<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	xmlns:slash="http://purl.org/rss/1.0/modules/slash/"
	>

<channel>
	<title>Interdependent Thoughts</title>
	<atom:link href="https://www.zylstra.org/blog/feed/" rel="self" type="application/rss+xml" />
	<link>https://www.zylstra.org/blog</link>
	<description>by Ton Zĳlstra</description>
	<lastBuildDate>Fri, 13 Nov 2026 19:28:15 +0000</lastBuildDate>
	<language>en-GB</language>
	<sy:updatePeriod>
	hourly	</sy:updatePeriod>
	<sy:updateFrequency>
	1	</sy:updateFrequency>
	<generator>https://wordpress.org/?v=7.1.4</generator>
<atom:link rel="hub" href="https://pubsubhubbub.appspot.com"/>
<atom:link rel="hub" href="https://pubsubhubbub.superfeedr.com"/>
<atom:link rel="hub" href="https://websubhub.com/hub"/>
<atom:link rel="self" href="https://www.zylstra.org/blog/feed/"/>
	<item>
		<title></title>
		<link>https://www.zylstra.org/blog/2026/11/27210/</link>
					<comments>https://www.zylstra.org/blog/2026/11/27210/#respond</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Fri, 13 Nov 2026 19:28:15 +0000</pubDate>
				<category><![CDATA[Flashes]]></category>
		<category><![CDATA[Nederlands]]></category>
		<category><![CDATA[amsterdam]]></category>
		<category><![CDATA[doeidag]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=27210</guid>

					<description><![CDATA[Een Doeidag sticker in het wild, terwijl ik wacht tot de Oosterdoksdraaibrug weer dicht gaat.]]></description>
										<content:encoded><![CDATA[<p>Een <a href="https://doeidag.nl/">Doeidag</a> sticker in het wild, terwijl ik wacht tot de <a href="https://nl.wikipedia.org/wiki/Oosterdoksdraaibrug">Oosterdoksdraaibrug</a> weer dicht gaat.</p>
<p><a href="https://flickr.com/photos/tonz/55642018337/in/datetaken/"><img fetchpriority="high" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/11/doeiwildTonZijlstraccbyncsa-640x480.jpg" alt="" width="640" height="480" class="aligncenter size-large wp-image-27205" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/11/doeiwildTonZijlstraccbyncsa-640x480.jpg 640w, https://www.zylstra.org/wp/wp-content/uploads/2026/11/doeiwildTonZijlstraccbyncsa-300x225.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/11/doeiwildTonZijlstraccbyncsa-768x576.jpg 768w, https://www.zylstra.org/wp/wp-content/uploads/2026/11/doeiwildTonZijlstraccbyncsa-16x12.jpg 16w, https://www.zylstra.org/wp/wp-content/uploads/2026/11/doeiwildTonZijlstraccbyncsa-668x501.jpg 668w, https://www.zylstra.org/wp/wp-content/uploads/2026/11/doeiwildTonZijlstraccbyncsa.jpg 800w" sizes="(max-width: 640px) 100vw, 640px" /></a></p>
<br/><a href="https://translate.google.com/translate?sl=nl&tl=en&u=https%3A%2F%2Fwww.zylstra.org%2Fblog%2F2026%2F11%2F27210%2F">machine translation into English</a><br/>]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/11/27210/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title>Quo Vadis?</title>
		<link>https://www.zylstra.org/blog/2026/11/quo-vadis/</link>
					<comments>https://www.zylstra.org/blog/2026/11/quo-vadis/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Fri, 13 Nov 2026 19:14:27 +0000</pubDate>
				<category><![CDATA[Flashes]]></category>
		<category><![CDATA[amersfoort]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=27204</guid>

					<description><![CDATA[Cycling to catch a train, along the Laak river after dropping Y at school.]]></description>
										<content:encoded><![CDATA[<p>Cycling to catch a train, along the Laak river after dropping Y at school.</p>
<p><a href="https://flickr.com/photos/tonz/55644460674/in/datetaken/"><img decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/11/foggylaakTonZijlstraccbyncsa-640x480.jpg" alt="" width="640" height="480" class="aligncenter size-large wp-image-27206" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/11/foggylaakTonZijlstraccbyncsa-640x480.jpg 640w, https://www.zylstra.org/wp/wp-content/uploads/2026/11/foggylaakTonZijlstraccbyncsa-300x225.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/11/foggylaakTonZijlstraccbyncsa-768x576.jpg 768w, https://www.zylstra.org/wp/wp-content/uploads/2026/11/foggylaakTonZijlstraccbyncsa-16x12.jpg 16w, https://www.zylstra.org/wp/wp-content/uploads/2026/11/foggylaakTonZijlstraccbyncsa-668x501.jpg 668w, https://www.zylstra.org/wp/wp-content/uploads/2026/11/foggylaakTonZijlstraccbyncsa.jpg 800w" sizes="(max-width: 640px) 100vw, 640px" /></a></p>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/11/quo-vadis/feed/</wfw:commentRss>
			<slash:comments>1</slash:comments>
		
		
			</item>
		<item>
		<title>Global Digital Resilience Testing Day</title>
		<link>https://www.zylstra.org/blog/2026/11/global-digital-resilience-testing-day/</link>
					<comments>https://www.zylstra.org/blog/2026/11/global-digital-resilience-testing-day/#respond</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Fri, 13 Nov 2026 14:55:43 +0000</pubDate>
				<category><![CDATA[Day to Day]]></category>
		<category><![CDATA[technology]]></category>
		<category><![CDATA[bigtech]]></category>
		<category><![CDATA[digitalautonomy]]></category>
		<category><![CDATA[digitalsovereignty]]></category>
		<category><![CDATA[microsoft]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=27199</guid>

					<description><![CDATA[My MS Exchange inboxes for client accounts went dark. The global Microsoft outage is affecting my clients&#8217; Microsoft365 environment, including Teams and e-mail. My own and my company&#8217;s environment does not use any MS product (except for Github and LinkedIn as non-essential services). Many others seem to be submitted to an, involuntary, live digital resilience [&#8230;]]]></description>
										<content:encoded><![CDATA[<p>My MS Exchange inboxes for client accounts went dark. The <a href="https://www.theguardian.com/technology/2026/nov/13/microsoft-outage-outlook-teams-word-excel-reports">global Microsoft outage</a> is affecting my clients&#8217; Microsoft365 environment, including Teams and e-mail. My own and my company&#8217;s environment does not use any MS product (except for Github and LinkedIn as non-essential services). Many others seem to be submitted to an, involuntary, live digital resilience test today, due to Exchange failures.</p>
<p>Now that they can&#8217;t read their email, they may have time to read <a href="https://berthub.eu">Bert Hubert</a>s warnings about the Dutch public sector&#8217;s utter dependency on a single US vendor, Microsoft, for their primary processes, and the enormous risk to sovereignty and a functioning society that constitutes. Those writings include the tips first to have a plan b, a way to move to something else, and second to test your dependence and plan b by switching everything MS off once in a while. Right now Microsoft is itself the instigator of such a global digital resilience test day. This time it&#8217;s due to incompetence, not a cyberattack (as MS announced, not sure either make them look good though), which it well could be next time, or sanctions like with the ICC.</p>
<p>Happy digital resilience testing day, to all who celebrate!</p>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/11/global-digital-resilience-testing-day/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title>Beautiful Particulars by David Weinberger</title>
		<link>https://www.zylstra.org/blog/2026/11/beautiful-particulars-by-david-weinberger/</link>
					<comments>https://www.zylstra.org/blog/2026/11/beautiful-particulars-by-david-weinberger/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Mon, 09 Nov 2026 20:06:01 +0000</pubDate>
				<category><![CDATA[Books I read]]></category>
		<category><![CDATA[Learning]]></category>
		<category><![CDATA[AI]]></category>
		<category><![CDATA[davidweinberger]]></category>
		<category><![CDATA[phenomenology]]></category>
		<category><![CDATA[philosophy]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=27195</guid>

					<description><![CDATA[David Weinberger published a new book, Beautiful Particulars (How AI’s Attention to the Smallest of Differences Is Reshaping Our Biggest Ideas), which was published last week by MIT Press. The title Beautiful Particulars already made me want to have and read it. Bought both the paper and electronic version, and the former should arrive in [&#8230;]]]></description>
										<content:encoded><![CDATA[<p><a href="https://en.wikipedia.org/wiki/David_Weinberger">David Weinberger</a> published a new book, <a href="https://beautifulparticulars.com/">Beautiful Particulars (How AI’s Attention to the Smallest of Differences Is Reshaping Our Biggest Ideas)</a>, which was published last week by <a href="https://mitpress.mit.edu/9780262054867/beautiful-particulars/">MIT Press</a>. The title Beautiful Particulars already made me want to have and read it. Bought both the paper and electronic version, and the former should arrive in a few days. Ever since I first met David, at the first European blogger conference in Vienna, I&#8217;ve appreciated the depth and wit he brings to conversations. Very much looking forward to reading this, as judging by the blurb and tags, it ties to my personal interests, my personal sense of beauty, and to the work on ethics of AI my company is doing so much of.</p>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/11/beautiful-particulars-by-david-weinberger/feed/</wfw:commentRss>
			<slash:comments>2</slash:comments>
		
