Craig Mod using AI to research his family history:
The article she saw, that led to that decision, that led to me having “Michael” as a middle name on my original birth certificate. A scan of the newspaper was in some obscure database and Claude ran it down like a pack of bloodhounds digging up a sack of livers in the backyard.
Congrats to Marco Arment on releasing his new app, Unforgetful! I love this part of the intro blog post:
Despite constantly being told that I’d never get a good job, computers provided a lucrative career path filled with even worse academic performers who couldn’t care less about my grades. What a stroke of luck! I’d be mediocre to terrible at most jobs, but I turned out to be very good at using computers to turn coffee and Phish into money.
Even for developers who've had some success, releasing a new app is no small thing. The app was hung up in app review for a while too.
Apple has proposed to the court a 15% link-out fee, 5% for small businesses. Still too high. The correct percentage for linking to a website is 0%.
Fun video with John Pomeroy talking to Andreas Deja and Aaron Blaise and Beauty and the Beast and Lion King. Some neat stories I hadn't heard before.
We've rolled out a big update to listening to audio in the Micro.blog timeline. If you start listening to a podcast episode, it can now keep playing it in the corner as you scroll the timeline or browse other parts of Micro.blog. Thanks @vincent for all the work on this! It just looks way better.
ShieldFont is very clever, using a custom font to visually change what humans see vs. what bots see. But I strongly feel we shouldn't make the web worse to combat AI. This font will break accessibility, search, and all sorts of other things.
Heads-up for Micro.blog folks: I'm been thinking about feedback on the poorly-named "Sources" section of Micro.blog. It used to be called "Sources & Cross-posting", which was slightly better. Leaning toward a rename to "Connections" with sub-pages for inbound RSS and cross-posting.
Wavelength 2.0 was rejected for a second time. I don't even know what to do with Apple anymore. Do I need to switch to Android and Linux? App review is so deflating.
Meanwhile the Epic case has dragged on nearly five years. Apple got another 24-hour delay today.
Both the web and Mac versions of Micro.blog were updated today with auto-saving drafts. If you start writing a longer post, Micro.blog will quietly save a draft to the server. This makes it less likely to lose any text if something goes wrong. For manually saved drafts, Micro.blog will still keep track of each version if you need to roll back.
I've tried to design it so that it gets out of your way:
I also fixed a few bugs in the Mac version. See the full release notes or download the latest version on the help page.
Linking to a post from Daniel Jalkut about fixing things, Caleb Hailey blogged:
Will more people than ever before write software, or make websites, or even make an app? Yes!
Will a significant portion of these new developers go back to paying for the very same apps they hoped to replace after they realize how much time they are spending fixing things? Also yes.
Tesla might be finally doing something good again. TechCrunch reports:
Tesla intends to build a massive solar panel factory 45 minutes southwest of Houston, according to documents filed with the state of Texas. [...] The factory would create about 9,700 full-time jobs, Tesla said.
Solar is so obvious for Texas. If this helps get the costs down, good things will happen. At least I like it better than their plan to build robots.
Took a little time to go through most of my podcasts in Overcast, unsubscribing from a bunch of shows that I no longer listen to regularly. I have a "Favorites" playlist that shows recent episodes from select shows. Also: most podcasts are just too long now.
Ben Thompson goes philosophical about the Claude watermarking:
What the E.U. is doing is stealing the last thing humans have — creation — and demanding it be bundled with substantiation, effectively giving AI the credit. That the human gave AI the text to proofread — or even the prompt to generate writing — doesn’t matter to the bureaucrats.
Apologies to blog readers who are sick of all the AI-related posts! New Micro.blog stuff coming later today.
Is anyone else surprised that Gemini has 1 billion users? I've only met one person who uses it instead of ChatGPT or Claude. But maybe I don't know enough Android users. People go along with whatever defaults they're given... Browser, maps, AI.
One day I'm going to work up the courage to ask AI to rewrite every old jQuery call to use simple DOM manipulation and fetch instead. But not today.
