Over the weekend, Automattic/Beeper had the pleasure of hosting and sponsoring the Build the Future MCP Hackathon in New York City organized by Adam Anzuoni, alongside Anthropic, Stainless, WorkOS, Smithery, and MongoDB. As someone deeply passionate about open-source technology and the potential of AI, I was thrilled to not only support this event but also to speak at it and serve as one of the judges. For those unfamiliar, MCP stands for Model Context Protocol and enables AI models to interact with multiple tools in a more efficient and structured manner, allowing developers to create sophisticated agents that can handle complex tasks. And it’s only 8 months old! The energy in the room was electric—hackers collaborating, iterating, and demoing groundbreaking projects in just a few hours. (I would love to see the results of a longer overnight or all-weekend hackathon.)
For a wrap-up of the winners, check out this great X/Twitter thread by Alex Reibman. The first-place winner, Levels AI, orchestrates different AI/ML models to solve specific business requirements without needing heavy manual coding or specialized teams. Essentially, it’s models creating more models—a meta-approach to AI that could streamline operations for companies of all sizes. Yusuf Olokoba said the name was inspired by Pieter Levels, which is pretty cool. I wanted to note one team in particular: BeepResearch, which used our early access Beeper API and local MCP.
6/ BeepResearch (Bonus)
MCP server to chat with your messaging apps like iMessage, Whatsapp, FB, X, etc. over @beeper. And Ollama + gemma3 to query photos in natural language
Judging the entries was tough! People built so much cool stuff in a really short period of time. I always say technology is best when it brings people together, and this was a great example of that, and makes me very excited for Contributor Day at WordCamp US next week. (Which has such a stacked AI program as well.) Hopefully some people who enjoyed the office might consider getting 24/7 access.
I also just want to take a beat and say how amazing it is that I’m blogging this, on a United flight from Houston to San Francisco I’ve taken a million times, on a plane 36,000 feet in the air and technology is amazing. I can’t wait until United has Starlink!
Scott Alexander, of Slate Star Codex / Astral Codex Ten fame, ran an Everything-Except-Book Review Contest 2025 in February. The prompt: “Submit an ACX-length post reviewing something, anything, except a book.” The submissions were collected anonymously in a giant 450-page Google Doc. I don’t think the winners have been chosen yet, but there is one essay that has been making the rounds and getting shared more, and that’s Alpha School and “2-hour Learning” powered by AI, a parent’s perspective on Alpha School, a set of “AI-powered” schools in Texas and Florida.
It’s worth reading the entire essay, but I wanted to excerpt a few points I found interesting:
After twelve months I’m persuaded that Alpha is doing something remarkable—but that almost everyone, including Alpha’s own copywriting team, is describing it wrong:
It isn’tgenuine two‑hour learning: most kids start school at 8:30am, start working on the “two-hour platform” sometime between 9am-930am and are occupied with academics until noon-1230pm. They also blend in “surges” from time to time to squeeze in more hours on the platform.
It isn’tAI in the way we have been thinking about it since the “Attention is all you need” paper. There is no “generative AI” powered by OpenAI, Gemini or Claude in the platform the kids use – it is closer to “turbocharged spreadsheet checklist with a spaced‑repetition algorithm”
It definitely isn’tteacher‑free: Teachers have been rebranded “guides”, and while their workload is different than a traditional school, they are very important – and both the quantity and quality are much higher than traditional schools.
The bundle matters: it’s not just the learning platform on its own. A big part of the product’s success is how the school has set up student incentives and the culture they have built to make everything work together
…Yet the core claim survives: Since they started in October my children have been marching through and mastering material roughly three times faster than their age‑matched peers (and their own speed prior to the program).
One of the surprises doesn’t come until Part 4 of the essay:
Incentives
People REALLY don’t like the idea of incentivizing kids to learn.
Roland Fryer, who has done extensive work on what works in incentivizing students, quotes a 2010 Gallup poll that found that only 23% of American parents support the “idea of school districts paying small amount of money to students to, for example, read books, attend school or to get good grades” (76% opposed the idea with only 1% undecided).
