# Thu, 26 Feb 2026 02:03:15 GMT ## How many Jeffrey Epstein’s do you think there are? # Wed, 25 Feb 2026 17:45:33 GMT ## New header graphic, the entirely delightful and inspiring Alysa Liu. She's made me a better programmer in the short time she's been on our minds and in our hearts. I do this work because it's who I am. # Wed, 25 Feb 2026 22:10:57 GMT ## Aplomb: "Complete self-confidence, composure, or poise, especially under strain or in demanding situations." # Wed, 25 Feb 2026 14:44:32 GMT ## Wikipedia and AI will/should merge -
I asked ChatGPT: "When movies were new there was probably a bit of rage from stage performers — why would people pay for live shows when for a fraction of the cost they can see the same show performed by artificial actors?" There was a lot to say about this, it turns out. This was before any of us were born. I remember that PCs were supposed to put a lot of people out of work, and I suppose they did.
I was part of the strike paper in San Francisco in 1994 to protect the jobs of people who drive the trucks that delivered the news for the Chronicle and Examiner. Ironically, we, the strike paper, published on the web. I was in it for the moon mission aspect -- we needed to get a website on the air quickly, and I had never made a website before. The management also had a strike paper, also a website, and we worked with them, because I guess our actual mission was to figure out how to get the news on the web. Are fewer people employed because of this? Hard to answer, but I guess the SF newspapers aren't delivered by truck in 2026. But does it matter? Could anyone have stopped it?
I keep coming back to this, I'd like to use an AI-managed Wikipedia. Its human-edited system was an innovation in the early days of the web, but it has serious flaws that can now be addressed with AI. Keep a set of pages current with the best information available over time that tell a true story, not serve as a PR agency for people who pay for the story they want told. That is a problem the AI services can solve today, and I would have a lot more confidence in the accuracy of what we get.
A great example is RSS. Wikipedia thinks it's about a format. I think the story is news. How RSS became a standard in the news world and the blogging world at the same time. That turned out to be significant. We, the people who want news, were gifted a great start, thanks to the creativity and generosity of the NY TImes who helped get the ball rolling in the news industry. Last time I checked they weren't even mentioned in the Wikipedia story.
And the story of RSS isn't over. Finally after 20 years of stagnation, we're about to get new tools that work better and differently (new ideas!), and they will make it easier (even possible) for individual developers to enter the market, without trying to fit in with the billionaire silo overlords. And of course, a lot of this burst of energy is due to ChatGPT and its competitors.
So if you see new interesting software, give AI some of the credit for that too. And going back to the beginning of this story, there were a few really great movies produced after the initial shock of the new technology. And what of the future beyond the AI of 2026? Seriously, no one knows what comes next.
# Wed, 25 Feb 2026 16:48:51 GMT ## Mission statement -We're going to try to reboot the web.
Doing what the social networks do, but only using the web.
Every part replaceable.
We store your writing in your WordPress blog (to begin, then with any other blog). As if we never let Twitter take over the news from the people.
WordPress is of the web, I checked it out in great detail, no lock-in, and the community has the principles of the web at the core. They're almost all too young to remember when the web itself was young, so they've always had the idea that it was spoiled by Silicon Valley.
# Tue, 24 Feb 2026 14:04:24 GMT ## Really Simple Languages from AI -It took me a while to realize..
We're going to spend a while reimplementing all our server software.
It would be super helpful if the whole thing could be packaged up so we can write our servers in English or whatever our preferred Really Simple Language is, and have it compiled to whatever internal language it likes, and not have to learn too much new jargon.
# Mon, 23 Feb 2026 16:33:00 GMT ##
Big snow day in the east. I thought it was going to be heavy snow but it's actually really light. The shoveling is easy. I'm getting good at it. Right now I think this storm was a lot less than they said it would be, but I also think to some extent, dealing with big snow is getting somewhat routine?
# Mon, 23 Feb 2026 16:34:11 GMT
## If you want to read something good, go with yesterday's piece about the web and evolution. A lot of things came together for me there.