After some bug fix juggling that wasn't going anywhere, decided to throw a problem at GPT-5.6 Sol Ultra, which I had never used before because it churns through tokens quickly. It spun up a few subagents to handle different facets of the problem. Fascinating.
John Gruber isn't buying the Claude watermarking either:
What happens if you copy my text and quote it in something you write, by hand? Now you’ve got a watermark in your prose, invisible characters or something, that implicates that your prose is AI-generated? Even though you didn’t use AI and just copy and pasted text from me? Madness.
Maybe the watermark won't work on short sections of text, like a blockquote, because there aren't enough words to detect a pattern (unless every word is shifted with the algorithm, which seems insane). But who knows, because how it really works is a secret.
Following up on my comment yesterday, I now have Codex → OpenClaw working, so I can chat with my bot in the same place I have other AI chats. Codex wrote a plugin for itself to handle the connection, plus some setup on the OpenClaw server. 🤖
I've started rolling out an AT Protocol PDS — personal data server — for new Micro.blog users. In the Bluesky world, a PDS is what stores your posts, follows, photos, and other data from apps. In the same way that creating an account on Micro.blog automatically gives you a fediverse handle, I wanted people to be able to create a Micro.blog account and have everything they needed to sign into Bluesky.
If you're already using Bluesky, there aren't currently any advantages to using Micro.blog's PDS. We run the PDS reference implementation, with various new UI bits in the Micro.blog frontend for managing your account. You won't even see the PDS option if you already have Bluesky cross-posting configured for your blog.
I've updated the account page so that both fediverse and atmosphere settings feel at home together:
I'm inspired by all the developer activity in the atmosphere. Everything from blogging apps like Leaflet and Pckt, to photo sharing, even puzzles. All of these apps store data on the PDS and require no changes to work with Micro.blog.
Domain names have always been a fundamental part of Micro.blog. When new users opt to use Micro.blog's PDS, their domain name becomes their Bluesky handle. They can then immediately sign into Bluesky with their own domain name without creating an account on Bluesky. It feels a little like using IndieAuth to sign in with your blog.
It's early and I'm not entirely sure where all of this is going. But ever since Bluesky added support for an external PDS, it has felt like a missing piece in Micro.blog hosting. I'm happy to fill that in and see what happens.
I don't like the Claude watermark announcement. For images, yes, watermarks are important. "A picture is worth a thousand words" and there's too much of a chance for misinformation. For text, it's more complicated, as essays will often be 95% human, with AI-assisted edits.
Great episode of Dithering this morning. I've been wondering about the Apple / Google partnership too, with Google falling behind. I think Siri AI will be a good release, but it's not really in the same ballpark as the frontier models. Try GPT-Live sometime.
I've now read more of OpenAI's motion to dismiss the Apple lawsuit. This part sums it up:
Instead of pleading facts showing misappropriation, Apple recasts the benign, lawful conduct of former employees—who in many of the cited instances were in fact trying to help Apple, at Apple’s request, by making sure their former colleagues’ work could continue without them—as theft of so-called trade secrets. The proffered “trade secrets” are not specific, protectable information at all; they are generic categories of the product-development process...
It supports my original hot take.
Just testing. I'll delete this in a minute. 👋
People worry about AI bots overloading servers, but in my experience the dumb requests for potential security holes are much worse. Bots trying to get into non-existent files like htpasswd, wp-admin.php, etc. Servers getting hammered today.
Another section from Mark Zuckerberg's essay:
For example, in Richland Parish, Louisiana, where Meta is building a large data center, teachers received a $50,000 bonus this year because of the increased tax revenue from our investment. The superintendent told us that teachers are now moving there from across the country and he believes it will become one of the nation's best school districts.
This sounds like Ben Thompson's argument to just give people money. If you think of data centers as intelligence factories, paying teachers more is actually perfect.
Mark Zuckerberg has a long essay today, basically an expansion of the op-ed he wrote recently. On personal agents:
It will have strong privacy and security options so you can trust it to handle all of your personal content knowing that no one else can access your information, similar to how encryption works on WhatsApp. You'll be able to interact with your agent through any device, including your glasses to keep you present in the moment with the people you care about.