In 2005, being a remote-first company was anathema to investors and business leaders* at the time; it was a scarlet letter that combined with our embrace of Open Source and the relative inexperience got us some funny looks and a lot of skepticism. I will be forever grateful to the true contrarians who bet on Automattic in our earliest days.
Even when it was clearly working the first few years, there was always the dismissal of “that won’t scale” that loomed like a remote startup Great Filter. These days I hear from friends who run incubators or do seed investing that almost every company they look at taps into remote talent.
It makes me think about what uncommon things Automattic does today that will be standard in the coming decades. We do our best to balance idealism with pragmatism, because even if you are on the right side of history, being too early can be as bad as being wrong.
I can’t predict everything that will change over the coming decades, especially with AI making the next few years particularly hard to predict. Still, I do know a few things that won’t change: everything flows from our people, open source is still the most powerful idea of our generation, growth is the best feedback loop, and no matter how far away the goal is, the only way to get there is by putting one foot in front of another every day. People will always want fast, bug-free software; instant, omniscient customer service when they need it; and experiences so intuitive that they usually don’t. And once they’ve had a taste of freedom, it’s hard to return to their previous state. (For more, see our creed.)
Our industry is highly cyclical, and I feel fortunate to have gained the perspective of a few bubbles and crashes, along with all the emotions that go with them. It’s undeniable we’re in the very early days — the command-line times — of an AI era, and though it will probably have its own bubble and crash cycles, it feels as significant to me as anything since we started. It’s more important than ever that we fight for open source and the freedom-enhancing side of technology. I’m committed to doing whatever I can to democratize publishing, commerce, and messaging, but there are many other areas of the human experience to cover… pick one to work on! It’s hard and rewarding work.
When I was working on an early version of one of our internal stats systems, it was really important to me that it showed rolling windows of the last 24 hours (daily), 168 hours, 4 weeks, and of course yearly. The rolling was important so you could see the impact of your changes as soon as possible. Then I felt called to add another: decade.
Some thought it was silly at the time, and it’s true that it initially served mainly as a way to display the cumulative number. But I wanted every time someone looked at one of these stats pages that they were reminded that we’re building for the long term. Our users and customers deserve nothing less. And now we have some statistics with 20 years of history, it has some useful comparisons as well!
There’s a lot more to do, and I can’t wait to see what a “20 Years of Automattic” post says. I’m a lucky guy.
Now we know! I’m still a very lucky guy, and can’t wait to build, learn, and share alongside a talented crew of like-minded hackers, dreamers, and doers.
* I’ll note that pioneers like Bob Young (Red Hat), Stephen Wolfram (Wolfram Research), Jason Fried (37Signals), and Mårten Mickos (MySQL) were big inspirations. Also, the entire Open Source community and most projects operated at least partially this way, which is why it seemed so natural to us as a second-generation Open Source company.
Gravatar has always been about giving people control over their identity online. One avatar, one profile, synced across the web, verified connections, with a fully open API.
Gravatar is a true open identity layer for the internet, and now for AI.
For developers, we’ve rolled out mobile SDKs and a revamped REST API that lets you fetch avatars and profile data with just an email hash. Whether you’re building a blog, a community, or an AI agent that needs to understand who it’s talking to, Gravatar provides the infrastructure to make identity seamless and user-centric.
It’s free, open, and built with developers in mind. We believe identity should belong to the individual, not be locked behind proprietary platforms. Gravatar is our contribution to that vision.
If you haven’t checked it out lately, now’s a great time to explore what Gravatar can do for your app or your online presence. And think about how your apps can drive more Gravatar signups.
I’ve checked off a bucket list item: I’m attending a Berkshire Hathaway shareholder meeting. It’s really an event! Thousands flock to Omaha, Nebraska, for the legendary Q&A sessions with Warren Buffett and shareholder deals. They’ve made it quite the circus, with every Berkshire Hathaway company having a booth of some sort, and typically selling their goods at a discount or with exclusive items you can only buy there, like Warren Buffett and Charlie Munger Squishmallows (which of course I got, to complement my bronze busts).