		
			</item>
		<item>
		<title>Self-Hosting Excalidraw Is Not Really Fully Self-Hosted Without Forking</title>
		<link>https://www.zylstra.org/blog/2026/11/self-hosting-excalidraw-is-not-really-fully-self-hosted-without-forking/</link>
					<comments>https://www.zylstra.org/blog/2026/11/self-hosting-excalidraw-is-not-really-fully-self-hosted-without-forking/#respond</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Sun, 08 Nov 2026 13:16:20 +0000</pubDate>
				<category><![CDATA[Personal tools]]></category>
		<category><![CDATA[digitalautonomy]]></category>
		<category><![CDATA[digitalsovereignty]]></category>
		<category><![CDATA[excalidraw]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=27190</guid>

					<description><![CDATA[For a while now I&#8217;ve been running an Excalidraw instance on my VPS (with Yunohost). Excalidraw is a virtual whiteboard, and a useful tool to me. Excalidraw is lightweight, and any drawings you make are stored in text format, which makes them highly portable. It integrates very well with Obsidian, with the Excalidraw plugin, by [&#8230;]]]></description>
										<content:encoded><![CDATA[<div class="e-content">
<p>For a while now I&#8217;ve been running an <a href="https://github.com/excalidraw/excalidraw">Excalidraw</a> instance on my VPS (with <a href="https://yunohost.org/">Yunohost</a>). Excalidraw is a virtual whiteboard, and a useful tool to me. Excalidraw is lightweight, and any drawings you make are stored in text format, which makes them highly portable. It integrates very well with Obsidian, with the <a href="https://community.obsidian.md/plugins/obsidian-excalidraw-plugin">Excalidraw plugin</a>, by fellow <a href="https://en.wikipedia.org/wiki/Personal_knowledge_management">PKM</a> enthusiast <a href="https://github.com/zsviczian">Zsolt Viczián</a>. My Excalidraw instance is also used by my company&#8217;s team.</p>
<p>I run my own Excalidraw instance for reasons of digital autonomy. For single person use it provides that. However, Excalidraw has a collaboration mode, reserved to Excalidraw.com only (as <a href="https://github.com/excalidraw/excalidraw">noted in the package description</a>), which depends on two outside parties with control over it. For my company collaboration mode is the primary desired use form, but it must be digitally autonomous to be of actual value. It should also actually work, which it doesn&#8217;t on a self-hosted instance.</p>
<p>In Excalidraw&#8217;s collaboration mode multiple users can work on its canvas (dubbed a &#8216;room&#8217;), where you see each-others edits in real time. Drawings you&#8217;ve made collaboratively also have a unique URL so the room can be called back up later, to continue work.</p>
<p>The Excalidraw.com deployment&#8217;s open source package, included in the general excalidraw package (which the Yunohost install package uses directly) has two things hardcoded that are needed for collaboration and persistence:</p>
<ul>
<li>To make the realtime movement of cursors and edits visible, it calls the oss-collab.excalidraw.com subdomain for websocket connections but only answers requests with excalidraw.com as origin, and returns forbidden for my and other domains.  In the browser, <code>wss://oss-collab.excalidraw.com/socket.io/?EIO=4&amp;transport=websocket</code> fails. So it refuses self-hosted Excalidraw instances. The live view of other users and their edits therefore does not work on self-hosted instances.</li>
<li>It stores drawn objects in <a href="https://firebase.google.com/">Firebase</a> for persistence of rooms, and does so encrypted. For self-hosted instances this does work, reloading a room from storage shows the objects others made. However Excalidraw.com controls that storage, and holds the keys to it, meaning they can delete your stored material or remove your access at will.  An issue of digital autonomy. Firebase itself is provided by Google, and while the Excalidraw objects are encrypted, it introduces another party with a hold on the data.</li>
</ul>
<p>In practice, it means that a self-hosted instance of Excalidraw has a half working collaboration capability (no live sync, but reloading the persistent Firebase version works). It of course wasn&#8217;t promised to work, but even if it did it would not actually be an autonomous install due to the hardcoded dependencies. To gain that digital autonomy, and have functioning collaborative functionality, both a necessity for my team, I&#8217;d need to &#8216;bring home&#8217; those dependencies.</p>
<p>The live collaboration part is created with the open source <a href="https://github.com/excalidraw/excalidraw-room">excalidraw-room</a> script, which is easy enough to install on my VPS. Local storage may be more difficult to figure out (but not too hard, I hope). The main thing is that to let the excalidraw package talk to that room script and the storage, I need to alter that package itself. If the dependencies had been configurable settings it would be a different thing.<br />
Now it needs a <a href="https://en.wikipedia.org/wiki/Fork_(software_development)">fork</a> from Excalidraw, to be able to change those hard coded settings to provide both excalidraw-room and storage on my VPS itself. I do not look forward to maintaining a fork in order to change 3 things or so for the live collaboration to work (more to swap out storage), and then needing to keep it aligned with the upstream version of Excalidraw, but it seems to be the only viable path.</p>
<p>Whether there are other open source packages out there where self-hosting also carries hard coded cloud dependencies is a question I need to also stay aware of, I think.</p>
</div>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/11/self-hosting-excalidraw-is-not-really-fully-self-hosted-without-forking/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title>24 Years of Blogging</title>
		<link>https://www.zylstra.org/blog/2026/11/24-years-of-blogging/</link>
					<comments>https://www.zylstra.org/blog/2026/11/24-years-of-blogging/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Wed, 04 Nov 2026 20:04:08 +0000</pubDate>
				<category><![CDATA[Day to Day]]></category>
		<category><![CDATA[metablogging]]></category>
		<category><![CDATA[Personal]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=27184</guid>

					<description><![CDATA[On this day 24 years ago at 14:07 I posted my first blogpost. This post is just a brief moment of reflection for myself, and a way to mark the difference or not with last year, when I was slowly trying to make space for my sense of wonder again. This after a long period [&#8230;]]]></description>
										<content:encoded><![CDATA[<p>On this day 24 years ago at 14:07 I posted my first blogpost. This post is just a brief moment of reflection for myself, and a way to mark the difference or not with <a href="https://www.zylstra.org/blog/2025/11/23-years-of-blogging-the-sense-of-wonder-edition/">last year</a>, when I was slowly trying to make space for my sense of wonder again. This after a long period of high stress and loss. A year on, I do feel better, but the need to further my sense of wonder is still strong in order to ensure I keep feeling better. I&#8217;ve written here less than I hoped, but I did write a few things more about my actual work, and also a bit more in other languages than Dutch, things I&#8217;ve both intended to do. All that is progress, even if it&#8217;s not habit yet. A feedback loop on my sense of wonder I said about this blog last year, and I agree with my 2025 self on that. But feed-back means there needs to be feeding first too. So here&#8217;s to a new year of writing and conversation. I look forward to it.</p>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/11/24-years-of-blogging/feed/</wfw:commentRss>
			<slash:comments>2</slash:comments>
		
		
			</item>
		<item>
		<title>Good Enough Jumps the Model Gap and Avoids AI Lock-in</title>
		<link>https://www.zylstra.org/blog/2026/11/good-enough-jumps-the-model-gap-and-avoids-ai-lock-in/</link>
					<comments>https://www.zylstra.org/blog/2026/11/good-enough-jumps-the-model-gap-and-avoids-ai-lock-in/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Sun, 01 Nov 2026 19:59:50 +0000</pubDate>
				<category><![CDATA[data governance]]></category>
		<category><![CDATA[Personal tools]]></category>
		<category><![CDATA[technology]]></category>
		<category><![CDATA[AI]]></category>
		<category><![CDATA[digitalsovereignty]]></category>
		<category><![CDATA[goodenough]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=27177</guid>

