IndieWeb
Great article on The New Yorker about The Rehearsal:
The first season of “The Rehearsal” seemed to exhaust all possibilities for the conceit that acting—the science of the artificial—could provide a prophylactic for life. It did not. The second season is, somehow, even more berserk than the first, but it’s also more disciplined and coherent.
Do not read this if you haven’t seen season 2. It spoils everything in the show.
John O’Nolan celebrates 12 years of Ghost. Congrats to the team! Amazing success, and the revenue is an inspiration for much smaller platforms like Micro.blog to aspire to:
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.
Nice guide for setting up Phosphor icons with a Micro.blog site, using the mnml theme.
Reminder that next week (Sunday) we’re starting a new photo challenge that will run through all of June. Follow @challenges for the prompts, which will be posted each day. 📷
The Rehearsal season 2 is a masterpiece. I can tell the final episode is going to be something I think about for a while, and again anytime I fly.
Gus Mueller blogs about being indie and Brent Simmons’s upcoming retirement. This is accurate:
Only crazy people are willing to put up with having to file business taxes, mess with social security, find healthcare, deal with all the stuff you have to handle to be indie. And you don’t even have to be a particularly good programmer. You just have to be persistent.
Also, congrats Brent! 🎉
Simon Willison has a fantastic, detailed look through the Claude 4 system prompt. Personality, safety, tools, and artifacts.
Enjoyed Mission: Impossible. A couple minor nitpicks, but I liked the throwbacks to earlier movies. Crazy that it’s been a nearly 30-year series. 🍿
We rewatched Star Wars episode 4 last night, with frequent pausing to think about the context with Rogue One and check Wookieepedia for random details. So fun. Also lots of nitpicking about the special edition changes.
Glitch announced they are shutting down hosting. Kudos to them for allowing redirects after everything winds down. That’s more than most companies do:
In the coming days, your dashboard will get a new feature to set up redirects for your project subdomains, so all your links will keep working. Make sure your redirects are set up before December 31, 2025. (We’ll make sure they stay active at least through the end of 2026.)
Built up the courage to run Redis’s MEMORY PURGE in production this morning. Sadly made no difference. This week I’ve been running a cleanup script that has trimmed nearly 10 GB of memory usage, but still lots reserved by the system.
With Pocket shutting down, I’ve updated Micro.blog’s bookmarks import for Pocket’s latest export format. Also supports Instapaper, Raindrop, and Pinboard. Happy bookmarking! 🔖
Stephen F. Austin State Park. Short hike on a trail near the Brazos River.
Write.as creator Matt Baer wants to focus more on writing for 2025. I can relate to a lot of this:
Our apps have always been built to help you get your words down with nothing in the way. That’s why Write.as always opens to our editor — so you can start writing immediately, and not get distracted by notifications, comments, and superfluous things like how many “likes” you received.
Stephen Hackett throwing some cold water on the io products hype:
To be clear, the failure of the Rabbit R1 and Humane AI Pin does not mean that there’s no room in the market for an AI-powered device with no screen. However, people really like their phones, and creating a product that will compete with the smartphone is a hill no one has successfully climbed to date.
I don’t agree that the Rabbit R1 was a failure. It fell short of expectations, but that team is still churning away. A new kind of AI-first device is still more likely to come from someone new, without the smartphone baggage.
You might’ve noticed there’s a new status line in Micro.blog’s publishing progress, providing a little more context for what the platform is doing. I’ll continue to tweak this so the progress is more useful.
Almost lost in all the OpenAI hype this week was that MCP is now supported in OpenAI’s responses API. MCP was a big part of Microsoft’s keynote and at Google I/O.
Om Malik blogging about today’s video of Jony Ive and Sam Altman:
The slick video harkens back to Ive’s glory days at Apple when he would talk about the chips, designs, and aluminum on videos extolling the iPhone, the watch, and the laptops. In a way, what he and Altman are indicating, through words, and subliminal marketing, is that we are building the next Apple.
What stood out to me in the announcement about io is that there will be a “family” of devices. Maybe one without a screen, a couple with small screens? I assume voice will be a big part of this, but also perhaps not. It could truly surprise us.
Finished watching Andor. Loved it. Now watching Rogue One. As expected, flows perfectly together with the series.
Big update out for Tapestry today, adding bookmarking for Micro.blog accounts among other features.
This is a lengthy post from Ben Werdmuller, but everyone who cares about a sustainable, vibrant ecosystem of both free and commercial products for the social web should queue it up to read. He goes over different types of funding and a lot more. Really good.
Anil Dash on the growing popularity of MCP and adopting standards:
It’s cool that other platforms adopted the same spec that Anthropic made for their system. There’s a generosity of spirit to a technology platform choosing to be the second to adopt a protocol, if they do it in a faithful fashion.
This was echoed in the interview with Microsoft’s Kevin Scott this week. Some people at Microsoft would’ve designed MCP differently, but that doesn’t matter. Just roll with it, and the compatibility across platforms makes up for any shortcomings.
John Voorhees blogging about Apple’s AI predicament:
Hardware plays to Apple’s design and supply chain strengths. In contrast, the rapid iteration of AI models and apps is the antithesis of Apple’s annual OS cycle.
I agree. A year is a lifetime right now.
Some of the news out of Microsoft Build is what developers hope Apple will do: access to AI models in the cloud, on devices, or even from web apps inside Edge. There’s also MCP everywhere. I’m reading through NLWeb, created by RV Guha, whose semantic web work for decades seems a nice fit AI search.
If you use Chrome or Arc with Micro.blog, check out the latest update to our Chrome web extension for bookmarking web pages. It improves on the last version, now better saving HTML to archive the page. Still working on making it compatible with Safari and Firefox.
Not gonna lie, I’m close to dropping $200 to try OpenAI’s Codex. But I don’t think most of my code is well suited to it. Not enough automated tests! It figures that would come back to bite me.
I mostly use AI as a machine that can generate unlimited example code. I learn best from editing examples.
Working on server memory usage, discovered a massive leak of Redis keys we use to rate-limit clients hammering the server. Hope I can reclaim tons of memory by fixing it.