It’s strange to have a Dairy Queen booth selling $1 ice cream (cash only!) next to NetJets, but those juxtapositions are part of the Berkshire vibe—it’s very high/low, like Costco (a big Berkshire holding). There’s also an element of WordCamps or a Salesforce Trailblazer event in that you can tell there’s a “type” of person that’s easy to spot who’s a Berkshire enthusiast. A lot of Berkshire brands are also WordPress users: Duracell, GEICO, Acme Brick, Berxi, MiTek. I think there is a lot of mimetic overlap between the values of open source and the values of building a Berkshire company.
As with any big gathering, the side events are also great, and I was honored to have a fireside chat with a friend and Buffett protégé, Tracy Britt Cool. To an audience of about 60+ CEOs in the Kanbrick community, we talked about Automattic’s history and some of the latest happenings in tech; AI was definitely on people’s minds in the Q&A. They had questions for me, but I also feel like I have a ton to learn from this group that has built founder or family-owned businesses with an average of 80-100M of revenue, the kind of thing that is the engine of the American economy.
It makes me pine for the day when we can have more shareholders in Automattic; I think it would be an amazing cohort of folks that believe in open source and the open web, invested together and learning from each other, and I could imagine an event very much like these shareholder meetings. It’s so much more powerful when you build a business where your customers are also a community.
Update: I knew this would be a special one because it was Warren’s 60th, but he really went above and beyond by announcing his intention for Greg Abel to take over as CEO at the end of the year. The standing ovation was a special moment, 60 years of 19.9% compounding returns! I think the future of Berkshire is very bright because he’s shared so much of his worldview that there are others that have made it their own.
It’s a funny time in WordPress because there are a lot of really interesting open questions:
Can we iterate faster with canonical plugins?
What’s the fun thing we can put in to celebrate 7.0, and when will that be? (I was rooting for real-time co-editing like Notion/Canva/Google Docs.)
How can we use AI to automate our manual work around WordPress.org?
Can AI help us make 60k+ open source plugins and themes in the directory more secure? (I think so.)
What should we do with our 13k issue backlog? (That’s a lot of bug gardening.)
How will AI change how people build and update sites?
Just like RSS and web standards supercharged WordPress for the podcasting and search revolutions, what standards or APIs can we ship to help 40%+ of the web work with AI agents? (Plus an entire rabbit hole of all the new sloppy crawlers using so many resources.)
Some of these broad changes are mixed. At one point, I used Google to search for things and would visit their top result, which is great for website owners. Nowadays, Google pulls almost everything I need into the results page, so I don’t see as many random sites. But on Perplexity, sometimes I’ll read the answer and then visit 4-5 of the sources it cites to learn more, so I’m visiting 4-5x more random websites, usually powered by WordPress, than I would have even in the early days of Google. We don’t know how this all plays out yet.
These questions are also against the backdrop of some of the brightest minds in WordPress spending more time with legal code than computer code, which could last until 2027 or longer with appeals.
Speaking for myself, I was in my first deposition today. I really appreciated the due process and decorum of the rule of law, and just like code, law has a million little quirks, global variables, loaded libraries, and esoteric terminology. But wow, after a full day of that, I’m mentally exhausted. Hence, I’m posting about 6.8 after it’s had 6 million downloads. I’m more impressed than ever by what smart lawyers do, and the entire thing, though sometimes imperfect and frustrating, is a blessing to our democracy. However, I can’t wait to return to spending the plurality of my days with engineers and designers again. I’m sure many other folks in the WordPress community would agree.
The long-anticipated “Big Sky” AI site builder on WordPress.com went live today. It combines several models and can create logos, site designs, typography, color schemes, and content. It’s an entirely new way to interact with and edit a brand-new or existing WordPress site. This AI agent will make WordPress accessible to an entirely new generation and class of customers, and it will be a power tool for professionals to do things in minutes that used to take them hours.