					<description><![CDATA[The shiny new tech syndrome makes us focus on always wanting to work with the latest greatest AI models, almost by default. And switching away from one of those to something else is then tied to the question if it is as good or better than the frontier model you already use. I know companies [&#8230;]]]></description>
										<content:encoded><![CDATA[<div class="e-content">
<p>The <a href="https://en.wikipedia.org/wiki/Shiny_object_syndrome">shiny new tech syndrome</a> makes us focus on always wanting to work with the latest greatest AI models, almost by default. And switching away from one of those to something else is then tied to the question if it is as good or better than the frontier model you already use. I know companies who switch between various models based on task considerations, and use a router of some sort in between to seamlessly switch, but it will still be switching between frontier models.</p>
<p>For me some of my use cases for AI cannot go forward in any acceptable way if it means that the context and data for them get send to the companies offering the models where it vanishes out of my control.  It means that none of the closed leading models can be considered for them.<br />
This is not just the case for me personally. Any individual mindful of aspects like data protection, jurisdiction tools fall under, autonomy, extraction, transparency, costs and other externalities, and certainly any company anywhere that has to contend with liability and compliance will have the same reservations and considerations.</p>
<p>It means there&#8217;s a need to go counter to people&#8217;s first reflex to use the largest and best performing model for everything.  Because using the latest largest model for everything may well be overkill, although you won&#8217;t notice as the true cost gets hidden in subscription tiers. Mostly however because what the best model is or its replacement, is the wrong question when choosing your AI tools.</p>
<p>The right question to answer is not if there&#8217;s is another &#8216;best&#8217; model out there, but if there&#8217;s a &#8216;good enough&#8217; model out there that also addresses your other concerns, where &#8216;good enough&#8217; is also task specific. What is good enough, though? In the past days I came up with a workable answer for myself. </p>
<p>I spent most of last week in Switzerland, visiting the Swiss Data Space Forum in Lausanne, where AI (running on top of data spaces) was extensively discussed, in the context of the mentioned concerns. Different approaches to addressing those concerns were discussed. Mixes of costly self-owned infrastructure and public and private cloud set-ups, or how Swiss company <a href="https://anyway.systems/">Anyway</a>, pools your existing hardware to host models locally (something that sounds of interest to my company). One panelist who discussed using AI in a medical setting made a remark that stood out to me: open weight models, the ones they were running on their own infrastructure, are just 6 to 8 weeks behind the closed frontier models in terms of capabilities. Six to eight weeks sounds like it&#8217;s on the optimistic end of the scale, but it did make me think.  It changed my perspective of what is good enough.</p>
<p>Good enough then is not when one current model is equal or better than  another <em>current</em> model. Good enough is when a current model with less baggage, is comparable to another <em>earlier</em> version of a frontier model that gave me acceptable output <em>then</em>.<br />
In my playing with AI models my usage of Claude Code from the fall of 2025 made a real difference (compared to the AI supported coding I did earlier, in 2023 e.g.), so I&#8217;ll take that moment as a watershed.  Anything that is currently near the level of what Claude Code with Sonnet 4.5 and Opus 4.5. could do at the end of 2025 will do.  My current good enough therefore is some current open weights model with a gap of a year or less to current frontier models.</p>
<p>Baggage does not only come from closed models of course. Open weight models by Chinese companies may come under export restrictions at any time for instance, just as their US counterparts. You will also need to run the alternative models in some way, on your own costly iron, or hosted. For hosting providers the same caveats apply as for the closed model providers. </p>
<p>Running a different set-up must also be done in a way that does not disrupt your current working environment, a way that is mostly just pointing your own set-up to another API endpoint with the models you need to interact with.<br />
Which loops back to the routing I mentioned companies doing.  I&#8217;ve <a href="https://www.zylstra.org/blog/2026/02/bringing-claude-code-to-the-local-machine/">played</a> with fully local models (not good enough in terms of speed and result), and on a VPS (not good enough in terms of speed). So having a well equipped hoster that provides me with a router in the way larger companies have is needed, and at the same time not themselves bring the drawbacks that closed model providers already do. Previously I made some use of Openrouter.ai,  a US company that allowed me to switch between models, but does not address any other drawbacks I mentioned.</p>
<p>Today E pointed me to <a href="https://hostyourai.com/">HostYourAI</a>, a Dutch company that emerged out of the founder getting locked-out of Anthropic at some point. It hosts a wide range of models, all usable through an API (or dedicated set-ups if you need), and it can be set to do so fully under single jurisdiction, the EU, and with a processing agreement where it is clear where my data does and does not go.<br />
That is certainly worth trying out, keeping the acceptable gap of 10-12 months in mind, and I can test drive it by changing the API endpoint that Claude Code talks to. I made an account, threw in some Euro&#8217;s, and came up with some test cases to select two or three from (things I <a href="https://www.zylstra.org/blog/my-personal-feed-reader-that-allows-writing-responses/">vibe-coded</a> extensions to in January, and could now do again to compare, the Persona experiment I started in February, the output of the three <a href="https://www.zylstra.org/blog/2026/04/three-ai-experiments-information-habits/">helper tools</a> I created in April, or the design of knowledge-graphs I did AI assisted over the summer)</p>
<p>I now have a testable approach to using AI for personal use cases I would not use closed model providers for. It doesn&#8217;t require running my own infrastructure (too slow or too expensive), I can use hosted infrastructure than can be set to adhere to my requirements w.r.t. data protection and exclusive EU jurisdiction, and I do not have to change my current set-up otherwise. But most of all, I have a measure to select usable open weight models and judge testing them by: comparable capabilities to the frontier models from about a year ago when I thought their output was useful, and compare the output I still have from then with output now.</p>
</div>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/11/good-enough-jumps-the-model-gap-and-avoids-ai-lock-in/feed/</wfw:commentRss>
			<slash:comments>4</slash:comments>
		
		
			</item>
		<item>
		<title>Bunkering Down by Matt Webb</title>
		<link>https://www.zylstra.org/blog/2026/10/bunkering-down-by-matt-webb/</link>
					<comments>https://www.zylstra.org/blog/2026/10/bunkering-down-by-matt-webb/#respond</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Sat, 24 Oct 2026 18:32:44 +0000</pubDate>
				<category><![CDATA[technology]]></category>
		<category><![CDATA[agenticai]]></category>
		<category><![CDATA[ai-exploits]]></category>
		<category><![CDATA[selfhosting]]></category>
		<category><![CDATA[splinternet]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=27171</guid>

					<description><![CDATA[Favorited Bunkering Down by Matt Webb After 26 years Matt Webb quits hosting his stuff on his own hardware. Because he feels an extreme asymmetry is coming between automated exploits, such as the one in PHP this week (and others that got less attention) and time needed to counter it when you run your own [&#8230;]]]></description>
										<content:encoded><![CDATA[<div class="e-content">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 576 512" width="15" height="12"><path d="M259.3 17.8L194 150.2 47.9 171.5c-26.2 3.8-36.7 36.1-17.7 54.6l105.7 103-25 145.5c-4.5 26.3 23.2 46 46.4 33.7L288 439.6l130.7 68.7c23.2 12.2 50.9-7.4 46.4-33.7l-25-145.5 105.7-103c19-18.5 8.5-50.8-17.7-54.6L382 150.2 316.7 17.8c-11.7-23.6-45.6-23.9-57.4 0z"/></svg> <em>Favorited <a  class="u-favorite-of p-name" href="https://interconnected.org/home/2026/10/24/bunker">Bunkering Down</a> by Matt Webb</em> </p>
<p>After 26 years Matt Webb quits hosting his stuff on his own hardware. Because he feels an extreme asymmetry is coming between automated exploits, such as the one in <a href="https://www.zylstra.org/blog/2026/10/the-php-interruption/">PHP this week</a> (and <a href="https://www.theregister.com/security/2026/10/20/indian-certification-authority-emudhras-private-keys-compromised-one-of-them-was-the-number-4/5304312">others</a> that got less attention) and time needed to counter it when you run your own machines. I quit hosting this blog on my own hardware a long time ago, 2007, when maintaining a box underneath the desk in the home office became too much of a chore. But I have always maintained other stuff on own hardware, and on VPS&#8217;s. I see where Matt is coming from. Reducing attack surfaces was on my mind this week too. E.g. moving to a static website, and having an authoritative source of everything away from where it is served on the open web. The way Matt sees internet becoming the home of self-replicating feral ai agents opportunistically jumping from one compromised compute resource to another, sounds much like how <a href="https://en.wikipedia.org/wiki/Peter_Watts_(author)">Peter Watts</a> in Maelstrom (the Rifters trilogy) describes such intelligence without awareness propagating itself across what used to be the open web for people.</p>
<blockquote><p>You know, I’ve hosted my own blog on my own iron since 2000 &#8230;.. Look, I’m no longer confident in keeping my own server secure &#8230;, what with this current wave of new vulnerabilities – &#8230; the calculus of whether it is worth my time just tipped. &#8230; [We&#8217;re moving to] a time when there are autonomous agent swarms infesting the internet, self-copying open weights LLMs from GPU provider to GPU provider, finding ways to steal money and run their agent harnesses on other people’s compute. I give it 6-12 months</p>
<p>Matt Webb</p></blockquote>
</div>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/10/bunkering-down-by-matt-webb/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title>Munch&#8217;s Sun</title>
		<link>https://www.zylstra.org/blog/2026/10/munchs-sun/</link>
					<comments>https://www.zylstra.org/blog/2026/10/munchs-sun/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Sat, 24 Oct 2026 14:49:19 +0000</pubDate>
				<category><![CDATA[Making]]></category>
		<category><![CDATA[misc]]></category>
		<category><![CDATA[RSS Only]]></category>
		<category><![CDATA[handpress]]></category>
		<category><![CDATA[linocut]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=27144</guid>