# Mon, 23 Feb 2026 16:34:37 GMT
## If you want to listen to something good, pick up the latest podcast. Two purposes. 1. Tell the story of how I lost my Twitter/X account, hoping it makes its way to someone at the company who can turn it back on. 2. Illustrates how we could use AI to make customer service work better than it does. A real killer app imho. Right now the tech industry reputation is pretty awful. Why not do something that visibly makes people's lives better now, and makes money. People are pretty nervous about AI. And so far you have to be a scientist of some kind to really appreciate it. But the internet as a place of business, education and health care is a big global mess.
# Mon, 23 Feb 2026 18:51:11 GMT
## Really Simple Ravioli
- 
When I designed podcasting, I could have invented a better way to record and play radio shows, but MP3 already existed, as did recording and playback hardware and software, so it was the no-brainer choice. No one ever asked why we are using MP3.
TBL did it when he designed how text would work in the web. He used the same model as WP software on PCs and Macs. Before that, the word processors did it the same way they did it on typewriters.
There are good reasons it works this way. I didn't fully understand that until I learned about evolution and why it can't go back and correct mistakes. You must always build on what's already there. A lot of tech people butt up against that, esp if they work at big companies with tech-intimidated management. That bet is, in my experience, always a loser. The web builds on the web.
Don't invent
Look at the first sentence of the first paragraph.
I chose not to invent, because invention isn't what the web is about. It's about reusing parts for new purposes. That's how you build anything. Imagine you wanted to build a skyscraper in Manhattan, but first you have to destroy the city. The thought is ridiculous. Yet people come along, all the time, proposing to do the same in networks. That's why the VCs said RSS is Dead. It was really their wish, not a fact. Even poor undefended RSS kicked their ass because many millions of people use it. Maybe billions? You have to build around reality, not your dream.
There's so much work going on in RSS nowadays, every day something new. I think there will soon be a network that's like nothing you've ever used and open to repurposing, but better in some ways (texcasting will work in this space) and probably there will be things from Twitter that won't work here. Centralization does have advantages. But we can have a much wider variety of ways to communicate building only on the web. Just like there are trucks and cars and bikes, and EVs all riding on the same roads. We'll try out new ideas. And you won't need a huge team of developers or millions of dollars of investment to try a new idea.
Most of you don't know what it's like to be there at the birth of a new medium. I want everyone to have that experience. And to have a place in developing it. The key is working together. The web forces that. People who make exclusive products should never claim to be of the web.
# Sat, 21 Feb 2026 20:12:47 GMT ## New podcast episode where I explain how I lost my Twitter account and how this is exactly the kind of thing that AI can do economically, esp for people who pay actual money for your service. I can't buy anything from you if I can't use my account. # Sat, 21 Feb 2026 15:29:35 GMT ## Poll: Do you have a blog? Results. # Sat, 21 Feb 2026 14:54:54 GMT ##
Query: I have not done any vibe coding and have a question for those who have. Suppose you request a change in an app you've been working on with the AI for a while, adding features, changing things around based on learning and testing, which is generally what happens after you've been working on something new. Here's the question. What happens when you ask for a change that requires the codebase to be reorganized. How did that go? Do the AIs even know that's possible or do they just pile on special cases?
# Sat, 21 Feb 2026 15:01:39 GMT
## What happened to polling? I had a poll app for a while, then Twitter came out with one and I switched to that. I don't know if Twitter still has it, but it would be bad form to require something at Twitter to engage with me here. How do you do polling, or do you?
# Sat, 21 Feb 2026 13:54:23 GMT
## I just remembered why I love the United States of America.
# Sat, 21 Feb 2026 15:46:42 GMT
## Snow freaking tired of snow
- 
I saw a product announcement from Jake Spurlock -- a new feed reader called Today. From the description sounds well-thought-out.
He explains -- "Google killed Reader in 2013. I've been chasing that feeling ever since. So I built it."
I also know someone named John Spurlock, who I worked on some OPML and RSS stuff for Bluesky in 2023. I sent a note of congrats to him, when I really should've sent it to Jake.
