With the release of various interesting text generation tools, I’m starting an experiment this and next month.

I will be posting computer generated text, prompted by my own current interests, to a separate blog and Mastodon account. For two months I will explore how such generated texts may create interaction or not with and between people, and how that feels.

There are several things that interest me.

I currently experience generated texts as often bland, as flat planes of text not hinting at any richness of experience of the author lying behind it. The texts are fully self contained, don’t acknowledge a world outside of it, let alone incorporate facets of that world within itself. In a previous posting I dubbed it an absence of ‘proof of work’.

Looking at human agency and social media dynamics, asymmetries often take agency away. It is many orders of magnitude easier to (auto)post disinformation or troll than it is for individuals to guard and defend against. Generated texts seem to introduce new asymmetries: it is much cheaper to generate reams of text and share them, than it is in terms of attention and reading for an individual person to determine if they are actually engaging with someone and intentionally expressed meaning, or are confronted with a type of output where only the prompt that created it held human intention.

If we interact with a generated text by ourselves, does that convey meaning or learning? If annotation is conversation, what does annotating generated texts mean to us? If multiple annotators interact with eachother, does new meaning emerge, does meaning shift?

Can computer generated texts be useful or meaningful objects of sociality?

Right after I came up with this, my Mastodon timeline passed me this post by Jeff Jarvis, which seems to be a good example of things to explore:


I posted this imperfect answer from GPTchat and now folks are arguing with it.

Jeff Jarvis

My computer generated counterpart in this experiment is Artslyz Not (which is me and my name, having stepped through the looking glass). Artslyz Not has a blog, and a Mastodon account. Two computer generated images show us working together and posing together for an avatar.


The generated image of a person and a humanoid robot writing texts


The generated avatar image for the Mastodon account

Peter has experimented for a while with Mastodon (and the ActivityPub protocol behind it) and decided that it’s not for him.

Well, this has been fun, but it turns out that the effort-vs-reward for the fediverse doesn’t balance for me; I need fewer reasons to be tethered, not more. @mastohost, recommended by @ton, was an excellent playground. In 24 hours this account will self-destruct. But, now and forever, https://ruk.ca is where you’ll find me.

I very much recognise his point. The disbalance he mentions I felt strongly in the past month, where it was absent in the five and a half years before it. The enormous influx of people, positive in itself, and the resulting growth in the number of people I followed made my timeline too busy. In response I started following topics more and am evaluating rss feeds from ActivityPub servers. The disbalance expresses itself in spending too much time in the home timeline, without that resulting in notable things. (I mean literally notable, as in taking notes) Unlike my feedreader. It does result in some interesting conversations. However such interactions usually start from a blogpost that I share. Because of the newness of AP and Mastodon to the large wave of people joining, many posts including mine are of the ‘Using Mastodon to talk about Mastodon’ type. This is of course common for newly adopted tools, and I still have a category on this blog for metablogging, as blogging about blogging has been a 20 year long pattern here. Yet it is also tiring because it is mostly noise, including the whole kindergarten level discussions between petty admins defederating each other. There’s a very serious discussion to be had about moderation, blocks and defederation, to turn it into a tool that provides agency to individual users and the groups they are part of. These tools are important, and I’m glad I have them at my disposal. Ironically such serious discussion about Mastodon isn’t easy to conduct in a Tweetdeck and Twitter style interface, such as Mastodon provides. I moved the home timeline over to the right in my Mastodon web interface, so I don’t see it as the first thing when I open it up. I’ve concluded I need to step away from timeline overwhelm. Much as I did on Twitter years ago.


A tired purple mastodont lies on the ground sleeping while groups of people are talking in the background, sketchbook style. Dall-E generated image.

There are however two distinct aspects about AP and the recent incoming wave of people that I am more interested to be engaged with than I was before this started.

First, to experiment personally with AP itself, and if possible with the less known Activities that AP could support, e.g. travel and check-ins. This as an extension of my personal site in areas that WordPress, OPML and RSS currently can’t provide to me. This increases my own agency, by adding affordances to my site. This in time may mean I won’t be hosting or self-hosting my personal Mastodon instance. (See my current fediverse activities)

Second, to volunteer for governance related topics in the wider Dutch user group of Mastodon. Regardless of my own use of Mastodon, it is an environment in which many more people than before have new choices to make w.r.t. taking their online presence and tools in their own hands. A step from a global silo such as Twitter to e.g. a larger Dutch instance, while not the same as running one’s own, can be a significant step to more personal agency and networked agency. I’m involved in a group discussing how to establish governance structures that can provide continuity to the Dutch instance Mastodon.nl, lets people on the instance have an active voice and role in its internal governance, and raises awareness of the variety of tools and possibilites out there while purposefully avoiding becoming a new silo (through e.g. providing pathways away from the instance). Such governance is not part of the Mastodon instance, but structured around it. Such involvement is an expression of my experience and role in using tech for the past 33 years online as being inherently political.