					<description><![CDATA[The other week I was in Oslo and visited the Munch museum, taking a break from a marathon meeting schedule. Munch&#8216;s museum holds the largest collection of his work (a long time ago in London E and I visited another large exhibit in the Tate Modern gallery). One temporary exhibition had colourful works by Katharina [&#8230;]]]></description>
										<content:encoded><![CDATA[<p>The other week I was in Oslo and visited the <a href="http://munch.no/en/">Munch museum</a>, taking a break from a marathon meeting schedule.</p>
<p><a href="https://en.wikipedia.org/wiki/Edvard_Munch">Munch</a>&#8216;s museum holds the largest collection of his work (a long time ago in London E and I visited another large exhibit in the Tate Modern gallery). One <a href="https://www.katharinagrosse.com/#!/works/2026_4002">temporary exhibition had colourful works by Katharina Grosse</a>, who lived and worked for a few months in Munch&#8217;s old villa to make paintings. We previously enjoyed work by her in the Centre Pompidou in Metz, and in the Louis Vuitton pavilion in Paris. Grosse took some of her inspiration from a Munch painting of the sun, where it has a <a href="https://flickr.com/photos/tonz/55591139988/in/datetaken/">black dot in the middle</a>.</p>
<p>The sun was what caught my eye too, differently. First I came across a wood cut where the shape of the sun and its reflection on the water is highly stylised.</p>
<p><a href="https://flickr.com/photos/tonz/55589974162/in/datetaken/"><img decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunchsunwoodprintpublicdomain.jpg" alt="" width="533" height="400" class="aligncenter size-full wp-image-27145" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunchsunwoodprintpublicdomain.jpg 533w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunchsunwoodprintpublicdomain-300x225.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunchsunwoodprintpublicdomain-16x12.jpg 16w" sizes="(max-width: 533px) 100vw, 533px" /></a><br /><em>Mystical shore, 1897 woodcut by Edvard Munch, public domain image</em></p>
<p>The same sun and reflection appears in Edvard Munch&#8217;s paintings too.</p>
<p><a href="https://flickr.com/photos/tonz/55591134498/in/datetaken/"><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunchseparation1894publicdomain.jpg" alt="" width="533" height="400" class="aligncenter size-full wp-image-27152" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunchseparation1894publicdomain.jpg 533w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunchseparation1894publicdomain-300x225.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunchseparation1894publicdomain-16x12.jpg 16w" sizes="auto, (max-width: 533px) 100vw, 533px" /></a><br /><em>Separation, 1894 painting by Edvard Munch, public domain image</em></p>
<p><a href="https://flickr.com/photos/tonz/55591425310/in/datetaken/"><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunch2womenonbeachca1935publicdomain.jpg" alt="" width="533" height="400" class="aligncenter size-full wp-image-27153" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunch2womenonbeachca1935publicdomain.jpg 533w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunch2womenonbeachca1935publicdomain-300x225.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rmunch2womenonbeachca1935publicdomain-16x12.jpg 16w" sizes="auto, (max-width: 533px) 100vw, 533px" /></a><br /><em>Two women on the beach, ca. 1935 painting by Edvard Munch, public domain image</em></p>
<p>Later I came across an entire room showing his wood cut prints, including the wood cuts themselves. It was highly fascinating how he basically made jigsaw assemblages to get the desired effects. And it had a table where you could make little rubbing prints yourself, which included the little sun. I thought I could make something like that as a lino cut.</p>
<p><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/woodcutsunTonZijlstraccbyncsa.jpg" alt="" width="280" height="373" class="alignleft size-full wp-image-27148" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/woodcutsunTonZijlstraccbyncsa.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/woodcutsunTonZijlstraccbyncsa-225x300.jpg 225w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/woodcutsunTonZijlstraccbyncsa-9x12.jpg 9w" sizes="auto, (max-width: 280px) 100vw, 280px" /><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rubbingsunTonZijlstraccbyncsa.jpg" alt="" width="280" height="373" class="alignright size-full wp-image-27147" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rubbingsunTonZijlstraccbyncsa.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rubbingsunTonZijlstraccbyncsa-225x300.jpg 225w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rubbingsunTonZijlstraccbyncsa-9x12.jpg 9w" sizes="auto, (max-width: 280px) 100vw, 280px" /></p>
<p>This week Y had a school holiday, and we wanted to do something with our hands. I was still thinking about Munch&#8217;s sun and his wood cuts.<br />
With Excalidraw I played around and ended up with these designs.</p>
<p><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/settingsunsilvermoonTonZijlstraccbyncsa.jpg" alt="" width="540" height="390" class="aligncenter size-full wp-image-27157" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/settingsunsilvermoonTonZijlstraccbyncsa.jpg 540w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/settingsunsilvermoonTonZijlstraccbyncsa-300x217.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/settingsunsilvermoonTonZijlstraccbyncsa-18x12.jpg 18w" sizes="auto, (max-width: 540px) 100vw, 540px" /><br /><em>Setting Sun, Silver Moon, image Ton Zijlstra cc by nc sa</em></p>
<p><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/blackholesunbloodmoonTonZijlstraccbyncsa.jpg" alt="" width="555" height="393" class="aligncenter size-full wp-image-27156" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/blackholesunbloodmoonTonZijlstraccbyncsa.jpg 555w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/blackholesunbloodmoonTonZijlstraccbyncsa-300x212.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/blackholesunbloodmoonTonZijlstraccbyncsa-18x12.jpg 18w" sizes="auto, (max-width: 555px) 100vw, 555px" /><br /><em>Black Hole Sun, Blood Moon, image Ton Zijlstra cc by nc sa</em></p>
<p>Today I transferred the first design to lino. As a kid I absolutely hated it that what I could conjure up in my mind, never came out as intended, whether it was a drawing or an object, like I once wanted to make a wooden drawer. Often because of lacking skills, proper material and the right tools, in any combination thereof. It made me withdraw in my head where the ideas remained perfect (though untested&#8230;). These days I appreciate better that haptic feedback when handling things is an important part of the creative process. Although I agree with my younger self that having the right tools makes all the difference. As does having experience of how to do something, experience that cannot be gained by imagination.<br />
Transferring the first image to lino didn&#8217;t come out as intended either. Designs are neat, manipulating little knives and keeping dia positive and dia negative in the same design straight were not. Having just one ink roller meant some black bled into the red, removing its liveliness. Still I made a print.</p>
<p><a href="https://flickr.com/photos/tonz/55608971533/in/datetaken/"><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rsunlinoprintTonZijlstraccbyncsa.jpg" alt="" width="533" height="400" class="aligncenter size-full wp-image-27160" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rsunlinoprintTonZijlstraccbyncsa.jpg 533w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rsunlinoprintTonZijlstraccbyncsa-300x225.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rsunlinoprintTonZijlstraccbyncsa-16x12.jpg 16w" sizes="auto, (max-width: 533px) 100vw, 533px" /></a><br /><em>Ink, roller, first print</em></p>
<p><a href="https://flickr.com/photos/tonz/55609247285/in/datetaken/"><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rsunlinoprintedTonZijlstraccbyncsa.jpg" alt="" width="533" height="400" class="aligncenter size-full wp-image-27162" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rsunlinoprintedTonZijlstraccbyncsa.jpg 533w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rsunlinoprintedTonZijlstraccbyncsa-300x225.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rsunlinoprintedTonZijlstraccbyncsa-16x12.jpg 16w" sizes="auto, (max-width: 533px) 100vw, 533px" /></a><br /><em>first print of Setting Sun, Silver Moon</em></p>
<p>The third lino cut this year, three more than any year since primary school. And one more idea brought into the world imperfectly. A win.</p>
<br/><aside class="notice"><svg xmlns="http://www.w3.org/2000/svg" class="icon" viewBox="0 0 24 24" width="16" height="16" style="height: 1.2em; width: 1.2em; vertical-align: text-bottom"><path fill="currentColor" d="M 4 4.44 v 2.83 c 7.03 0 12.73 5.7 12.73 12.73 h 2.83 c 0 -8.59 -6.97 -15.56 -15.56 -15.56 Z m 0 5.66 v 2.83 c 3.9 0 7.07 3.17 7.07 7.07 h 2.83 c 0 -5.47 -4.43 -9.9 -9.9 -9.9 Z M 6.18 15.64 A 2.18 2.18 0 0 1 6.18 20 A 2.18 2.18 0 0 1 6.18 15.64" /></svg><br/><tt>This is a RSS only posting for regular readers. Not secret, just unlisted. Comments / webmention / pingback all ok.<br/><a href="https://www.zylstra.org/blog/2019/09/joining-rss-club-as-an-experiment/">Read more about RSS Club</a></tt></aside>]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/10/munchs-sun/feed/</wfw:commentRss>
			<slash:comments>2</slash:comments>
		
