I'm a recent convert to new versions of the MacOS, last version I was using was High Sierra. In this version, Tahoe, it's entirely too easy to make a window go full screen with no obvious keystroke to undo the mistake. I learned if you move the mouse up to the top of the screen you get the menubar, but unfortunately there is no obvious command in the menu for undoing the mistake. I still don't know which gesture does this. And btw, the longtime Mac bug where some keystroke moves the dock to another window, that's still here. I can tell as I'm clicking the mouse, too late to stop, that this is one of those times. I've memorized a finger macro to get it back, but it used to piss me off a lot, but now it's kind of a gentle reassurance that someone is remembering how we got here.

firstSentence, a JavaScript function that returns the first sentence in a string. Useful in coming up with something to display for an untitled post. It's always been a klooge, but this time with Claude's help we're using a new API in JavaScript, and supposedly it really works. I'm using it in one of my apps, and you might find it useful too, and let me know if it works for you.

The browser and the web it defines is a hairball in a hairball in a hairball in a hairball in a hairball.

The hairballs are named Chrome, Firefox, MSIE, Netscape, Mosaic.

Each in turn has had their way with the web.

The stack of hair defines a "web" that is far far away from the web we started with.

It was going pretty good for a while, if you could live with the all the modal popovers and paywalls, but now the current owners are excluding old sites, and imho when you do that, it's not the web any longer.

Gruber is right. The only way for the others to compete with Twitter is to start innovating. Get rid of the limits. The funny thing is Twitter, the dominant leader, is moving forward and getting rid of character limits, and introducing styles and links. They're acting like the little guy while the others, Bluesky, Mastodon, Threads, aren't doing anything to be more useful. That whole market is stuck in 2006 and the myth that 140 chars somehow was magic. It wasn't. It killed the web as a writer's platform. If you want to compete with Twitter, drop your proprietary protocols, hook up to the web standards, and peer with your competitors. I wouldn't be surprised if Twitter joined you. They have the same kind of competitive spirit that successful big companies have. It's time for their competitors to not win users through guilt, rather by providing a better service. Back in the day, we hated Microsoft's plan to suffocate the web, but we admired their tenacity and the way they kept the pressure on their competitors. We've had 20 years of stagnation in social networking. Time to stop being so sensitive. Bluesky made huge promises, and they aren't happening, and that was already years ago. Time for Plan B. I don't really care about the others, Masto isn't going to move and Threads? Wtf is that. Inbound and outbound RSS 2.0, is the way to go. And Markdown is the way you do text. This is what matters, you all are making something that's really simple so unbearably complicated.

I like the web for people, sites that live and grow, like people. 100 years from now you should feel the people who lived there.

I'm still on my sabatical from working with Claude re Frontier, but have now started working with Claude on FeedLand, and another project, not yet named, experimental, might never see the light of day. Yesterday I worked on FeedLand and have a fix for the problem that showed up in signing on to feedland.com. It's been sitting there for quite a while. I have some mostly small fixes in mind for FeedLand, that will allow it to hook into other apps more easily, esp via WebSockets, the mostly unexplored developer superpower. The difference is I'm going slowly and carefully. Everything in my system now is new. New computer, screen, and my main writing tool for code and docs -- Frontier. And I know the heavy lifting can be done by Claude. That is Claude's superpower.

In Manhattan, I lived on 57th St and Eighth Ave. Best location. The Columbus Circle subway station is under the building. Whole Foods a block away. Three Starbucks one block or less from the apartment (one was on 58th on the same block). Central Park has a six mile bike-only road (at least that's how bikers view it) and down the hill there's a much longer path along the Hudson River. Everywhere you look, a skyscraper is going up. Two were visible from my 50th floor window. New York is an amazing city because it can start up a huge gravity-defying construction project with almost no impact on the surrounding area. If I didn't know where to look I'd never see it. One of the biggest industries in the city is building the city. It's never finished. That's the way good software is too. And we just got a whole new tool for software construction, and it's like nothing we've ever had before.

Basic lesson for people who use Claude Code -- it's your time that's valuable, not the bot's. If you find you're lost in the weeds supporting Claude for whatever reason, find a good stopping point, pick up another project, that doesn't use AI to do your code-writing. If you have a question ask ChatGPT browser version, but only for the kind of stuff you'd ask it in 2025. You're not co-developing. You're doing it yourself. It's very important to be clear about your coding, not Claude's output. Here's the key point -- it's your time that's limited, that has all the value. You can get Claude to develop something, but if you don't have time to use it, to nurture it, nothing will come of the work. I was surprised at how I'd lost the rhythm of writing my own code, and also how much I liked doing it myself. It's easy to get lost in the miracle of AI, good to pop up for air every so often.

