One thing I look for when listening or reading someone’s commentary, is their opinion the most obvious conclusion from a very surface-level reading of the source material? The obvious is often wrong, otherwise we’d never have anything interesting or surprising happen in the world! Gotta go deeper.

On the latest Pivot, Kara and Scott seem way off on their AI take. OpenAI wouldn’t have leaked the hardware details to Bloomberg this week if the product wasn’t going to be revealed. I just can’t imagine Jony Ive personally hyping a vision for something that we never see.

Watching the news during breakfast… Major flooding in Texas again. Hope everyone’s safe.

This ad from Anthropic is so strange. I saw it a couple days ago and rewatched it just now. Starts with a burning house, ends almost hopeful. I can imagine what they were trying to go for, but in this edit it doesn’t land for me.

Codex Micro is a clever little hardware add-on with buttons and knobs. I’m not a clicky keyboard person, so this isn’t for me, especially at $230. Also I’m working at a coffee shop right now and can’t imagine taking this with me anywhere. But I bet they sell a bunch.

Wasn’t expecting this, but I guess it’s obvious now that I see it. Sarah Perez writing at TechCrunch about Apple Intelligence in China:

Apple Intelligence, the iPhone maker’s generative AI offering, is coming to China. On Wednesday, Reuters reported that China’s regulator, the Cyberspace Administration of China, approved Apple’s AI services in the country, on the back of a deal to integrate Alibaba’s Qwen AI model into Apple’s operating systems, including iOS, iPadOS, macOS, and visionOS.

If Qwen proves “good enough”, it also reduces Apple’s dependence on Google.

After working on something that not everyone will agree on, I love returning to something universally good: books! Here’s another view of the upcoming books calendar feature, with an agenda-style list that works well even for not many books. For your blog, you can choose this or the calendar view.

Dave Winer has deployed a separate RSS.chat instance that other people can join to play around with: demo.rss.chat.

A couple days ago, OpenAI removed the 5-hour usage limit. I think it was meant to be temporary, but I wonder if they’ll make it permanent. It’s one less thing to worry about. Lots of numbers in a UI introduce friction, making your brain think about things that don’t really matter.

Three separate posts dropped yesterday with OpenAI’s side of the story in the Apple lawsuit. First, a more complete official statement from OpenAI, reported by various outlets: While we take these allegations seriously, we’re not aware of any evidence that this complaint ha...

Thinking about Om Malik again. We only exchanged a handful of emails. I was able to interview him at Micro Camp, and I put an edited portion of it online here for my book, which I was skimming through today. It’s from 2023 but still feels relevant. Such a loss.

Demis Hassabis published a post — on X and LinkedIn — about needing a framework to steer AI development, with AGI possibly just a few years away:

Even if we solve these hard technical challenges, there will be further complex economic and philosophical questions to tackle: what sorts of new economic models will be needed to help everyone thrive in a post-scarcity world? What values do we want to live by, what will meaning and purpose be, and how might even the human condition itself change?

I put a lot of thought into reevaluating the X API from a blogging perspective. I also wrote a whole book about the problems with social media, and I updated it for Elon Musk. But it is stressful, and I’m in no hurry. For people who have asked for this to be re-enabled, let’s give it some more time.

Great analogy from JA Westenberg about SaaS and bread machines:

A company that decides to build its own version using AI coding tools is buying a bread machine. The ingredients are cheap, and the machine does most of the work, but they’re now the baker. They own the maintenance, edge cases, and security gaps that AI-generated code tends to introduce…

Most people just buy sliced bread ready to eat. Most people will still buy apps. Although the quality bar is higher now.

Congrats to Manu Moreale on 150 interviews for his People and Blogs series. It’s an incredible archive, what makes the web great. He’s downsizing now:

My digital life follows a somewhat regular rhythm and I alternate through phases of expansion, where I buy domain names, ship new projects, start newsletters, and chase a million ideas, and phases of contraction, where everything happens in reverse: domains are left to expire, projects are archived, newsletters are deleted, services are cancelled.