(This post should be read while listening to Wish by Joshua Redman. The writing is synchronized to the music reading speed.)
Contributor day just wrapped up for Portland for WordCamp US. If you ever have a chance to visit a WordCamp, I recommend it. It’s an amazing group of people brought together by this crazy idea that by working together regardless of our differences or where we came from or what school we went to we can be united by a simple yet groundbreaking idea: that software can give you more Freedom. Freedom to hack, freedom to charge, freedom to break it, freedom to do things I disagree with, freedom to experiment, freedom to be yourself, freedom expressed across the entire range of the human condition.
Open Source, once ridiculed and attacked by the professional classes, has taken over as an intellectual and moral movement. Its followers are legion within every major tech company. Yet, even now, false prophets like Meta are trying to co-opt it. Llama, its “open source” AI model, is free to use—at least until “monthly active users of the products or services made available by or for Licensee, or Licensee’s affiliates, is greater than 700 million monthly active users in the preceding calendar month.” Seriously.
Excuse me? Is that registered users? Visitors to WordPress-powered sites? (Which number in the billions.) That’s like if the US Government said you had freedom of speech until you made over 50 grand in the preceding calendar year, at which point your First Amendment rights were revoked. No! That’s not Open Source. That’s not freedom.
I believe Meta should have the right to set their terms—they’re smart business, and an amazing deal for users of Llama—but don’t pretend Llama is Open Source when it doesn’t actually increase humanity’s freedom. It’s a proprietary license, issued at Meta’s discretion and whim. If you use it, you’re effectively a vassal state of Meta.
When corporations disingenuously claim to be “open source” for marketing purposes, it’s a clear sign that Open Source is winning.
Actual Open Source licenses are the law that guarantees freedom, the bulwark against authoritarianism. But what makes Open Source work isn’t the law, it’s the ethos. It’s the social mores. It’s what I’m now calling Ecosystem Thinking: the mindset that separates any old software with an open source license from the software that’s alive, that’s humming with activity and contributions from a thousand places.
Ecosystem Thinking has four parts:
Learn
Evolve
Teach
Nourish
Learn is about keeping ourselves in a beginner’s mind, the curiosity to always engage with new ideas and approaches.
Evolve is where we apply those learnings to our next iteration, our next version. We see how things work in the real world: it’s the natural selection of actual usage.
Teach is actually where we learn even more, because you don’t really know something until you teach it. We open source our knowledge by sharing what we’ve learned, so others can follow on the same path.
Nourish is the trickiest, and most important part: it’s where we water the garden. If you’ve done the previous three steps, you’ve been very successful; now your responsibility is to spread the fruits of your labors around the ecosystem so that everyone can succeed together. This is the philosophy behind Five For the Future, which you’re going to see us emphasize a lot more now.
That’s the ecosystem. But if it’s the yin, what’s the yang? This openness and generosity will attract parasitic entities that just want to feed off the host without giving anything back. There are companies that participate in the Learn/Evolve/Teach/Nourish loop like a FernGully rainforest, and there are those who treat Open Source simply as a resource to extract from its natural surroundings, like oil from the ground.
Compare the Five For the Future pages from Automattic and WP Engine, two companies that are roughly the same size with revenue in the ballpark of half a billion. These pledges are just a proxy and aren’t perfectly accurate, but as I write this, Automattic has 3,786 hours per week (not even counting me!), and WP Engine has 47 hours. WP Engine has good people, some of whom are listed on that page, but the company is controlled by Silver Lake, a private equity firm with $102 billion in assets under management. Silver Lake doesn’t give a dang about your Open Source ideals. It just wants a return on capital.
