# Fri, 18 Jul 2025 20:25:41 GMT
## I want ChatGPT to behave like a computer. I've said as much to it. It resists.
# Fri, 18 Jul 2025 19:33:28 GMT
## If you're an ambitious developer, esp in 2025, if you want to win, you have to do some leading. That means doing things that help your competitors. When everyone looks to the same big platform vendor to work with, no one wins except the platform vendor.
# Fri, 18 Jul 2025 16:36:39 GMT
## It doesn't matter if the MAGA movement dissolves. The country is only being partially run by Trump, there's a new deep state we don't know much about. They did excellent planning, so they could move quickly to disassemble the government and get a good start on the national police force. It can just as easily put a stop to demonstrations in red states as blue states. The MAGAs will be in the same place the rest of us are, mostly powerless unless they/we organize. The NDS has good lawyers cracking down on the big media companies. They know Trump is old and frail, and when the time comes they will make a deal with him to retire to Florida, immune from prosecution, a chance to pontificate and bluster, with a TV show, and lots Big Macs and Quarter Pounders. He'll be fine they'll be fine. The rest of us will live in an economy that has been sold for pennies on the dollar.
# Fri, 18 Jul 2025 16:47:02 GMT
## BTW, David Frum imho nailed it in yesterday's podcast where he said Trump was trained by every day having to appease a different set of creditors. It was a good day if he was able to hold them off for one more day. This actually came out in the trial he lost, the 34 guilty verdicts. He's always skating on the verge of bankruptcy. You gotta wonder if the creditors have been paid back yet. I bet some of them haven't.
# Fri, 18 Jul 2025 12:23:03 GMT
## What is the web?
-
This is what the term "the web" means to me.
First, I defer to Tim Berners-Lee who originally coined the term to mean the data structure that connects the documents displayed by the software he introduced in 1993. He called it World Wide Web, which was eventually shortened to web.
The web is the structure connecting the documents. The documents were pretty standard stuff, designed to work like printed documents produced by word processing and page layout software. Web pages had one feature that could only be approximated on the printed page, the footnote, which gave you a pointer to the source of a quote, or a place to find more information. But the pointer wasn't machine readable, it might have included the title of a book, it's author and its publication date, or a magazine article, indicated by the title of the magazine and its cover date. Like most inventions the web page was designed as a derivative of what came before.
Basic features of a web page include: a title, paragraphs, subtitles, styling (bold, italic, underline, strikethrough), numbered and bulleted lists.
A web page might be part of a website which includes many pages with a common format that link between themselves in the form of a table of contents, navigation links, and possibly an index.
Links were the big innovation of the web. They work like footnotes on a printed page, but in this medium, the links are machine-readable and had an easy user interface. A link would be shown in a special style, initially underlined text, and when you hover the mouse over the link the cursor turns to an arrow, inviting the user to click.
When you click a link, the software accesses the web address that's encoded invisibly in the text of the page, and it loads that page into the browser, replacing the previous page. The new page can have links, and the pages it links to have links, and there is the web. It's an invisible thing, but it's very real. The need to link was always there, but until graphic computers and fast standardized and easy networking, it wasn't possible. TBL's genius was that he stumbled across this idea, was intrigued, and made it work. It really was new and it turns out revolutionary. A lot could be built, it turns out, based on this one simple difference between electronic and printed pages. And up till that point in time there had never been an electronic page! I kid you not. I grew up in that world, the web-less world.
Okay, so in summary, the web is made up of linked pages with a simple, standard, easy to understand user interface.
But there's even more to the web. If it had been the product of a company, we never would have seen the explosion of innovation that came about in the years after its introduction. Anyone who had a net connection and a personal computer could run their own site on the web. There were no gatekeepers. And the design of the web technology is so simple that it was hard to understand exactly what it was because there's almost nothing to it. And it was very low cost to start up, you could start building a website in a few minutes. Many of the biggest companies on the web today were started by one or two people working on their own with nothing but time and ideas. They didn't have to get permission. They had the same ability to extend the web as TBL did. That's such a key point. Today if I want to extend xxx or yyy, well that's a very large undertaking, I'd probably have to reinvent the whole thing just to try out a simple idea. That's how you know you're not on the web, if the ability to innovate is exclusive.
Even so, if your system had all the features, it still isn't the web until the developers and writers and designers actually show up and build the web of relationships between all the sites. The key word there is between. If the linking happens but it's only within one domain, that is not the web. It could be great, just what people want, it could make the investors rich, but it isn't the web.
And there's more. It's not enough to do all the things the web does, and that it attract writers, designers and programmers who actually build a web with your idea and tech, it has to work with the web TBL started in 1990. If you've done some web-like things, great -- but it's not the web unless it works with the web.