Life seems at once fleeting and yet almost long enough. At 50, I feel lucky to be able to look back on what feels like a complete life already. My kids are adults. I’ve had a good career with a few interesting twists. Joy, heartache. But also I want decades more. There’s so much to see and do.

Great memorial to Don Iwerks who passed away at 96:

An allergic reaction to photographic chemicals forced him out of the Process Lab and into the studio’s Machine Shop. It was a twist of fate that proved transformative for both Iwerks and the company. Over the next three decades, Iwerks designed cameras, optical printers, special effects systems, and projection technologies that became essential to Disney productions and attractions.

Not as well known as his father, Ub Iwerks, but what an amazing career. Rest in peace.

Mixed feedback on my post about X, mostly good. I’m going to hold the rollout another week so I can consider everything. No deployments yet.

Also there’s some confusion about whether all paid Micro.blog subscriptions send money to X. Nope, this would be a change only for users who want to use X.

Sometimes it feels like my job is just to manage thousands or millions of queued background tasks and make sure all the servers are perfectly in balance.

Rainy morning. Walked from the mechanic around the pond to Central Market for coffee. Reminds me of when my kids were little and we’d play over here. 😢

To John Gruber’s point earlier today, I just opened the App Store and literally the entire screen is just ChatGPT and an ad. Hugely valuable. Millions of people see this.

When building a product that does so many things, no one is going to agree on every decision, but I think Micro.blog and our users are aligned on the big things. The mission has never changed: post short posts (and everything else!) to your own blog, which makes the web better, and also opens opportunities for smaller, safer communities.

I use an old Pixel 4 for testing Android. I really like the buttons: only a power button and up/down volume. All the buttons are on one side. Not sure how I would redesign the iPhone buttons, but I don’t think they’re quite right.

A month ago I wondered out loud if it was time to add Twitter cross-posting back to Micro.blog. After a lot of thought and feedback from users, I think yes. The goal has always been to help people publish on their own domain first, regardless of where they also choose to sha...

There is always going to be some water spilling, might as well water the plant with it! Great idea at Lazarus on Airport. ☕️

Deployed some fixes and new settings for Mastodon this morning. Always feels good to make things just a little better. 🎉

Satya Nadella writes on his blog about what we potentially give to models for training. He focuses mostly on the enterprise, which I don’t care much about, but this part does describe a real tension of commercial models:

While the great innovation that comes from model providers having fair use rights to train models on public data is needed, I find it ironic that the status quo is to then turn around and impose restrictive terms on distillation, and to reserve the right to learn from customer usage and interaction data. If learning flows in only one direction, economic value converges toward the owners of the learning infrastructure rather than the creators of the knowledge itself.

I think this is something that both AI fans and skeptics could agree on.

Good article today from Ben Thompson about Apple vs. OpenAI:

AI is changing the world, and Apple’s contribution is to deliver two year-old technology differentiated by exclusive access to data it won’t share. It certainly has no interest in delivering on a new AI-driven paradigm that endangers its iPhone franchise in any way; any of its employees who wish to do exactly that, meanwhile, are being told loud and clear that Apple doesn’t just think it owns your laptop, but also your mind and knowledge.

Simon Willison blogs about how Anthropic keeps extending Fable access, presumably because too many people would cancel their subscriptions and use GPT-5.6 instead:

At this point I think Anthropic should change track and keep Fable permanently available on those plans. OpenAI are winning users simply due to the uncertainty that surrounds Fable access.

Anthropic should do this, but also if they do, it will make their original claim that they can’t support Fable in subscriptions look like a lie. Very interesting to watch. 🍿

TwoMillionKit is a clever solution from Guilherme Rambo (via Daring Fireball) to use private cloud compute by routing through the command-line fm tool. I was accepted into the Small Business Program a few days ago and had forgotten why I had applied until reading this.