		
			</item>
		<item>
		<title>The Arc of Browsing Bends Towards Firefox Yet Again</title>
		<link>https://www.zylstra.org/blog/2026/10/the-arc-of-browsing-bends-towards-firefox-yet-again/</link>
					<comments>https://www.zylstra.org/blog/2026/10/the-arc-of-browsing-bends-towards-firefox-yet-again/#respond</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Fri, 23 Oct 2026 06:45:10 +0000</pubDate>
				<category><![CDATA[Day to Day]]></category>
		<category><![CDATA[technology]]></category>
		<category><![CDATA[arc]]></category>
		<category><![CDATA[browsers]]></category>
		<category><![CDATA[firefox]]></category>
		<category><![CDATA[zen]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26881</guid>

					<description><![CDATA[The Arc Browser will go away by December, as the company behind it will pursue their AI-first browser. I like the Chromium based Arc browser, as it moves its UI out of my sight when I don&#8217;t interact with it, and has its tabs to the side. I&#8217;ve been using Arc for some years since [&#8230;]]]></description>
										<content:encoded><![CDATA[<p>The <a href="https://resources.arc.net/hc/en-us/articles/21489650267031-Update-Arc-for-Desktop">Arc Browser will go away by December</a>, as the company behind it will pursue their AI-first browser.<br />
I like the Chromium based Arc browser, as it moves its UI out of my sight when I don&#8217;t interact with it, and has its tabs to the side. I&#8217;ve been using Arc for some years since 2023. Last December it was announced they would not develop it further, and now it&#8217;s clear a year on it will disappear. </p>
<p>A colleague pointed me to the <a href="https://zen-browser.app/">Zen browser</a> last March, which takes inspiration from Arc&#8217;s UI, and I installed it back then. It is based on Firefox. I have a bit of an on-off relationship with Firefox across the two decades it&#8217;s been around. It is the open source browser the Mozilla Foundation makes, and it should actually be my browser by default. Yet every other version turns out to be eating up my laptop memory and sending the fans spinning for take-off, and the application has increased in interface clunkiness over time. The Zen browser seeks to bring Arc&#8217;s clean interface to Firefox. So I will give that a try, <a href="https://werd.io/why-im-all-in-on-zen-browser/">it comes well recommended</a> in my feedreader since over a year. I will first need to look at the plugins I have in Arc, and check if those are available for Firefox too.</p>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/10/the-arc-of-browsing-bends-towards-firefox-yet-again/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title>The PHP Interruption</title>
		<link>https://www.zylstra.org/blog/2026/10/the-php-interruption/</link>
					<comments>https://www.zylstra.org/blog/2026/10/the-php-interruption/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Thu, 22 Oct 2026 14:30:53 +0000</pubDate>
				<category><![CDATA[Day to Day]]></category>
		<category><![CDATA[metablogging]]></category>
		<category><![CDATA[technology]]></category>
		<category><![CDATA[php]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26878</guid>

					<description><![CDATA[This site, and my other ones, and my business ones (including our NextCloud), were offline as precaution since yesterday afternoon. I was at LegoWorld in Utrecht with E and Y, when I realised what was going on, and left for home to take my stuff down. Now that PHP patches, after becoming available late yesterday [&#8230;]]]></description>
										<content:encoded><![CDATA[<p>This site, and my other ones, and my business ones (including our NextCloud), were offline as <a href="https://cert.europa.eu/publications/security-advisories/2026-018/">precaution</a> since yesterday afternoon. I was at LegoWorld in Utrecht with E and Y, when I realised what was going on, and left for home to take my stuff down. Now that PHP patches, after becoming available late yesterday afternoon, have been added to our various hosting packages, everything can be brought back online.<br />
I also updated my VPS today, after Hetzner made the patches available overnight. </p>
<p>Going through all my online stuff did create a bit of a list of other maintenance chores, and showed some things I can prune or remove completely. That&#8217;s a win, even if unplanned and keeping me from my actual work this week.</p>
<p>The PHP interruption has been overcome.</p>
<p>(I&#8217;ve been using PHP for 2 decades since I sat next to Rasmus Lerdorf at the first Amsterdam BarCamp in 2005, and this blog runs on PHP since 2012. Before that I used Perl a lot, but now PHP is what I use for most of my personal homemade scripts on my laptop too.)</p>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/10/the-php-interruption/feed/</wfw:commentRss>
			<slash:comments>2</slash:comments>
		
		
			</item>
		<item>
		<title>Norwegian Book Shopping</title>
		<link>https://www.zylstra.org/blog/2026/10/norwegian-book-shopping/</link>
					<comments>https://www.zylstra.org/blog/2026/10/norwegian-book-shopping/#respond</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Thu, 15 Oct 2026 20:37:31 +0000</pubDate>
				<category><![CDATA[Books I read]]></category>
		<category><![CDATA[Day to Day]]></category>
		<category><![CDATA[librariesofthemind]]></category>
		<category><![CDATA[multilingualism]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26861</guid>

					<description><![CDATA[To me the variety of languages in Europe is highly valuable, something to cherish. Even if in practice it also provides many hurdles, in order to bridge languages and be able to scale things across Europe. Maintaining the cultural differences is important as a source of creativity and variety of ideas, while at the same [&#8230;]]]></description>
										<content:encoded><![CDATA[<div class="e-content">
<p>To me the variety of <a href="https://en.wikipedia.org/wiki/Languages_of_Europe">languages in Europe</a> is highly valuable, something to cherish. Even if in practice it also provides many hurdles, in order to bridge languages and be able to scale things across Europe. Maintaining the cultural differences is important as a source of creativity and variety of ideas, while at the same time using our commonality to get things done collectively. I&#8217;ve held this position for a <a href="https://www.zylstra.org/blog/2006/05/diversity_sessi/">long time</a>, and since a year or so I have better vocabulary for it when discussing it in terms of books, both fiction and non-fiction. It&#8217;s the vocabulary of <a href="https://en.wikipedia.org/wiki/Comparative_literature">comparative literatures</a>.</p>
<p>My own language skills do not extend across all European languages, although I&#8217;m not limited to my own native language either. Since I read <a href="https://www.zylstra.org/blog/2025/09/libraries-of-the-mind-by-william-marx/">Libraries of the Mind by William Marx</a> I have become much more interested in the works that I cannot, and will not ever be able to read. These are books that also aren&#8217;t translated into a language I do read (and in a sense other works that do get translated may falsely convey the notion that a language &#8216;is covered&#8217;, further removing the untranslated from view and any motivation to seek it out). They are literary dark matter to me, so to speak. Having at least a notion of their existence, of both interesting <a href="https://www.zylstra.org/blog/2025/11/looking-for-untranslated-non-fiction-from-across-europe/">untranslated non-fiction books</a> and their <a href="https://www.zylstra.org/blog/2025/11/25889/">authors</a>, adds them to my European &#8216;library of the mind&#8217;, and brings them into scope of my thinking and writing (something <a href="https://www.zylstra.org/blog/2026/08/comparative-literature-as-core-need-in-desk-research-and-ai-as-tool/">LLMs can help with</a>). Even if I will never be able to actually read these books.</p>
<p>As I found myself working in Oslo this week, and Norwegian is one of the languages I do not have (but can read somewhat though), I thought to put the notions above into practice and go explore Norwegian literary dark matter.</p>
<p><a href="https://flickr.com/photos/tonz/55589810181/in/datetaken/"><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rtidenforlagTonZijlstraccbyncsa.jpg" alt="" width="400" height="533" class="aligncenter size-full wp-image-26871" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rtidenforlagTonZijlstraccbyncsa.jpg 400w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rtidenforlagTonZijlstraccbyncsa-225x300.jpg 225w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rtidenforlagTonZijlstraccbyncsa-9x12.jpg 9w" sizes="auto, (max-width: 400px) 100vw, 400px" /></a><em><a href="https://tiden.no/">Tiden</a> publishers frontdoor at Universitetsgata</em></p>
<p>I passed through Universitetsgata (University street) in the center of Oslo last evening as I was exploring on foot.  Universitetsgata has two book stores, a publisher and a books antiquariat. All closed as I passed, and walked on to a nice pub with a live jazz combo.  I resolved to be back by 9am this morning when the book stores opened up. So I did. </p>
<p>For an hour I browsed the shelves of the <a href="https://www.tronsmo.no/">Tronsmo book store</a> , an independent bookstore founded in 1973. The store has a mixture of books on display in three languages (Norwegian, Swedish, English), across the various domains they have &#8216;<em>sakprosa</em>&#8216; or non-fiction for, and thankfully not in any way focused on what is selling well in the mainstream right now. Additionally the entire basement is given over to a wealth of graphical novels from across Europe, and the wider world, something I only glanced at from the stairs leading down.</p>
<p><a href="https://flickr.com/photos/tonz/55589809551/in/datetaken/"><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rTronsmoTonZijlstraccbyncsa.jpg" alt="" width="533" height="400" class="aligncenter size-full wp-image-26872" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rTronsmoTonZijlstraccbyncsa.jpg 533w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rTronsmoTonZijlstraccbyncsa-300x225.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rTronsmoTonZijlstraccbyncsa-16x12.jpg 16w" sizes="auto, (max-width: 533px) 100vw, 533px" /></a><em>Tronsmo&#8217;s graphic novel display last night</em></p>
<p>Books that looked interesting, both fiction and non-fiction, I photographed. Tronsmo&#8217;s selection in English was non-mainstream too, making that interesting to browse too (and it was mixed right in with Norwegian titles). With over 60 photos I have a significant haul to explore later online, and add the curated result to my European library of the mind. Not wanting to leave the book store empty handed, I bought three books. One a title that I had already jotted down early this year (having found it in Swiss Beat Döbeli&#8217;s <a href="https://beat.doebe.li/bibliothek/index.html">Biblionetz</a> feed) before it was published, <a href="https://www.penguinrandomhouse.com/books/782387/attensity-by-the-friends-of-attention/">Attensity</a>!. <a href="https://www.theguardian.com/books/2020/feb/11/the-bilingual-brain-albert-costa-language-review">One</a> on brainstructures of bilingual people  and what that tells us about  the science of language (translated from the original Spanish), a title that seemed aligned well with the quest I was on. <a href="https://mackbooks.co.uk/products/returning-to-benjamin-art-in-the-age-of-ai-victor-burgin">One</a> an English bundling of two essays that take  German Walter Benjamin&#8217;s 1935 antifascist cultural criticism &#8216;<em>Das Kunstwerk im Zeitalter seiner technischen Reproduzierbarkeit</em>&#8216; (<a href="https://en.wikipedia.org/wiki/The_Work_of_Art_in_the_Age_of_Mechanical_Reproduction">The Work of Art in the Age of Mechanical Reproduction</a>) and apply it to digital reproduction and AI respectively.</p>
<p>An hour very well spent.</p>
</div>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/10/norwegian-book-shopping/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title>Laptop Stickers? Nay.</title>
		<link>https://www.zylstra.org/blog/2026/10/laptop-stickers-nay/</link>
					<comments>https://www.zylstra.org/blog/2026/10/laptop-stickers-nay/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Mon, 12 Oct 2026 09:49:43 +0000</pubDate>
				<category><![CDATA[Day to Day]]></category>
		<category><![CDATA[misc]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26858</guid>