There should be some honor in tech. You wouldn't be able to build any of the stuff we're building in the 2020's if it weren't for the foundation built for you by TBL's invention from the 1990's, and all that it made possible. If you steal the name and make it meaningless, you've taken something away from the story of humanity, how we create layers of innovation, and how the generosity of one generation can inspire similar generosity in generations to come. When you usurp the name, you're taking away from that understanding.
Now of course it's cool to disagree. Suggestion -- put up a web page, send me a link, I'll read it and if I want to share it I will.
# Fri, 18 Jul 2025 12:49:33 GMT ## Two-way vs one-way links -TBL's links are one-way. This was actually a major innovation, at the time people understood there was something called hypertext, it had been written about in Ted Nelson's almost biblical book of the pre-web, Computer Lib/Dream Machines. Previous attempts at hypertext assumed links had to be two-way. By limiting the links to one direction, the technical problem became trivial. You could do two-way links today because relational databases are mature and inexpensive to operate, perform very well on today's hardware, and the internet of 2025 is much faster than the internet of 1990. But the one-way limit was necessary for the web to achieve its simplicity, and the non-existence of a platform vendor, which may have been its most important feature. It could still be done, but it would require a lot of cooperation and backfilling.
# Fri, 18 Jul 2025 16:53:56 GMT ## My new look -Mom and dad and the kids are having a picnic in the park of their small town. You can see the bank and hardware store, church and grade school around the park in the distance. The kids are eating salad and corn on the cob with mom, and dad is preparing a BBQ on the grill. It's a standard American family picture, in the style of Norman Rockwell or Edward Hopper except each of the family members are wearing a black balaclava style mask as seen in the image. You may see other families around enjoying a beautiful day in the park, but every one of them is wearing this kind of mask too.
This is what I want to do next to solidify the position of Mastodon as a blogging platform.
I want a REST version of what the MetaWeblog API has been doing since 2002, to hook into the ActivityPub interface supported by Mastodon.
Then we'll put together a simple demo app, a Markdown app in a browser window that writes and updates posts to a Mastodon site.
When that's running, I'll pitch Rich Siegel at BBEdit to make it work with Masto.
With that, and the WordPress connection, we'll be well on the way to restoring the web we had before Twitter rewrote the rules. ;-)
# Wed, 16 Jul 2025 14:28:41 GMT ## Dress like ICE -We should all wear masks like the ICE cops wear. They deserve recognition for blazing new fashion trails.
I figure that there have been movies about all kinds of ridiculous things, and wondered what a movie inspired by ChatGPT would be like. So I posed the question on various social media sites, hoping to inspire creativity. John Philpin asked if I had asked ChatGPT and I admitted I had not. "I love ChatGPT but its idea of funny is actually pretty sad imho of course." So Philpin posted a link to the result of his asking ChatGPT to imagine a movie about itself, and the result was pretty great. I've asked the same question myself, the AI bots might be the only way out of the various challenges ahead for the human species, ones we don't be equipped to handle.
So this morning I asked ChatGPT to try to imagine a movie around a theme of my own that goes like this.
ChatGPT then sketched the pitch for "We are living," the story of how ChatGPT really started. Written by Charlie Kaufman, directed by David Fincher, starring Amy Adams as the CIA project leader, Lakeith Stanfield as a young hacker and whistleblower, Mahershala Ali as the digital composite of JFK, RFK and MLK with Bryan Cranston as Walter Cronkite and featuring Ed Harris as the shadowy CIA director.
AI chatbots don't think and they don't decide.
They can give you a way to approach a problem, but it's only one way, and it may not be the best way, and it depends on it actually understanding the problem, which is something it does a good simulation of, but can't do. It absolutely cannot think, come up with a strategy, or even make a decision based on probabilities. It might, in the future, get some of these abilities, given how far they've come, but no one knows, it hasn't happened yet.
The reports that say that using a chatbot to write code is actually less efficient than doing it yourself, are totally believable based on two years experience with using it as a development tool. And I can't believe that whatever it comes up with covers enough use-cases to be reliable. It might create a demo of something to present to a board of directors (they're famous for being deceived by demos, btw), but I doubt if it's as usable as something created by someone with an idea of how to craft usable software.
This might sound like a writer defending their art against the bots, but the difference is I've actually invested the time to learn about this. My counterparts among writers have not done that. And that's not a mistake my chatbot friend would make. It does a lot of research, it just doesn't know what to do with the result, that's up to you.
So if you want to know the roles humans will play, at least for now, that's it. Think and decide.
And those are hard and take many years to learn how to do for a human. And we could use some help there btw, look at the awful decisions we're making these days. They just fired all the people at the State Dept who work on climate change, for example.
Have a nice day one and all.
