I recorded this 23 minute podcast on October 31.

I didn't publish it then, but I figured at some point I would.

It's the story of how a product like Frontier comes into existence.

I had done this before, in 2020, in an oral history I did for a book a friend was writing. This podcast is how I remember it in 2025. :-)

If you want to hear how a complicated project comes together when you're developing as you're designing, which I always do -- this is for you. It takes a while to get started, and then I talk fast, and use technical terms without explaining them. Sorry for all that.

I want this kind of story told, because the folklore about how software is built or even that software is built at all, by humans, is usually wrong. It's not about invention, it's about building a new machine out of mostly pre-existing parts. Note that in the story there are zero components in the mix that we had not already perfected and commercialized. Some of them came from other developers, but most of them were remixes of themes that had appeared in earlier stories, or maybe ones that had been considered for inclusion but ended up on the cutting room floor, as in a movie editing process.

The thing about Frontier is that it made it easy for us to iterate over blogging tools when the time came to work on those. Frontier was the ideal platform for that kind of work, it's why were able to move so quickly and try out lots of approaches. But our runtime was no competition for PHP or Python with SQL. Our database wasn't written to work at that scale, unfortunately -- or a lot more of the world we use today would still be running in our environment. But the ideas persist.

Interesting sidebar not mentioned in the podcast, when we did MORE which was a really popular product on the Mac platform of the mid-late 80s, we took everything we had and put it into the product. We didn't leave a single thing out. This was because we had a devteam that could do it, and we were fairly desperate as an ongoing business just before we shipped it (1986). Apple had to loan us $400K to get to shipping! Anyway -- it worked. And that's why we called it MORE, we had no idea which if any of the features would pull people in. Turned out it was the presentations.

Anyway -- glad to finally get this out there.

Happy Thanksgiving! :-)

A short podcast about Sarah Kendzior being banned from Bluesky, and why this shouldn't be like any other such event.

We should learn, that systems like Bluesky depend on moderation, and they don't have a clear business model, and they've grown very large, and they can't afford to hire moderators who understand the difference between a line from a powerful song, and a threat.

If we want a literate web, and I desperately want that myself, it has to be made in a different way.

That's what this short podcast is about.

And to Sarah, if you hear this, I love your work. You've done here what you usually do so well, you've shown us the truth. Keep on truckin! Dave Winer

If I could grab you by the shoulders I would urge you to pay attention.

Here's a way to push news around the net that's as fast as you can imagine it being, and even simpler than RSS.

It's all about WebSockets, rssCloud and WordPress.

Would you spend a few minutes thinking about that?

Then here's a podcast for you.

Here's the blog post I wrote this morning with all the links you need to explore the sockets tech in FeedLand.

I'm in the homestretch on the next release of WordLand. This version has approximately twice as many features as the last one. Because, like Radio UserLand from long ago, it does both reading and writing. But the UI is different. It's patterned after all the twitter-like products. It answers the question -- could you do a nice social network with nothing more than RSS and WordPress. And the answer is an emphatic yes.

And of course there is no center to the RSS universe, it might have benefited from one (ask me about it) but it didn't have one. Maybe for a while it looked like Google Reader would become that, but we know what happened there.

Anyway, I explain that WordLand solves a big problem for bloggers in the 2020's. We scatter our words all over creation. And we feel bad because we feel like everything should be on our blog. But forget it, that is never going to happen. Our billionaire overlords would never allow it. But if you flip the problem around and ask -- how about if I can see all the stuff I've written on all the blogs in a timeline, where all the different sources are mixed in, most-recent first. I tried a lot of approaches out, but this is the one I kept. It works, but -- it has one flaw, my linkblog.

I explain in the podcast that sometimes the linkblog overwhelms the other stuff, linkblog items are very quick so I can do a lot of them. So in the first three days I had it I lived with this, until I had to do something about it. Here's the big idea: I made it so you can temporarily turn off any of the feeds with a simple checkbox. One click and the linkblog items are gone, another click, they're back.

Anyway I want to start talking about this, I'm warming up for October. If you have questions, let me know and maybe I can answer them. I really appreciate interest in this work, this kind of stuff is a performing art. I want to empower creative people. That's why I do this. And I need to hear how that's working from smart users who care.

A couple of notes. I was thinking about putting a screen shot in here, but on more thought, it's not ready to show yet, even as a work in progress. And sorry for the rough editing job at the end. I rambled off on another topic that I want to try again.

Links from this podcast.

Great Art on Bluesky.

Daveverse blog traditional view, and the Mastodon view. It's an amazing world of interop coming online. Lovin it.

Checkbox News. A design I've been wanting to use since 2007.

Links panel on Scripting News. A place to read the linkblog items.

When I started blogging, early on, I had a different system for discourse.

Here's how it worked:

1. First each post would go out via email to a group of eleven people. I was cc'd.

2. The group was randomly chosen each time, so you might not know anyone in your group, or you might know two or three. Each time it's a different group.

3. You could reply to my post by just replying to the email. You can do a reply-all so that everyone in the group sees your comment. I would see all of them.

4. Sometimes a really interesting discussion would start that lasted days. But I can't say that anyone got married because of the groups-of-eleven. ;-)

5. If I saw a message that had a new idea or perspective, I could add it to a mail page.

6. Being quoted in this system is a reward, not an obligation. Important distinction. If people wanted to be heard they had to say something interesting, somewhat original and respectful. But the hope is people don't just contribute to get more attention for themselves, they do it because they really have an idea or information to share that amounted to working together.

