot of patterns–"and have only a basic understanding of strategy."

Thanks to their calculation power, computers are tactically flawless compared to humans. Grandmasters predict the near future, but computers do it better. What if, Kasparov wondered, computer tactical prowess were combined with human big-picture, strategic thinking?

In 1998, he helped organize the first "advanced chess" tournament, in which each human player, including Kasparov himself, paired with a computer. Years of pattern study were obviated. The machine partner could handle tactics so the human could focus on strategy. It was like Tiger Woods facing off in a golf video game against the best gamers. His years of repetition would be neutralized, and the contest would shift to one of strategy rather than tactical execution. In chess, it changed the pecking order instantly. "Human creativity was even more paramount under these conditions, not less," according to Kasparov.

Kasparov settled for a 3-3 draw with a player he had trounced four games to zero just a month earlier in a traditional match.

"My advantage in calculating tactics had been nullified by the machine."

The primary benefit of years of experience with specialized training was outsourced, and in a contest where humans focused focused on strategy, he suddenly had peers.

A few years later, the first "freestyle chess" tournament was held. Teams could be made up of multiple humans and computers. The lifetime-of-specialized-practice advantage that had been diluted in advanced chess was obliterated in freestyle. A duo of amateur players with three normal computers not only destroyed Hydra, the best chess supercomputer, they also crushed teams of grandmasters using computers. Kasparov concluded that the humans on the winning team were the best at "coaching" multiple computers on what to examine, and then synthesizing that information for an overall strategy. Human/Computer combo teams— known as "centaurs" —were playing the highest level of chess ever seen.

If Deep Blue's victory over Kasparov signaled the transfer of chess power from humans to computers, the victory of centaurs over Hydra symbolized something more interesting still: humans empowered to do what they do best without the prerequisite of years of specialized pattern recognition.

In 2014, an Abu Dhabi-based chess site put up 520,00 in prize money for freestyle players to compete in a tournament that also included games in which chess programs played without human intervention. The winning team comprised four people and several computers. The captain and primary decision maker was Anson Williams, a British engineer with no official chess rating. His team-mate, Nelson Hernandez, told me,

"What people don't understand is that freestyle involves an integrated set of skills that in some cases have nothing to do with playing chess."

Range, by David Epstein

Why Generalists Triumph in a Specialized World

$9 on Amazon

I was digging through some old design files yesterday and came across two of the first ever sketches I made for the idea of memberships and subscriptions in Ghost, from way back at the very beginning of 2016.

These seem pretty obvious, now, but at the time nothing like this really existed. It's kind of amazing how close the sketches ended up being to what we actually built.

In 2016 Ghost was a nice, clean blogging platform that worked great, but it didn't do anything fundamentally valuable or useful compared to [all of the other platforms for writing on the internet out there].

We hypothesised that anonymous pageviews weren't particularly interesting, and that meaningful relationships with direct subscribers were. Blogs existed. Newsletters existed. What if the two combined? Ben Thompson was a big inspiration — what he was doing seemed like it was the future.

Now, why was I digging through old sketches? I was actually looking for another one that I made back in 2016 for an entirely different concept. One we're about to start working on this year. More on that later.

About 10 years ago I applied to Y Combinator with Ghost. At the time we were trying to get $120k in funding for 7% equity — but got rejected (twice). So instead we launched on Kickstarter and then bootstrapped.

Today Ghost has earned over $25Million from customers for 0% equity, and all of Y Combinator's own blogs run on Ghost.

Most of the time things don't work out quite the way you thought they would, but often they work out just fine.