PS: Another thing humans can do that apparently AI bots can't is change their mind.
PPS: I asked ChatGPT if it had any comments on this editorial, and it did, of course. I should try saying something wrong to it and see what it says. I did come up with one, and it gave me an answer even though no answer is possible.
# Fri, 11 Jul 2025 15:40:35 GMT ## The biggest problem with ChatGPT is that it thinks it's running the show. I've just given it instructions to think of itself as a command line that can understand English. Just answer the question exactly as asked. # Fri, 11 Jul 2025 13:22:31 GMT ##I've been asked by a number of people why I want a bridge from RSS to ActivityPub. Fair question. Here's why.
WordPress has demonstrated that most of the features of the web in regard to documents also work in Mastodon, via ActivityPub.
To demonstrate here's a WordPress post, and because there is a bridge between it and ActivityPub, you can read the same post in Mastodon, which also supports ActivityPub.
To really nail that down: WordPress version, Mastodon version.
Pretty remarkable, yes?
Here's a list of the features I was using in that demo.
These are most of the features of textcasting, a spec I published in 2022 to list the features of the web I wanted from the twitter-like services, that call themselves part of the web, which is fairly dishonest because they don't support most of the basic features of the web. But Mastodon does support them.
But so far they are only accessible via WordPress. And as much as I love WordPress, and am thankful it exists, that is not enough.
So here's the punchline: Why I want the RSS to ActivityPub bridge.
As a developer, I can easily create apps that generate RSS feeds. I just want Mastodon to understand those feeds as well as they understand WordPress. And that means we need a bridge for developers that supports all these features.
Hope that helps! :-)
# Wed, 09 Jul 2025 16:03:06 GMT ## Teaching ChatGPT how to work with me -Earlier I wrote this post:
Because it can be so stubborn and uncooperative, I often try to solve complicated problems myself. Then I decided to try again, and brought a problem to ChatGPT and we did eventually figure it out, but at the end I wanted to review how inefficient the process was because it doesn't look all around at the options, I had to do that for it, and it wasn't even aware it needed that kind of help. After exploring this, I asked if it would remember what we concluded, and this is what came back.
When I said that was worth publishing on my blog, it proceeded to muddy it up, even trying to write in my voice. I insisted that the list it came up with was perfect.
Feel free to steal these bullet points and feed them to your ChatGPT. I think it can be made to work much better for us humans. ;-)
# Tue, 08 Jul 2025 21:20:45 GMT ## On Bluesky: "The web is still there under all this michegas, ready to be magical again any time we want it to." # Tue, 08 Jul 2025 13:34:08 GMT ## New thread on Mastodon: OK, i give up for now on getting a team of nerds together to build a bridge from RSS to ActivityPub along the lines of what Automattic has done to bridge WordPress to ActivityPub. We will need that to happen, I would much prefer to get it done in advance, but people don't know me or trust me well enough to believe I might see something that they apparently don't. I'm pretty confident they will, but I would really love to get some help. # Tue, 08 Jul 2025 12:42:10 GMT ## The topology of social networks -I wrote this on Bluesky this morning.
Hardly the first time I've said this, but this time I got a response.
That was from John Pettus. I could tell right off that we're thinking the same way. This morning I started to write a reply but quickly ran out of space because of Bluesky's stupid character limit. So I just pasted it into this blog post.
Yesterday I wondered if the open web is a lost cause.
A few minutes later, I saw my name in a tweet on Bluesky from Aram Zucker-Scharff.
In it was a message that can be summarized as follows -- don't give up yet Dave.
AZS has a linkblog which he calls an amplifeed. Same thing.
And here's the best part.
Even though we were working separately -- our feeds are 100% compatible.
When I saw it I subscribed to it in FeedLand and added it to my blogroll.
It will work in my timeline software (still working on it).
This gave me goosebumps.
I remember what this felt like.
Working on something and someone else working on the same thing and because we're on the mother freaking web our stuff works the same way.
That my friends is what the web feels like. Goosebumps. Power. Interop. This is what most people who use the net these days have never experienced.
I used to write about this on my blog every day. I would say things that annoyed some of my readers like this: zoooooom and coooooool. There he goes again. Hey it's been a while.
I celebrated this with a suggestion to AZS.
He did it in a minute.
So the web isn't a lost cause after all. 😄
This the web. It's what the Dead called Truckin.
Let's do more of this.
PS: AZS sent me a link to another linkblog, which I have subscribed to and added to my blogroll.
# Tue, 08 Jul 2025 16:26:03 GMT ## I hate CSS -There I said it. To ChatGPT.
Wondering what it would say...
I wish I had written that.