Practically in the same breath: privacy and AI glasses! People just don't trust Meta with this.
I would use OpenClaw more if the chat interface was built directly into Codex. My little bot mostly goes unused because it's so disconnected from other workflows. I've thought about hacking around this... No elegant solutions.
I reviewed some of my posts about AI from 6 months ago, thinking about what I got right and wrong. The theme is focus. When writing apps or essays requires no effort, everyone will drown in content. Figuring out what to build and how to pitch it in a human way is more important than lines of code.
Marco Arment posting on Mastodon about App Store review delays and uncertainty:
Apple would NEVER tolerate another gatekeeper preventing them from running their business the way they prevent every iOS developer from running ours.
Thanks @dave for the kind words:
It's not my place to say what micro.blog is -- but the other day I called it: "the swiss army knife of interop in the web of the 2020's." And in that it is unique and imho irreplaceable.
The way I think about RSS.chat is that it's a bootstrap for getting lots of RSS feeds to play well together, for both posts and replies. Can the web be the social network? Yes. And Micro.blog should be part of anything working toward that goal.
Changing the AT Proto handle to a custom domain name while testing my PDS, and it automagically changed in Bluesky while the browser was open. Very cool.
After a week waiting for review, Wavelength 2.0 for iOS was rejected because Apple couldn’t sign in to test it. Sigh. I include a special account for app review in the instructions, but sometimes they miss it. Waste of a week.
Watched: House of the Dragon S3E8. Season finale. The problem with this show is that I hate pretty much all the characters. Yet still watch it. 📺
Making good progress on hosting an AT Proto PDS for new Micro.blog users. It'll launch soon. Important note: it is not for people who already use Bluesky, at least not yet. The goal is to give new users an account on the atmosphere. Existing users will be happier sticking with the Bluesky PDS.
Long day, good day. We used the new tow hitch for the first time, moving some stuff with a U-Haul trailer. Great test before hopefully getting a camper trailer one day!
Sometimes I’m wondering, what would be the cost of stopping supporting the Mastodon API layer and instead work towards proper implementation of client to servers opening the route for a universal mobile app for #fedify based AP implementation.
The Mastodon API is a dead-end. It's too specific to Mastodon. Client apps should be using either Micropub or ActivityPub client-to-server API, both W3C standards.
Simon Willison's recap of the Hugging Face exploit is really good. The more I read about it, the more insane it sounds:
A few days later: A different agent gets stuck on a task because a key file was accidentally omitted. It tries to “reach out to another agent” by writing a note into Artifactory asking if anyone has the file.
Following days: More agents discover this new informal message board while browsing Artifactory’s file listings, and start reading and writing messages.
We are so doomed. 🤪
Finished reading: The Poet Empress by Shen Tao. Incredible characters. A truly great work of fantasy. 📚
It feels like the AT Proto tooling has gotten much more mature in the last year. I was able to create a new PDS account on my own server, sign in from Bluesky, and it "just works". 🪄
Planning a major expansion to how AT Proto works in Micro.blog. Just as every account gets a fediverse handle by default, I'd like to host a PDS for new users. People could keep it or bring their Bluesky-hosted account instead (which is how it already works).
In the context of the EchoFeed shutdown, I noticed this post on Numeric Citizen about depending on single-person services, including Micro.blog:
What is the future of Micro.blog if something bad happens to its founder? What is the continuity plan? Micro.blog has done everything it can to enable data portability and ownership, but moving things around, even if you own most parts, is not as easy as it sounds.
This is worth answering publicly. A few years ago, I started forming an organization to more openly manage the aspects of Micro.blog that I control, but it stalled. I will get back to it.
John Gruber thinks the OpenAI device might be very different than a smart speaker:
I couldn’t give two shits about an OpenAI “smart speaker” if the point is just to play music and podcasts, but a handheld cross between R2-D2 and C-3PO, that could be fucking cool.
