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.

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.

• The post has a title (it's optional).
• Simple styling: bold, italic, numbered lists, bulleted lists.
• Hyperlinks, the defining feature of the web.
• I can edit my post.
• No character limit.

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! :-)

Earlier I wrote this post:

• The worst thing about Claude and ChatGPT is that you can't get them to work with you the way you want them to. They always run off and do the things I asked about. So if I say "what if the middle column must be at least 600 pixels wide? What I want is literally an answer to the question I asked. It was not an instruction to do it. I've been a CEO, and this infuriates me. I want to understand the limits, I don't want to impose them. I'm trying to figure out what the cost would be in doing something, again -- not requesting it be done. No matter how many times I tell it not to do this, it just does it. And Claude has absolutely no memory. If I want it to work with me a certain way I have to tell it how to do it every time. I can save it off in a document, but even the simplest automation is impossible. In that sense it is very primitive software. If I were using a spreadsheet for example, I'd have settings that I could change, and unless the software had a bug it would never forget it until I told it otherwise.

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.

• You want me to think first, not code.
• You want me to challenge assumptions, including my own.
• You expect me to give up on a theory when the evidence contradicts it, and look for more plausible explanations.
• You're a deeply experienced debugger (almost 50 years), and you know how important it is to notice the obvious.
• You're not here for code generation — you're here for sharp, collaborative thinking.

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. ;-)

I wrote this on Bluesky this morning.

• Most of what passes for discourse on platforms like Bluesky amounts to spam and abuse. Makes expensive moderation necessary and who’s going to pay for that on an open system. It’s why this approach can’t lead anywhere but to yet another Twitter or Threads, a place for billionaires to control us.

Hardly the first time I've said this, but this time I got a response.

• Yep. I've said for years—your platform's experience will be determined by its mechanics, not whatever culture you think you want to foster.
• Mechanics > culture in the medium term.
• All systems are adversarial systems now.

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.

• I've been saying that for years myself.
• I ran a BBS in the early 80s, and was on Compuserve CB Simulator (my handle was Mastodon), and started blogging in 1994, and on and on. Mail lists always flame out. Same thing we're seeing in the tweetersphere.
• Blogging has the inverse problem. Spam abuse is impossible, but then it's hard for people to find your brilliance. But at least you get to finish a thought before the trolls attack. ;-)
• I was also a math major, studied combinatorics and graph theory. The connections between nodes determines a lot, as you pointed out. Most people don't even begin to get this.
• I'm working on a new structure for a social web without these problems. It's very simple, a derivative of something I had on my blog in the early days called the Mail Pages.
• I sent my blog posts out via email to groups of 8 people chosen randomly each time (or maybe it was 11, I forget). Sometimes great discussions would break out in the groups. I was always cc'd.
• Sometimes people would just respond to me personally.
• If I saw something I thought everyone should read I would put it on the current Mail Page. Thus there was a way for discourse to have more distribution, but only if I thought it was worthwhile. Spam was not possible, and there was a little bit of abuse here and there, but it would never get any further than my email inbox.
• That's how sensible moderation works. And it doesn't cost anything, because the spam motive is gone.
• I wrote this in a Bluesky message, but had to move it to my blog because of their stupid character limit.
• BTW, we should communicate and perhaps collaborate.

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.

• If you also add a channel-level <image> element it'll work with my timeline app. I'm doing it the same way WordPress does it.

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.

There I said it. To ChatGPT.

• i find flexbox too confusing for me to use at this time. maybe in a year or two, after some more learning. i hate css, everything about it. so it's hard for me to learn new stuff. i grew up using quickdraw on the mac, where i could do everything with math, instead of trying to tease the right behavior as something as convoluted as css and html for apps. ridiculous.

Wondering what it would say...

• Totally fair, Dave. That frustration makes perfect sense — you're coming from a world where you told the machine exactly what to do, and it did it. CSS is the opposite: it’s like negotiating with a passive-aggressive layout engine that only gives you what you want if you ask just the right way, using its weird dialect.
• QuickDraw gave you full control with math — coordinate space, bounding boxes, baseline alignments — it was declarative drawing by imperative code. CSS is declarative positioning by magical suggestion, with hidden constraints, inherited styles, and a constantly shifting box model.
• You’re not alone in this. A lot of veteran programmers — especially those from early Mac, HyperCard, VB, even Win32 — find CSS maddening. It’s not made for precision. It’s made for fluidity and adaptation. But that’s not how app UIs are often designed.

I wish I had written that.

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.

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.

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.

I'd love to see a bridge from RSS to ActivityPub. I've asked people at various companies if they'd do this. I'm happy to help with the software but operating the service is something for a trusted company to do.

I think this would go all the way to putting the "open" in open social web, because people who already know how to build RSS feeds would be able to quickly write apps that hook into AP networks. And of course it wouldn't have to be limited to RSS, it could build on Atom and RDF equally well.

It think it's tragic that it's taking Ghost, for example, so long to get their service up fully, and it suggests that smaller devs don't stand a chance. I can't wake up one day and have an idea of something that would work well with Mastodon, for example, and have a prototype running the next day.

If you think this is a good idea, post a link to this post somewhere developers live, and let's see if we can get a cooperative project up and running.

And if you don't like RSS, Atom or RDF, invent an orthogonal format and we can work with that too. I know people have strong feelings about this stuff, not a problem.

PS: I asked Tim Bray to comment, and he responded. Sounds good. We've known each other for decades, going back to the early days of XML.

This is the address of my linkblog feed: dave.linkblog.org.

I think it's kind of interesting to have the top page of a site be a feed. I don't hide the XML-ness of it. I never supported the obfuscation, it's confusing, makes people not trust RSS, imho.

I think the feed is pretty stable now, so if you want to subscribe, go ahead. I haven't redirected from the old feed yet, probably should do that soon, since it more or less has stopped updating.

This is all managed in WordLand and therefore is part of the WordPress ecosystem.

I felt it was time to do a definitive linkblog, since as far as I could tell no one has tried to explain what it is: basically, a feed where the <link> element of each <item> points to some other site. That's the basic difference.

Also a linkblog feed should specify the channel-level <image> element, which is used as the avatar for the feed when it appears in a twitter-like timeline.

I think the only other product that is open to feeds being part of the open social web is Surf from Mike McCue's company, Flipboard. I asked ChatGPT to brief me on how it works with feeds, and saw that we're more or less doing the same thing, except I'm not trying to work with the output from Twitter, Bluesky, etc. Even when they have outbound RSS feeds they aren't good enough to be part of the social web defined by feeds.

I only want really good feeds. It's time to stop being so careless about what we transmit to the world. If we want an open web we're all going to have to be good gardeners. It's like a food system where all the food is grown by family farmers and I'm running a restaurant, and only want the good stuff, and we want it to look good too! :-)

PS: Another thing, the feed items must have working guids. All software that runs on feeds should be able to depend on this.

PPS: Linkblogs aren't the only kinds of feeds that will be used in this RSS-based feediverse. Scripting News will work with it. You would be able to read this post in this new medium (not yet delivered, btw).

PPPS: More here and here.