A purple mastodont is conversing with a crowd of people, sketchbook style. Dall-E generated image.

Today I heard the EU High Value Data list in its first iteration is finally decided upon. In September 2020 we submitted our advice on what data to include in the thematic areas of geographic data, statistics, mobility, company information, meteorology, earth observation and environment. Last week the Member States submitted their final yes/no vote, and the final text was approved. The EC will now finalise the text for publication, and it should be published before the end of the year. It will enter into force 20 days after publication and government data holders have 16 months until April/May 2024 to ensure compliance. It’s been a long path, and this first list could have been better concerning company information. Yet, when it comes to geographic data (addresses, buildings, land parcels, topography), meteorology and that same company information, it draws a line under two decades of discussion, court cases and studies to help dismantle the revenue model of charging at the point of use. Such charges are a threshold to market entry, and are generally lower than the tax revenue otherwise gained from the activities it’s a threshold to.

It’s easy to just move ahead and think about how this is not enough, what still needs doing, how to implement this etc. But it’s good to acknowledge that when I first started working on open government data in 2008 I heard the stories of those who had been at it for many years since well before the first PSI Directive was agreed in 2003. Some of those people have by now been retired for quite some time already, and I worked on it standing on their shoulders. The implementation act for EU high value data sets is a big step, even if in the field we thought it a no-brainer for decades already.

De Nederlandse overheid gaat mogelijk van Facebook af. In Duitsland is het overheidsgebruik van Facebook door de Duitse autoriteit persoonsgegevens stilgelegd in 2021, omdat Meta zich (uiteraard) niet aan de AVG houdt. De AP doet dat soort uitspraken sinds 2018 niet meer zelf, omdat Facebook in Ierland is gevestigd. De Duitse instantie heeft dat soort consideratie niet, en gaat uit van een eigen verantwoordelijkheid. In plaats van veel te lang wachten op een Iers oordeel over Facebook, steekt ze de hand in eigen boezem en stelt dat de eigen overheid in ieder geval niet aan de eigen AVG verplichtingen kan voldoen door Facebook pages te hebben.

Staatssecretaris van Digitalisering Alexandra van Huffelen bereidt nu mogelijk een besluit voor langs dezelfde lijnen. Terecht lijkt me. Meta houdt zich zelf niet aan de AVG, en bovendien is de algemene uitwisseling van Europese persoonsgegevens met de VS geheel niet juridisch gedekt op dit moment.

De overheid moet zelf het goede voorbeeld geven bij online interactie met burgers en de omgang met eigen gegevens. Dit geldt voor Meta, voor Twitter, maar ook voor cloud diensten en de Microsoft lock-in waar de overheid zich grotendeels in bevindt. Facebook zelf niet meer gebruiken is een bescheiden eerste signaal, dat al verrassend lastig lijkt voor de overheid om helder af te geven.

Ik hoop dat de staatssecretaris de knoop snel doorhakt.

Ein datenschutzkonformer Betrieb einer Facebook-Fanpage sei nicht möglich, schrieb Kelber in einem Brief an alle Bundesministerien und obersten Bundesbehörden.

Bundesdatenschutzbeauftragte

Today at 14:07 twenty years ago, I posted my first blog post. Well over 3000 posts later, this blog has been an integral part of quite a stretch of my life, to the point where it is unavoidable that if you’ve read along you now probably know more about me than I think I’ve actually shared in writing.

In the past few years I’ve taken this blog’s anniversary as a moment to reflect on some of my blogging practices. That yearly reflection started 5 years ago when I was just leaving Facebook. This time it coincides with #twittermigration, where many people are exploring federated options now that Musk has taken over Twitter. Whether that is something that will stick is uncertain of course, but it is interesting to watch playing out. Other earlier such reflections: 2018, 2019, 2020, 2021.

Last year I wrote:

For the coming time this note-to-blog pipeline, and making it easier for myself to post, will be my area of attention I think. Let’s see next year around this time, when I hit the two decade mark with this blog, how that went.

Indeed, that is exactly what I did from early this year: ensure that I could post directly to this blog in different ways. The key to that was create a Micropub client, which posts to this site. Once I had that I could create different paths to feed a post to that Micropub client. From inside a feedreader, directly from my notes in Obsidian, or through a simple web form. More recently I created different versions of that web form, to also post check-ins, and announce travel plans. In all fairness, my habits in how I post things haven’t fundamentally changed yet: I’m writing this in the WordPress back-end. But increasingly I am using those other paths to get content into this site.
Making it easier to post, puts the friction of blogging where it needs to be: wanting to write something.