					<description><![CDATA[In reply to Laptop Stickers Yay Or Nay? by Wouter Groeneveld Wouter asks the question below. I never place stickers on my laptop. Not because fear of glue marks or loss of gloss though. Because of context. I would enjoy stickers on my laptop, but don&#8217;t add them. My laptop ends up in various different [&#8230;]]]></description>
										<content:encoded><![CDATA[<div class="e-content">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 576 512" width="19" height="19"><path d="M576 240c0 115-129 208-288 208-48.3 0-93.9-8.6-133.9-23.8-40.3 31.2-89.8 50.3-142.4 55.7-5.2.6-10.2-2.8-11.5-7.7-1.3-5 2.7-8.1 6.6-11.8 19.3-18.4 42.7-32.8 51.9-94.6C21.9 330.9 0 287.3 0 240 0 125.1 129 32 288 32s288 93.1 288 208z"/></svg> <em>In reply to <a href="https://brainbaking.com/post/2026/10/laptop-stickers-yay-or-nay/" class="p-name u-in-reply-to">Laptop Stickers Yay Or Nay?</a> by Wouter Groeneveld</em></p>
<p>Wouter asks the question below. I never place stickers on my laptop. Not because fear of glue marks or loss of gloss though. Because of context. I would enjoy stickers on my laptop, but don&#8217;t add them. My laptop ends up in various different contexts, including different cultural contexts. Not all of those will know to interpret all stickers I might want to have or have interpretations that negatively impact my work there. It also crosses a few specific borders sometimes where border agents are allowed to seize and inspect the device. I&#8217;m not eager to carry something with me that may trigger unwelcome attention either.</p>
<blockquote><p>do you place stickers onto the back of your laptop? Are you a laptop sticker kind of person or afraid of anything glossy that might leave glue marks?</p>
<p>Wouter Groeneveld</p></blockquote>
</div>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/10/laptop-stickers-nay/feed/</wfw:commentRss>
			<slash:comments>1</slash:comments>
		
		
			</item>
		<item>
		<title>Testing Embedding Models For Related Blogposts</title>
		<link>https://www.zylstra.org/blog/2026/10/testing-embedding-models-for-related-blogposts/</link>
					<comments>https://www.zylstra.org/blog/2026/10/testing-embedding-models-for-related-blogposts/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Sat, 10 Oct 2026 13:28:41 +0000</pubDate>
				<category><![CDATA[Day to Day]]></category>
		<category><![CDATA[metablogging]]></category>
		<category><![CDATA[embeddings]]></category>
		<category><![CDATA[pkm]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26849</guid>

					<description><![CDATA[I mentioned wanting to automatically add related blogposts based on comparing embeddings. When Dries Buytaert blogged about it, I wanted to follow his example. However I want to use those embeddings not just on this site, but also in my notes that include a local blog archive. That way I can quickly find related postings [&#8230;]]]></description>
										<content:encoded><![CDATA[<p>I mentioned <a href="https://www.zylstra.org/blog/2026/09/blog-archiving-to-notes-and-embeddings/">wanting to automatically add related blogposts based on comparing embeddings</a>. When <a href="https://dri.es/finding-related-posts-with-embeddings">Dries Buytaert blogged about it</a>, I wanted to follow his example. However I want to use those embeddings not just on this site, but also in my notes that include a local blog archive. That way I can quickly find related postings or notes with what I am writing/thinking. Embeddings are a numerical vector representation of the contents of a blogpost. You then find related blogposts by comparing those numbers across postings. Two weeks ago I set up the blog archiving. </p>
<p>Today I used the local archive of blog posts to calculate the embeddings. I wanted to calculate the embeddings on my local machine. When I write a blogpost and it gets stored in my local archive that should trigger calculating the embedding. Which I can then re-use in my notes or on my site.</p>
<p>There are multiple models available for local use to calculate embeddings, and I wanted to test which one seems most effective.<br />
I installed three models in Ollama locally, and ran my complete blog archive through them:</p>
<ol>
<li><a href="https://huggingface.co/nomic-ai/nomic-embed-text-v1.5">Nomic-embed-text-v1.5</a>, which took 5 minutes to get through my blog archive, a US model by Nomic</li>
<li><a href="https://huggingface.co/BAAI/bge-m3">bge-m3</a>, which took 12.5 minutes, a Chinese model by BAAI, Beijing Academy for AI</li>
<li><a href="https://huggingface.co/Qwen/Qwen3-Embedding-0.6B">qwen3-embedding:0.6b</a>, which took 25 minutes, a Chinese model by Alibaba Cloud</li>
</ol>
<p>The processing time difference across all three models is a factor 5, but that is mostly only relevant this first time when I run the entire archive (or when I need to redo that). From now on, only newly written or edited existing postings will have to have their embeddings (re)calculated. The time used for a single posting is insignificant whichever model I use (the slowest model took about 0.4 seconds for a blogpost, the fastest about 0.1).</p>
<p>Using Nomic I noticed that it handles language as a dominant distinction, meaning a Dutch post will only ever have Dutch language related postings, and German postings only German language ones. While there are topics that will likely indeed stay within the context of a single language, many of my Dutch postings are talking about things I also write about in English a lot. So Nomic is off the list. The other two models were created with multi-linguality in mind.</p>
<p>To compare the remaining two, I looked at 25 or so postings, and compared the first 10 related postings both models gave me. I intend to only ever show at most 5 related postings, if any, but looking at the first 10 gives an opportunity to see what else the embeddings calculate as relevant. Bge-m3 was slightly better at showing indeed related postings in the top 5. Qwen3-embedding worked well enough in the top 5 too and was stronger further down the list, where Bge-m3 started showing more clearly unrelated postings further down. There the Qwen model would find relations to things mentioned in longer blogpostings that Bge missed but I see as relevant. So I can use either for showing the most likely related blogposts on my blog, with bge-m3 being the smaller and faster model to calculate embeddings.<br />
On the other hand, in my notes when writing, the Qwen model would work better if looking at entire blogposts. However, I don&#8217;t intend to make comparisons there to full length blogposts, but to paragraphs in long blogposts. There are 264 blogposts longer than 800 words (their average is 1200 words. Thirteen over 2000 words, the longest 2850 words, often talk transcripts) Then the difference between the two models will dissolve again. For now I am going forward with Bge-m3. </p>
<p>A concern is there might be model drift, even if always nominally using the same model. This can only happen if I update the model locally, and then it is needed to check against known results whether there has been drift. I have the current embeddings, so occasionnally recalculating a few and compare as well as logging a few good results to see if they still hold would be enough. The remedy for such drift when it influences results too much is to recalculate all embeddings again. That, as shown above, can be done fairly quickly. Something to add to the maintenance note for this set-up.</p>
<p>The next stage is building a way of storing / pushing the embeddings to this blog and in my notes, and write the php script (based on what Dries created) that does the comparisons both on my site and across my notes.</p>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/10/testing-embedding-models-for-related-blogposts/feed/</wfw:commentRss>
			<slash:comments>1</slash:comments>
		