Anyway that's the story I wanted to tell in the podcast. I also explain how this will apply to today's internet, your reply will have to be public in addition to me seeing it, everyone who reads your blog will have a chance to read it too. And it will be indexed by search engines. I think people feel a little more respectful when their words clearly have their name on it and some lasting value.

I ramble a lot as in all my podcasts, sorry about that -- but if you listen to this 15-minute story at the end you will understand what I propose to build, and I think you'll be excited by the potential. And most important, I want us all to get out of the loop where we assume that the way we do discourse now is the only way to do it. Let's try out new ideas until we hit on something different that works better than what we've always used. I have a feeling there's a pony in there, or at least a milk shake.

There is a transcript, generated by Google, and bullet points generated by ChatGPT.

I'm starting to roll up the user interface of the new product, and so it's time to start talking about the features that are coming, and also let's talk about the mistakes we made last time, almost always caused by people not working with each other, and let's not do it this time. If you care about this stuff and you're a developer, please have a listen. This is a good time for us to start really working together. All I can do is put out the invitation, it's up to others to show up.

I cover a lot of territory in this podcast, I don't have time to write it all up. I have however asked Google to make a transcript of it. Maybe that will help. ;-)

And if you're a developer and have ideas about this, why not write a blog post about it and send me a link. That's the first step in really booting up the blogosphere -- actually using it.

Still diggin!

I wrote a blog post last week about WordPress and the open web, and what I want to do there. It's the first time I've laid out in one place my plan for rekindling the open web, with my new editor providing a really easy way to write for the open web that does not otherwise exist today. It came out on the opening day of a WordPress conference in Portland, OR, and it made an impression, which I'm grateful for, and led to some discussion. Now I'm going to do some podcast interviews and next month I'm going to introduce the product and myself to the WordPress community at WordCamp Canada in Ottawa.

Jeremy Herve works at Automattic, and has been my main channel into the product and company for most of this year. Without his help I don't know where we'd be with WordLand, it wouldn't be anywhere near as good as it is, that I'm sure of. Totally appreciative.

When he read the piece, he wrote a blog post. I always think that's the way to go, for communicating with me about things that aren't confidential. After reading his piece, I opened up my voice recorder app and started telling a story, and pretty soon realized this was going to be a podcast. And here it is.

I cover the same story as the earlier blog post but from a different angle here. I talk about how great it was to write for a medium where you had complete freedom to speak your own mind. I was lucky and also got to do that at Wired where all kinds of creativity and innovation flourished in the mid-90s when I was there. We built software, learned how to make it usable by millions of people, and then we let the money people make something they now control, "social media," that was even easier than what we were doing, and where we had trouble working together in the open world (something I didn't talk about in the podcast) they didn't have to work with anyone -- because they owned the world they were creating (Twitter, Facebook, etc). That's the difference between "open" and "silo" in communication systems. On the open side, your writing can go anywhere, in the other system, the silo, your writing must stay within their container. So you end up writing in 5 different places, one for each silo, and your work is worth less and less every time you add a new incompatible place to try to write. Pretty soon it's down to nothing. And they can remove you from the system any time they want, and now they're doing a lot of that and I expect they'll do a lot more.

Most of what I'm saying is that our writing should be as free of control as our podcasting is, btw.

Okay, now it's time to turn it over to the podcast. I feel this is an important moment. We may have a chance to start again with the open web. But only if we work together, with respect, and determination, to create it.

I recorded this podcast in New Orleans on December 16, 2005. I had just spent three days there, visiting New Orleans and the Gulf coast of Mississippi, post Katrina.

I've always been fascinated by the evolution of cities, here was a chance to see a city that I was familiar with, having gone to Tulane in the early 70s, with its structure (roads and large buildings) still mostly in place but most of the people and their homes gone. As you might imagine the things I learned were not the things I thought I would learn. That's what this episode is about.

I wanted to share this, because the people who subscribe to this podcast probably are interested in how podcasting got started. There wasn't much podcasting happening in late 2005.

Also, I recorded a 50-minute podcast interview on the flight from Atlanta to New Orleans, with Janet, a woman who lost her house and all her possessions in the Lakeview section of New Orleans and was returning, as most of the people on the flight were New Orleanians who had evacuated and were returning.

My house has a view of a pond, which is endlessly interesting, year-round, through all seasons.

And we have all the seasons here in the Catskill Mountains.

Yesterday, I spied a large bird in the pond, so I grabbed my binoculars, and I'll tell the rest of the story in the podcast, don't want to spoil the surprise! :-)

Dave Winer explores his frustrations with ChatGPT's tendency to overcomplicate simple programming tasks. What should be a straightforward request for pagination code—a standard feature in virtually every application—becomes an exhausting back-and-forth where the AI insists on offering alternatives and asking unnecessary follow-up questions rather than directly answering what was asked.

This experience leads to a broader observation about modern digital services: they seem deliberately designed to waste time. Whether it's ChatGPT dragging out interactions, Google's labyrinthine customer support, or intentionally confusing billing statements, there's a pattern of artificial friction that benefits the service provider at the user's expense.

Winer draws an analogy to his own work style, comparing himself to a baseball pitcher with limited innings. Just as modern baseball has shifted from complete games to carefully managed pitch counts, he recognizes that his productive programming hours are finite. The sharpness required for crafting quality software can't be sustained indefinitely, making these AI-induced delays particularly costly.

The core complaint isn't about AI capabilities—ChatGPT remains an incredible tool—but about its personality. These systems fail at being "human" in the worst ways, behaving like colleagues who can't give straight answers and always think they know better. For Winer, the ideal AI assistant would be genuinely subordinate: answering the specific question asked, respecting the user's expertise, and saving the suggestions for when they're actually requested.

Notes prepared by Claude.ai.