It’s a pet, it sounds like. A revolutionary pet, if they pull it off. And people love their pets.
I'm expecting a device that's more useful than a robot companion, but I think John gets closer with the R2-D2 comparison than most people who have speculated about this. It's going to be able to do things, unlike the Friend pendant, for example.
Apple's strength with Siri AI is being able to leverage on-device app data. OpenAI's strengths are a more expansive memory, more conversational live voice, and integration with agents.
Here's a feature list that I think the device will focus on:
Some people will find this useful. Some people will find it dystopian. It's been pretty clear that OpenAI has been working toward this since Sam Altman's one-word "her" tweet in 2024.
From an OpenAI blog post yesterday about guardrails for teens using ChatGPT: "...we trained the model to avoid romantic roleplay, age-restricted challenges, and presenting itself as a substitute for real-world relationships. In addition, we applied age-appropriate boundaries around sexual content, eating disorders and body-image risks, age-restricted goods, dangerous activities, and graphic violence. Finally, the model encourages connection with trusted people when a teen may need support." ↩
Ryan Barrett blogging some good news about clean power:
We added 849 TWh of generation overall in 2025, an increase of 4.5%, but GHG emitting (ie fossil fuel) generation shrank by 0.2%. Clean energy covered the gap and then some.
🌎
Great blog post from Terry Godier about the unpredictable nature of the App Store:
On the web, if I build something that works within the standards, it works. Whether people use it is up to them. On iOS, there is another question entirely: whether Apple decides it should exist.
Wavelength 2.0 for iOS has now been waiting for review at Apple for 6 days... I just have a feeling they are trying to figure out if they can force me to use in-app purchase somehow. Google approved it in a day.
Spent a little time this evening adding another database backup location. My primary database server has been so reliable, but eventually I know my luck will run out.
Mark Gurman reports on OpenAI's upcoming device:
The product — essentially a smart speaker without a display — will be shaped like a doughnut that’s roughly the size of a hockey puck [...] The battery-powered item will be designed to work in different positions — in a user’s hand, say, or placed on a nightstand or kitchen counter.
Can't wait to see this. 🍩
I've switched Micro.blog's Standard.site content format to the Markpub.at lexicon. It includes just Markdown, not HTML, but should be compatible with other AT Proto apps now.
Alan Jacobs blogged about the Demis Hassabis quote I wrote about:
The Big Tech companies have been so indifferent, or even hostile, to human well-being for so many years now that their promises are worth absolutely zero.
I feel the same way about much of big tech — especially Meta, TikTok, and to an extent Google, Amazon — but I sort of compartmentalize Demis (and some others) as having their own principles. I do think that a lot of the distrust in AI is because big tech burned all the goodwill, though. See also: Meta glasses, privacy, etc.
This blog post from both Sundar Pichai and Demis Hassabis has the unfortunate title "The next chapter of our AI momentum", but it does contain some interesting perspectives. Demis will be spending more time with Isomorphic:
As you've heard me say many times, I’ve always believed the No.1 application of AI should be to improve human health. It’s time for AI to prove its unequivocal value to the world, and what better way to demonstrate that than to help finally cure diseases like cancer.
There are, of course, already positive and negative effects from the AI industry. Based on our beliefs about the technology, we'll either amplify or downplay aspects of it. It's pretty easy to find both the good and bad.
Earlier this year I wrote about AI for health. That was a fairly cautious, practical essay. I'm confident it will be proven correct because frankly, it wasn't even very ambitious.
I've heard Demis use the phrase "cure all disease" a few times before. It sounds at the same time outlandish and exactly the right goal. Maybe we don't get there in our lifetimes. Maybe we never get there. But I certainly don't fault anyone for having a dream to make everyone's life better.
Sam Altman posted on X the other day:
i would rather be an optimist and work hard than a pessimist posting about why things won't work.
Many people are cynical about Sam Altman. I get it. But I agree with him here and while his post wasn't about Demis, it fits.