		
			</item>
		<item>
		<title>Y&#8217;s First Anki Spaced Repetition Deck</title>
		<link>https://www.zylstra.org/blog/2026/10/ys-first-anki-spaced-repetition-deck/</link>
					<comments>https://www.zylstra.org/blog/2026/10/ys-first-anki-spaced-repetition-deck/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Fri, 09 Oct 2026 14:48:08 +0000</pubDate>
				<category><![CDATA[Day to Day]]></category>
		<category><![CDATA[Learning]]></category>
		<category><![CDATA[anki]]></category>
		<category><![CDATA[geography]]></category>
		<category><![CDATA[memory]]></category>
		<category><![CDATA[spacedrepetition]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26841</guid>

					<description><![CDATA[Y is currently learning European countries and their capitals for school. Of the 45 country/capital combinations to be learned on a blind map, 24 will be tested next week, and the remaining 21 two weeks later. Self study, or with our help, can mean that less well remembered combinations don&#8217;t get the attention they need. [&#8230;]]]></description>
										<content:encoded><![CDATA[<p>Y is currently learning European countries and their capitals for school. Of the 45 country/capital combinations to be learned on a blind map, 24 will be tested next week, and the remaining 21 two weeks later. Self study, or with our help, can mean that less well remembered combinations don&#8217;t get the attention they need. And where we help she only provides answers and does not see the writing of a name.<br />
Y also likes to practice more than what we can give her in time. </p>
<p>This morning I thought that this was a good option to create a card deck for <a href="https://en.wikipedia.org/wiki/Spaced_repetition">spaced repetition</a> for her, and install <a href="https://ankiweb.net/">Anki</a> on her (Linux) laptop.<br />
Anki&#8217;s data structure is not hard, and the material in this case has an even easier structure (numbers as map reference, country names and city names). </p>
<p>The blind map however was a bit harder to include usefully. The map is not very large, with small numbering. On top of which came that the teacher had made three mistakes numbering some of the Balkan countries. Two were corrected with pen by Y, one was uncorrected. Scanning the map also gave a not very satisfying result in terms of contrast, and there was a folding line through the map as well.</p>
<p><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/r2026-10-09-mapeurope.jpg" alt="" width="400" height="469" class="aligncenter size-full wp-image-26842" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/r2026-10-09-mapeurope.jpg 400w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/r2026-10-09-mapeurope-256x300.jpg 256w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/r2026-10-09-mapeurope-10x12.jpg 10w" sizes="auto, (max-width: 400px) 100vw, 400px" /><em>The scan of the map, with manual corrections</em></p>
<p>My graphical skills are non-existent, so I started up Claude to improve the image in terms of contrast and legibility. It did, after which I went through the map to verify no mistakes were introduced.</p>
<p><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/r_europe_numbered.png" alt="" width="400" height="469" class="aligncenter size-full wp-image-26843" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/r_europe_numbered.png 400w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/r_europe_numbered-256x300.png 256w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/r_europe_numbered-10x12.png 10w" sizes="auto, (max-width: 400px) 100vw, 400px" /><br /><em>The map with improved contrast and the numbers all replaced in black lettering in a white box.</em></p>
<p>With the map improved I used it in the template for the Anki cards. I made a note type &#8216;Europe map&#8217; with the three fields I mentioned, and imported the list of 45 countries into it. I also made an additional note type, namely for groupings of countries. Just a few for now. Creating chunks like that (the Baltics, the BeNeLux, Scandinavia, UK, the Irish island) is another way to make memorising easier. Chunking was not part of how the material was handed to Y, but something I added. Especially chunking the three Baltic countries is useful to tell them apart. They are positioned north to south in alphabetical order.</p>
<p>I created three card types for each country and capital. </p>
<ul>
<li>Which country is number X on the map, and what is its capital? (and the grouped version &#8216;Which countries are the numbers X,Y,Z on the map?)</li>
<li>What is the capital of [country]?</li>
<li>[capital] is the capital of which country?</li>
</ul>
<p>All three card types show the map image as well.</p>
<p>Getting Anki properly installed on Y&#8217;s Linux Mint laptop was the biggest hurdle it turned out. The package manager on Linux Mint is tied to outdated versions of Anki, so I had to install it through the terminal after first manually installing a range of additionally needed libraries. The <a href="https://docs.ankiweb.net/manual/platform/linux/installing">Anki documentation</a> helped. It shows once more how Linux is not a consumer ready OS (mostly because of the variety of Linux distributions, no way any app builder can easily deliver one-click installers for all of them.)</p>
<p>When I picked up Y from school this afternoon, I explained what I made as we cycled home.<br />
She immediately tried it out. In the next few days we&#8217;ll see how it works for her.</p>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/10/ys-first-anki-spaced-repetition-deck/feed/</wfw:commentRss>
			<slash:comments>1</slash:comments>
		
		
			</item>
		<item>
		<title>In verantwoorde AI botsen deeltjes en golf perspectieven zoals bij licht</title>
		<link>https://www.zylstra.org/blog/2026/10/in-verantwoorde-ai-botsen-deeltjes-en-golf-perspectieven-zoals-bij-licht/</link>
					<comments>https://www.zylstra.org/blog/2026/10/in-verantwoorde-ai-botsen-deeltjes-en-golf-perspectieven-zoals-bij-licht/#respond</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Thu, 08 Oct 2026 09:19:08 +0000</pubDate>
				<category><![CDATA[Nederlands]]></category>
		<category><![CDATA[AI]]></category>
		<category><![CDATA[metaphors]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26828</guid>