So it’s at this point that I ask everyone in the WordPress community to vote with your wallet. Who are you giving your money to? Someone who’s going to nourish the ecosystem, or someone who’s going to frack every bit of value out of it until it withers? Newfold, especially since its acquisition of Yoast and Yith, gives back. (I’ve asked them to consolidate their Five for the Future pages to better represent the breadth of their contributions.) So does Awesome Motive, 10up, Godaddy, Hostinger, even Google. Think about that next time it comes up to renew your hosting or domain, weigh your dollars towards companies that give back more, because you’ll get back more, too. Freedom isn’t free.
Those of us who are makers, who create the source, need to be wary of those who would take our creations and squeeze out the juice. They’re grifters who will hop onto the next fad, but we’re trying to build something big here, something long term—something that lasts for generations.
I may screw up along the way, or my health may falter, but these principles and beliefs will stand strong, because they represent the core tenet of our community: the idea that what we create together is bigger than any one person.
First, tech companies are typically best at adopting new technology, which leads to productivity gains.
AI may be an obvious example of this, though for all its hype it hasn’t had a huge impact on most companies yet. I agree with Sam Altman when he says there may someday be a billion-dollar company run by one person who is able to highly leverage future AI agents to automate most traditional roles at a company. That said, I think there are advantages to teams including allowing people to go on vacation or take time off, and provide business continuity and succession, so literally one-person is probably an exaggeration. We don’t need AI to see very small teams being valued highly: Instagram had only 13 employees when it sold for a billion dollars to Facebook, in 2012!
Some of this productivity gains just come from adoption of existing tools like Google Workspace or Office 365, issue trackers and version control with tools like Gitlab, Github, or Jira. At Automattic we don’t use email to work or communicate internally, it’s all Slack and P2. We also leverage our distributed nature to effectively have teams around the world coordinating several shifts of product work per day, and 24/7 coverage for things like systems and customer support without the need for “graveyard shifts.”
The way tech companies operate, the pace and culture, would be unrecognizable to people at many more traditional companies.
At tech companies some roles are highly leveraged, like systems, engineering, and design, and everything else in the company really exists to support these. These leveraged roles can create enormous amounts of value, and that’s why it’s not unusual to hear of machine learning engineers working on ads at Google with salaries in the seven figures. (There’s been a weird accounting thing where companies put a lot of their compensation into equity, but I think that’s going away as investors are learning to better account for dilution and employees appreciate the fungibility of cash.)
Creators are also highly leveraged, which is why Joe Rogan can sign a new $250M deal with Spotify (which smartly puts him back on Youtube) after laying off 1,500 people in December. Some people like Hagen Terschüren try to tie this together and say you should avoid Spotify for it, but there’s nothing wrong with a business becoming more efficient to serve its customers, it’s the whole point of capitalism. Capitalism is, as Nicholas Stern says (via Marc), the best way to take care of people we don’t know. There’s no honor in keeping people employed inefficiently, it’s better for them to find someplace in the market where their talents will be better leveraged for society and themselves.
There was a bubble in hiring because tech had so much money it tried to throw people at problems. But the unlock in technology can come from a single person, a single insight. It’s the mythical man-month. Tech-first companies are going to become leaner and more leveraged. Fewer people are going to create more value for society, in ways that will follow power laws and I think we should investigate things like Universal Basic Income to provide for all living beings. Technological progress creates abundance, where we have more than what we need.
At Automattic last year we did not do layoffs, but allowed performance management and natural attrition (voluntary regrettable was 2.9%, non-regrettable 6.8% for us in 2023) to allow our size to shrink down more naturally, on average two people left for every person we hired last year, from a peak of about 2,064 to 1,936 today.
Google News has ‘officially’ launched, insofar that it is now a tab on the now-crowded Google toolbar (not toolbar, I guess tab-bar). Although it’s still labeled as beta, those of you who remember what it used to look like know that it’s come quite a way. It seems to have some sort of algorithm running, so check it out. If you’re interested in seeing what else is cooking at Google, check out Google Labs, which currently has such goodies as a keyboard navigable interface, a glossary, something fancy with sets, and an interface that will return results based on your voice commands.