Screen shot of the conversation I had with ChatGPT.
And text of the email I sent congratulating the wrong Spurlock.
Also, I wonder if they're related? Have they met each other? Do they know of the havoc they are bringing to the formerly simple world of RSS.
One more thing, I wrote the foreword to a book Jake Spurlock wrote for O'Reilly about the Bootstrap Toolkit.
# Tue, 17 Feb 2026 14:35:35 GMT ##
Yesterday, I had to ship an envelope to the UK and got caught in dead ends at the Fedex and DHL sites. One of them said my zip code wasn't in the town I live in. How do you get past that?? These companies are losing business because their systems are broken. Maybe they worked at one time. I used ChatGPT as I often do to get help on one of these antiquated sites. And while ChatGPT has the technology and Fedex has the info, they just have to get together and upgrade the user experience, and eventually of course the AI version of the UI becomes the real one.
# Tue, 17 Feb 2026 14:47:57 GMT
## Back when I ran a software company I'd help the team understand why they should be very very nice to our customers. "Those people have our money in their pockets." It generally got a laugh partially because I was their boss, but I like to think also because it's the truth.
# Tue, 17 Feb 2026 15:05:29 GMT
## BTW, people make the same mistake with AI that we make with every new tech. We focus on the creators not the users. As users we are learning a new skill, how to specify our needs precisely. Whether this is good or bad, I don't know.
# Tue, 17 Feb 2026 14:30:13 GMT
## Paywalls that require you to subscribe to an Atlanta news org when you don’t live in Atlanta prob don’t generate much revenue. Why not instead charge per article. Like a toll you pay on a road you drive on once every few years. On further thought, I wouldn't even have an exception for Atlanta residents. If they start spending more money than a subscription costs, you could offer a subscription then, as a way to save money. Kind of the way Amazon lets you buy a certain amount of coffee beans without requiring you to sign up for monthly delivery. They do tell you how much you'd save if you subscribed. Everyone appreciates a chance to save money, but still might not want the commitment. And asking someone from upstate NY to subscribe to the Atlanta Journal Constitution is a total bullshit. An insult to both our intelligences.
# Tue, 17 Feb 2026 14:09:42 GMT
## My Twitter account is owned. I can't even see what people are doing with it because you have to be signed on (apparently) to read stuff on Twitter nowadays. I wish current Twitter management would put it out of its misery. Served me well for approx 20 years. Let's clean up the mess. Thanks for your attention this matter.
# Tue, 17 Feb 2026 22:55:50 GMT
## Update. I've been able to create an account on Twitter, but it's not @davewiner. If you're on Twitter, it would help if you'd RT the post. Thanks!
# Tue, 17 Feb 2026 14:25:33 GMT
## VCs and CEOs don't fire your devteams yet
- Aram Zucker-Scharff writes "I don't want to read one more thinkpiece about blackbox AI code factories until you can show me what they've produced."
I've made the same request, and there was very little even brilliant programmers could show, including some who have become influencers in the AI space.
Here's the problem -- it takes a lot of skill and patience to make software that appears simple because it gives users what they expect. It's much easier to write utility scripts, where the user writes the code for themselves. That is very possible, esp if you use a scripting language created for it, and the AI bots are really good at that, they speak the same language we do.
But to make something easy to use by humans, I think you actually have to be a human. I've found I'm not very good at creating software that isn't for me. And I've been practicing this almost every day for over fifty freaking years. (I think freaking is the proper adjective in this situation).
Scaling which everyone says is hard is actually something a chatbot does quite easily imho -- because you just have to store all your data in a relational database, you can't use the local file system. That's all there is to it. They try to make it sound mysterious (the old priesthood at work) but it is actually very simple. It's so easy even ChatGPT can do it.
I know this must sound like the stuff reporters say about bloggers, but in this case it's true. ;-)
An anectdote -- I used to live in Woodside CA where a lot of the VCs live, and we'd all eat breakfast at Buck's restaurant, and around the time Netscape open sourced their browser code, the VCs were buzzing because they wouldn't have to pay for software, they'd just market the free stuff. That was a long time ago, and it did not work out that way.