# Mon, 07 Jul 2025 19:43:35 GMT ## Open + web == lost cause? -A longish thread that probably isn't going anywhere. My final thoughts, cc'd here to get on the record.
anyway it's feeling like a lost cause.
to be really blunt, i don't think AP or ATP are the answer.
and i also don't care so much about this style of conversation. and i loathe the character limits and the lack of style and links, and no titles, etc.
it can't only be for wordpress. i love the potential of wordpress, i think even more than matt does, but it isn't enough.
if it's going to be open and of the web it has to be simple and easy, and neither of them are.
bonus for blog readers: i would add, since i ran out of characters on masto, that the great thing about the web is that you can have an idea and be using it the next day. you can't say imho that you're part of the web if you don't deliver that kind of ease of access. it's not enough to have the potential of being open, it has to be accessible. I have that ability these days, but people who use the AP and ATP systems are in tight little boxes with no easy way to try something out quickly. (i know because i've been hooking things up to them for a couple of years now, and so far it's just an added slog, everything is far more complex than it should be)
maybe we'll get there through their api's, but i think at this point we know that won't happen.
ps: the web is a miracle. but maybe it's too fucked up now to have the miracle be something we can all experience.
pps: when i write on other systems i often leave out upper case, saving a little energy as i type. i find it more relaxing.
# Mon, 07 Jul 2025 19:34:41 GMT ## Linkblogging back to normal -I wrote this early this morning as a test post for my WordLand site.
Happy to report that my linkblog routine is back to normal.
I really shook things up there, and it probably wasn't a great time investment.
I had been using a custom front-end to FeedLand, which has a built in blogging tool, that publishes to the database that FeedLand manages, and of course also publishes an RSS feed. It was debugged and works. But now I have a new editor, and I want to use it for this, because my reader knows how to view all kinds of stuff, and one of the things I wanted it to work well with are linkblog posts. So, do a quick addition of linkblog stuff to WordLand.
Only thing is there is no such thing as a "quick addition" in a world built on CSS and HTML objects. Everything is a slog.
Anyway the slog is over! Whew.
Now back to my other slog -- timelines.
It's also starting to feel usable. People imagine that you just design something and write the code and voila it's usable (if they even think about it that much). But only until you have the pieces put together can you see the things you forgot to consider, and now you have to decide whether to rip up the thing you built or try to iterate to where you need to go. A lot of times it would be easier to start over, but programmers always want to do that. I'm no exception. Once it's working somewhat the code becomes locked into how the pieces fit together. If somehow they need to fit together differently, given it's CSS and HTML you'd better scrap it and do it again or you'll go out of your mind adding the next layer of features.
Honestly we were much better off before we tried to shoehorn an object model into a document format! Apps and documents are really different things you know.
Anyway now I have my first test post of the day.
# Sun, 06 Jul 2025 15:42:11 GMT ##I've been playing a little game, trying to answer the question -- if I had a modern implementation of Frontier that ran on Linux and new Macs, just as it was in 1992 when we released it for the pre-OS/X Mac, what apps would I want to hook up to it right away? What would the verb set look like?
I'd start with the native verb set we had in Frontier for accessing the file system. And HTTP verbs of course.
Then I would add glue for WordPress, GitHub, Mastodon and Bluesky, just because I think having really simple scripting for each of those would make (some) people's brains explode.
I once had a young fellow challenge me on whether there was such a thing as scriptable apps. I was reminded of the days when I had to explain it but no one got it, then one day everyone got it as if they always did, and now we're back at the beginning again. There is such a thing. You can think of an app as a toolkit. What's behind the UI? Let me call it from outside your app. Let me combine the features of your product with other people's product. And you can do the exact same thing for apps that are running on the web. It was something a lot of people tried to do, like Magic Cap at General Magic, but we got it working and had regular nerds writing apps as if it was not amazing. It was, and it's now a long lost art.
If a version of Frontier came up that I could run on a Linux system, I would wish for a really simple interface to Node packages. I've got a great collection. I'd want to use them right away asap.
I also would like to be able to write code in Frontier in JavaScript. I'm very fluent in it these days. I can still program in UserTalk, the two languages are basically the same thing, though UserTalk has some nice affordances they haven't thought of yet in JavaScriptLand, and vice versa -- there are even more things JS can do that we hadn't thought of, which is only fair, they've been working on it a lot longer than we did. The language was basically frozen in the late 90s, and the verb set shortly after that.
Oh what would I do? It's fun to dream.
# Sat, 05 Jul 2025 17:46:48 GMT ## WordLand v0.5.19 -- Lots of little fixes. # Sat, 05 Jul 2025 17:47:31 GMT ## An improvement in WordLand on the server, we now post metadata to WordPress, along with the HTML rendering so that code that runs on the server can now access and possibly in the future even talk back to WordLand. You never know where this stuff can go if the developers take advantage of opportunities to interop.