Connecting things up into flows, blurring the lines between my site, online interaction and my notes for instance, stays an interesting thing to experiment with. In the past months I started using Hypothes.is more intensively, to annotate things I read on the web. Already all those annotations seamlessly end up in my local notes, from where I can work with them, and where they concern my own site I’ve made them visible here.

But most of all, aside from all the more nerdy things of tweaking this site and my information flows, this blog has been a source of conversation for twenty years now. It was my original hope, and my ongoing motivation to keep blogging.

Which brings me back to the earlier mentioned #twittermigration. Musk declared the bird is freed, but it seems quite a few people think the bird was caught and rather take wing on their own. Quite a few of those are the people I early on conversed with through their blogs too. If there’s a key difference between ActivityPub/Mastodon and Twitter, it’s that the federated version only ‘works’ if you actually interact with other people. Likes don’t matter in highlighting a message. Boosts do only share a message with your own followers, and has no other effect. It doesn’t mean it will be put higher in the timeline of others, it’s all in the now. There’s no amplification. Conversation is the key, if you interact then others may also see it and join the conversation. Twitter used to be like that too.

Conversation is key, and that is why I blog.
Here’s to another year of blogging and conversation.

As the Internet is alive with the sounds of #twittermigration these past days, I returned to some earlier thoughts and ideas, w.r.t to both self-hosting fediverse instances, and mapping those on to the business network of my company.

The resulting question is, would a set-up like this work?

If our company would set-up their own fediverse instance (m.tgl.eu here, with accounts for our team). This gives all of our team a ‘verified’, because of the company url, presence as part of their current work. That doesn’t mean we can’t have other accounts (see @ton in the image). And others in our network would do the same (names of organisations for illustration purposes).
If we would run one instance together (samenhankelijk.nl here), that is a relay for all the instances of the organisations involved, and the instance for any individuals in our network (@w… here).
Then we would have a fediverse network of our company’s actual network, where it becomes easier to interact more frequently across the entire network, where discovery is possible because of the shared public timelines through the relay. It’s bounded by being a representation of an actual network, but open within that and based on the permissive boundaries the various organisations themselves have.

I’m not sure if this is how ActivityPub relays are meant to work or are useful, but that’s what I want to explore.
A few of those building blocks are easy to set-up, a company instance and the instance to function as relay. Others are harder, getting our own instance used (we have internal asynchronous interaction through our own rocket.chat instance), getting others in our network to take the same steps.

Notions that play a role in this

My company is part of a network of similar groups and initiatives. Internally we call them friends of our company. These are the people and organisations we invite to events and parties, that we like to hang out with, jam about ideas with, and when possible work together with. That can be because we worked together in the past and thought that was fun and worth repeating, or because we share or shared office space, have similar perspectives or visions, and having overlapping or complementing activities. It’s a network of individuals in larger organisations that we interact with individually, and companies, non-profits and NGO’s that are Zebra’s, like us.

I think that technology should be smaller than us, in order to provide agency to us. With smaller I mean that the deployment and daily use of a tool must fall within the control and capabilities of the user or user group. Specifically the off-switch should be in control of the user group itself. That way a user group can use a tool under their control to address issues that group has by themselves in their own context. This is what I call networked agency. Different groups can strengthen their tools and work, by networking with other groups, yet tools stay useful on their own and get more useful when connected.

I also think that human networks of connections are similar to the structure of peer-to-peer internet structures. A network of many smaller nodes and areas where those connections are denser, individual nodes that are more intensively connected to others and form a local center. I’m convinced our digital tools work better if they deliberately mimic that human network structure, so that the digital affordances those tools provide flow naturally into the human network connections we all have. That’s what I call human digital networks, and distributed digital transformation.
Openness is a necessity in the networked age. But it also needs a limit. That limit is tied to our personal limits, the way we need to feel ‘at home’ in the context in which we exchange ideas. With the new influx of many new people on Mastodon I noticed how my timeline is feeling more alienating than before when it was more like hanging out in my favourite watering hole in town. That will settle, I’m sure, yet in social platforms that treat the entire globe as the same public square you are continuously exposed to the algorithmically amplified onslaught of all of it all the time. Which does not reflect human network reality anymore. Bounded openness matches that reality better.

All this maps on to the fediverse I think: if each company or group in our network has their own instance, that allows internal interaction and public interaction in parallel, and if that public interaction is always visible locally in all other instance in the network, then more direct and deeper ties between the people in the network may grow. Such interaction would create more ideas, more initiatives and help spot more opportunities to do things together I think (or equally quickly expose we’re not as nicely aligned or matched as we thought).