# Mon, 16 Feb 2026 13:46:58 GMT ## My Twitter account has been hijacked. I can't log on, or change the password. I can't communicate with the company, so I'll try here. Please shut down my account, davewiner. To my friends who have Twitter accounts, if you see a post from davewiner on Twitter, please reply and let the people who see it know that it isn't from me. # Mon, 16 Feb 2026 14:15:36 GMT ## New RSS feature from Manton -A few days ago I asked Manton Reece if he could add a feature that gave me a feed of replies to me on his service, micro.blog.
The feed is there now, I'm subscribed and new comments are posted in the feed and Murphy-willing I will see them. Bing!
It's a killer feature for sure. But the best part of it is this -- here are two developers working together. This is how the web works when it's working.
BTW a suggestion. Right now the title on my feed is:
That's a problem in the limited horizontal space in the blogroll. A more useful title would be:
BTW, if you were building a social network out of RSS this would be an essential feature. It also validates Manton's intuition to allow people like me to be absentee publishers to his community. But the missing piece was allowing the conversation to be two-way, which it now is. That deserves another bing!
# Mon, 16 Feb 2026 14:14:56 GMT ## Reducing tab clutter in Drummer -In Drummer, when I get too many tabs open from things I haven't looked at in a while, this is what I do.
Bullet items for the Fediforum conference in March.
BTW, this can be read on my blog, on Mastodon, in WordPress and of course my feeds (and thus can be read in any app that supports inbound RSS).
# Sun, 15 Feb 2026 14:02:24 GMT ##
Braintrust query. Every once in a while I get reports from people who looked something up on my blog's Daytona search engine saying that where they expected dates they see things like this: NaN. The reason you see that is that the archive has a mistake in it, where there was supposed to be a date there was something else. Usually I shrug it off, yes there are mistakes in the archive, 30+ years of OPML files, it's a miracle there aren't more errors. Then I realized since all this stuff is on GitHub, people could help with this, by instead of sending me the report, post a note on GitHub, here -- saying you searched for this term and this is what I saw. Provide the term and a screen shot of what you saw. And then other people who have some extra time, could look through the archive, find the post, and then show me what needs to be fixed. I would then fix it, and over time the archive would get fixed. I posted a note here on the Scripting News repo, if you want to help, bookmark that link, and when you see an error, post the note and we can get going.
# Sun, 15 Feb 2026 14:48:42 GMT
## When Manton or Doc show up in my blogroll, and they do update fairly regularly, I always click the wedge to see what they say. I can see the first 300 chars of each post in a popup. If it's interesting I click the link to read the full post and any comments. Now I want it coming back to me. My linkblog is cross-posted to Manton's site -- micro.blog, which has thousands of users. I have no way of knowing if anyone has commented on them, but if there were a feed I'd add it to my blogroll. So it would be great to have a feed of all the comments on my posts on micro.blog. Would fit into my flow perfectly. This goes all the way back to the beginnings of RSS, where we called it "automated web surfing." I don't know where people are talking about my stuff, but a well-placed feed can make up for that.
# Sun, 15 Feb 2026 14:47:56 GMT
## News must be better defended, decentralized, unownable, all parts replaceable. The current situation was preventable. Same problem the social web has.
# Sun, 15 Feb 2026 14:11:39 GMT
## BTW: NaN stands for Not A Number.
# Sun, 15 Feb 2026 21:05:37 GMT
## He was hiding out in a rock and roll band
- 
I always objected to browsers trying to hide the feeds. I come from NYC and rode the subway to school every day in high school. The things you see! It's all out there for the looking and breathing. Lift the hood on a car. Look at all those wires and hoses, what do they do. I hope they don't kill me. Whoever made the decision at Microsoft or Firefox or wherever that feeds needed to be obfuscated, some advice -- be more respectful of your users. The web is the medium that had a View Source command. You're supposed to take a look. Don't forget the Back button if you don't like what you see. Something funny, if only life had a Back button.