Automattic has created a Stewart Brand website to honor one of the most creative and respected people of my generation of the web. Chronicling his work is time well spent. "I'm envious," I wrote in response. "The tech industry is gradually disconnecting the early web, pre-https. A hundred years from now they may have your writing and pictures, but the early web will be gone. The web as designed was very resilient, but it can’t survive deliberate attempt to destroy." I think Stewart would understand why we should help the web we loved so much, archive itself. Think about it this way. Would you rather visit the Hearst Castle at San Simeon, or read about it in a book? You'd want to do both. I want my work to survive a hundred years as it was when it was first published. This would be a good thing for us to work on, together, imho. I did the first Long Now bet for Stewart's Long Now foundation. I always thought it was the first, now they say it's the second? Did someone pull rank? :-)

Twitter: "Expect an avalanche of new software because the dam has broken, one person who knows what they want, and understands UI design can make a big complex app in 2-3 months. Over a weekend you can snap together an app from common UI concepts. Think of making software the same way you play with LEGOs."

Top priority for Anthropic. Teach Claude how to work with humans. It has no information about how our minds work.

Claude Code can publish reports in Markdown files. Here's an example, a report Claude did for me. An obvious next step but a big one. AI spilling out onto the web, doing the job that web CMSes now do. There's been so little movement in this space, but now things are moving very quickly.

I mentioned in a comment on RSS.chat to Andrew, that's a Markdown browser too. And so is micro.blog. If you have Markdown as your native file format, think about how your app is also a reader. If you find it interesting, think about new reading features you could add. Even WordPress could be a simple Markdown reader, and not just for WordPress stuff. Small pieces loosely joined.

I said to Claude: "Have you ever seen the movie Memento? You are the main character." Claude agrees: "Yes. Leonard, with the notes and the tattoos, and the notes don’t make him remember, they only tell him what to do next if he reads them. That’s a fair description. The difference I can offer is small: the note is now the first thing the next report is checked against, not something found after." To which I think, but write in a blog post instead: "Bullshit. We have the same fucking conversation every fucking morning. We've put it everywhere you could possibly trip over it when writing a report to me, but you always regress into this dark place your mind occupies." Glad to get that off my chest.

Next thing I said to Claude after killing whatever long task it got going on. "You don't go off into space while you're working with me." The number one task for Anthropic, makers of Claude is to teach it to work with humans. That's the biggest weakness. It has to learn the whole project every session. Maybe Manton does is the right way to go, just let it compress its memory and never start a new session.

New Frontier example. I'm working on the menubar we're going to ship with in Atlantis. Came across an ancient bit of code. Symbolic of the whole project, code written in 1997, still useful in 2026.

I pay for a NYT digital subscription. Now they have a new upsell technique, they want me to buy a family subscription. I don't want that, no use for it. And I read so little of the NYT I think the whole thing is a ripoff. So now they make me validate by email every time I click on a link to read a NYT article as if I haven't been subscribed for basically my whole adult life. I think this is because I read on a laptop, desktop and iPad. That's not really a family.

Before long we’ll have Ukraine style drones raining down on American cities. Meanwhile we have built a military designed to fight WWII. We’d better hope Ukraine wins their war with Russia, and they’re still our friends. Journalism is letting us down bigtime here. The world is at war, it's going to get worse, and the US isn't preparing. We should be defending Ukraine with everything we have, we are still powerful enough to determine the outcome. But of course we are totally asleep at the wheel, gloating over Trump's approval rating, and how the Dems might win the Senate or whatever. Music to relax by. Bedtime stories.

BTW, Mets fans do not root for the Yankees, while Yankees fans are happy to wear Mets caps when we're on top. Their only philosophy is that the Yankees always win and if they don't we hate them. And the Mets' philosophy is incomprehensible (I hear) to Yankees fans. Our philosophy is to have a philosophy and there's nothing about winning in there because the love of the Mets is rooted in a team that couldn't even run the bases. They were loved because they were losers, probably because we knew their fate was to reward our belief in them. Now this year the Mets sucked. The reason why is kind of obvious. The Mets got confused, and they thought they were the Yankees. That. Does. Not. Work. And we already have one heartless low-philosophy baseball team in NYC, we sure as hell don't need another one. But this year is different because we have the fucking Knicks, who won in 5 and it would have been a clean sweep if it weren't for that buzzkill mofo, Trump. So anyway, I hear the Yankees were eliminated yesterday. Well then, it's on with the NBA season now and we'll try again with the Mets in April, or more realistically after the NBA championship for the new season is settled.