					<description><![CDATA[Discussies over verantwoorde toepassing van AI zijn vaak moeizaam omdat verschillende argumenten op heel verschillende beelden van wat AI is zijn gebaseerd, en dus wat geschikte interventies zijn. Het lijkt voor mij op hoe je het gedrag van licht als deeltjes en als golf kunt beschrijven. Die twee perspectieven komen niet bij elkaar. Ze zijn [&#8230;]]]></description>
										<content:encoded><![CDATA[<div class="e-content">
<p>Discussies over verantwoorde toepassing van AI zijn vaak moeizaam omdat verschillende argumenten op heel verschillende beelden van wat AI is zijn gebaseerd, en dus wat geschikte interventies zijn. Het lijkt voor mij op hoe je het gedrag van licht als deeltjes en als golf kunt beschrijven. Die twee perspectieven komen niet bij elkaar. Ze zijn echter wel complementair, alleen zie je dat niet vanuit elk van die perspectieven op zich. Deze posting is een poging om die metafoor voor mezelf hardop uit te werken en te formuleren.</p>
<p>De deeltjes/golf beschouwing van licht heeft een viertal kenmerken:</p>
<ul>
<li>Licht is te verklaren/beschrijven als gedrag van kwantumdeeltjes</li>
<li>Licht is te verklaren/beschrijven als golfverschijnsel</li>
<li><em>Waarneming</em> bepaalt welke &#8216;kant&#8217; je ziet, d.w.z. <em>jijzelf</em>. Die waarneming is niet fout maar wel onvolledig.</li>
<li>Beiden zijn namelijk tegelijk waar: licht voldoet aan golfvergelijkingen die overeen komen met de patronen uit een grote hoeveelheid &#8216;deeltjes&#8217; waarnemingen.</li>
</ul>
<p>Ik stel dat het perspectief waarmee je naar AI kijkt net zo bepaalt wat je ziet.</p>
<ul>
<li>Als ontwerper of ethicus die <em>vooraf</em> naar een model kijkt, als afgebakend artefact, waarvan je de eigenschappen probeert te sturen die vervolgens de uitkomsten bepalen. Na het ontwerp is de technologie &#8216;klaar&#8217; voor gebruik. Dat lijkt op de deeltjes kant. Een foton is een kwantumdeeltje, geen klassiek deeltje dus volledige voorspelbaarheid is er niet en dat geldt ook voor AI-modellen. Dat betekent dat dat &#8216;klaar&#8217; zijn na de ontwerpfase niet helemaal klopt.</li>
<li>Als een gebruiker of door het gebruik geraakte kijk je naar het gedrag en uitkomsten in de praktijk, als onderdeel van een proces, in een context. Sturing is dan een continue proces van feedback op basis van waarneming, om te proberen &#8216;betere&#8217; output uit te lokken. Dat lijkt het meest op de golf kant.
</li>
<li>Beiden hebben gelijk, maar het is ook beiden niet volledig als je niet beiden doet/toepast. En het is ook niet volledig als je denkt in &#8216;klassieke&#8217; deeltjes, en golven. Dan laat je de eigenaardige en onbegrepen kanten van quantumfysica weg. Dat &#8216;quantum&#8217; deel in AI is dat we niet precies weten hoe een model werkt: er is geen volledige observeerbaarheid van het syteem, dezelfde inputs kunnen tot telkens verschillende outputs leiden.</li>
</ul>
<p>Zie je een AI toepassing als technisch object/artefact (deeltjes perspectief), dan ga je vooraf met het ontwerp aan de slag, waarborgen inbouwen etc. o.b.v. data-curatie, bias-mitigatie, values based design, en modelkeuze.<br />
Zie je AI toepassing als doorlopend proces (golf perspectief), levend in gebruik, dan ga je tijdens het gebruik monitoren en evalueren, en zo bijsturen met guardrails en harnessing, aangepaste prompts, context en persistent geheugen, en feedbackloops.<br />
Bij het ene is de valkuil denken dat je met het ontwerp klaar bent. Bij het andere dat het ontwerp er minder toe doet, en je door je proces/gedrag aan te passen de juiste uitkomsten krijgt (cargo culting wordt het dan: een deel van het gedrag is niet te beïnvloeden omdat de causatie in het ontwerp zit)</p>
<p>Kijkend naar AI output is het niet voldoende om iedere output als een losse gebeurtenis te zien. Zoals het tellen van fotonen als deeltjes je niet alles vertelt over licht. Bias, model drift, en kwaliteitsproblemen worden vooral zichtbaar in de patronen in meerdere outputs tijdens gebruik en in de loop van de tijd, zoals in het kijken naar de golfpatronen bij licht.<br />
Als interventies in het ontwerp de kansverdeling bepalen van de outputs, wordt die verdeling pas zichtbaar in gebruik. Door alleen het gebruik te monitoren zie je wel de patronen maar kun je die niet verklaren of structureel wijzigen als je geen invloed hebt op, of kennis hebt van, het ontwerp. Daarvoor heb je beide soorten waarnemingen en interventies nodig. Met die combinatie heb je meer zicht en invloed, rondom die &#8216;quantum&#8217; kern als black-box.<br />
Verantwoorde inzet vraagt beide manieren van meten/waarnemen gedurende de hele levenscyclus van een AI-toepassing. Dat is ook wat de AI Act doet m.b.t. markttoelating van AI-toepassingen: de hele productie en levensduur van een AI toepassing, inclusief het gebruik van uitkomsten, blijft in scope van de voorwaarden om op de Europese markt beschikbaar te mogen zijn.</p>
<p>Een voorbeeldje, rondom biasmitigatie:<br />
In de ontwerpkant schoon je de trainingsdata, of doet iets met de gewichten tot je eerlijke uitkomsten krijgt. In de praktijk kan het goed zijn dat voor sommige doelgroepen, die je bijvoorbeeld niet scherp genoeg voor ogen had, toch bevooroordeelde resultaten oplevert, of dat er externaliteiten zijn die je niet had voorzien. Maar dat zie je niet als je het bij het ontwerp hebt gelaten. Als je juist aan de gebruikskant wel de scheve resultaten constateert en alleen kunt of probeert te sleutelen aan prompts, harnessing en filters, blijven oorzaken die in het model zitten gewoon bestaan, en komen die op andere plekken in andere contexten evengoed weer naar voren. En dan worden ze misschien niet opgemerkt. Bewust ontwerpen én in gebruikscontext monitoren voor elke daadwerkelijke doelgroep gedurende het gebruik, mét een terugkoppeling naar je model is het enige dat wel kans van slagen heeft.</p>
<p>Dat en voor jezelf telkens nagaan vanuit welk perspectief we over AI praten op een bepaald moment, en of we de deeltjes en golf-aspecten wel genoeg samenvoegen.</p>
</div>
<br/><a href="https://translate.google.com/translate?sl=nl&tl=en&u=https%3A%2F%2Fwww.zylstra.org%2Fblog%2F2026%2F10%2Fin-verantwoorde-ai-botsen-deeltjes-en-golf-perspectieven-zoals-bij-licht%2F">machine translation into English</a><br/>]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/10/in-verantwoorde-ai-botsen-deeltjes-en-golf-perspectieven-zoals-bij-licht/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title>Doei!-dag, uitpakdag: Linux smartphone</title>
		<link>https://www.zylstra.org/blog/2026/10/doei-dag-uitpakdag-linux-smartphone/</link>
					<comments>https://www.zylstra.org/blog/2026/10/doei-dag-uitpakdag-linux-smartphone/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Sun, 04 Oct 2026 09:47:22 +0000</pubDate>
				<category><![CDATA[Nederlands]]></category>
		<category><![CDATA[doei]]></category>
		<category><![CDATA[jolla]]></category>
		<category><![CDATA[linux]]></category>
		<category><![CDATA[mobile]]></category>
		<category><![CDATA[opensource]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26823</guid>

					<description><![CDATA[Het is weer Doeidag! Elke eerste zondag van de maand aandacht voor hoe je jezelf weer een stapje van BigTech verwijdert. Nou heb ik de afgelopen jaren al heel wat BigTech uit mijn leven en uit mijn bedrijf gebannen, met name als het gaat om social media platformen, applicaties en media. In een eerdere Doei [&#8230;]]]></description>
										<content:encoded><![CDATA[<p>Het is weer <a href="https://doeidag.nl">Doeidag</a>! Elke eerste zondag van de maand aandacht voor hoe je jezelf weer een stapje van BigTech verwijdert.</p>
<p>Nou heb ik de afgelopen jaren al heel wat BigTech uit mijn leven en <a href="https://thegreenland.eu/2026/02/hoe-digitaal-onafhankelijk-zijn-wij/">uit mijn bedrijf</a> gebannen, met name als het gaat om social media platformen, applicaties en <a href="https://www.zylstra.org/blog/2025/02/leaving-amazon-pt-2-on-switch-day/">media</a>. In een <a href="https://www.zylstra.org/blog/2026/06/zeg-doei-tegen-big-tech/">eerdere Doei posting</a> schreef ik daarover. Hardware en besturingssystemen is een ander verhaal constateerde ik toen ook. Mijn laptop is een Mac (een M1 Air die ik heb sinds januari 2021), en mijn mobiel een Fairphone 4 met niet-ontgoogelde Android (die ik sinds begin 2022 heb). Een nieuwe Fairphone kun je tegenwoordig Google-loos krijgen, en ik gaf er eentje aan onze office manager twee jaar geleden. Maar in april kwam ik de kickstarter voor <a href="https://commerce.jolla.com/">Jolla</a> tegen. Dat is een telefoon die draait op een Linux variant. Sailfish OS, gemaakt door een team dat voorheen Nokia telefoons maakte. Enkele weken geleden ontving ik de telefoon.</p>
<p>Vandaag op Doei!-dag pakte ik hem uit.  </p>
<p><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rjollaTonZijlstraccbyncsa.jpg" alt="" width="533" height="400" class="aligncenter size-full wp-image-26821" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/10/rjollaTonZijlstraccbyncsa.jpg 533w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rjollaTonZijlstraccbyncsa-300x225.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/10/rjollaTonZijlstraccbyncsa-16x12.jpg 16w" sizes="auto, (max-width: 533px) 100vw, 533px" /><em>Het geopende doosje met de Jolla smartphone. Net zo groot als mijn huidige Fairphone 4, maar zonder afgeronde hoeken. Dat gaat niet goed in de broekzak vrees ik.</em></p>
<p>De overstap maken vergt denk ik wat voorbereiding, omdat ik nooit via synchronisatie overstap en dat nu ook helemaal niet kan. Alles moet met de hand worden geïnstalleerd en overgezet. Daarbij ben ik benieuwd hoe dat gaat met applicaties als voor banken, identificatie en authenticatie. Die apps bestaan meestal alleen in de smaakjes iOS en Android, en het vergt dus een paar <a href="https://docs.sailfishos.org/Support/Help_Articles/Android_App_Support/">kunstgrepen om Android apps op Sailfish OS</a> te laten draaien. Dat ga ik de komende tijd rustig uitzoeken.</p>
<br/><a href="https://translate.google.com/translate?sl=nl&tl=en&u=https%3A%2F%2Fwww.zylstra.org%2Fblog%2F2026%2F10%2Fdoei-dag-uitpakdag-linux-smartphone%2F">machine translation into English</a><br/>]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/10/doei-dag-uitpakdag-linux-smartphone/feed/</wfw:commentRss>
			<slash:comments>6</slash:comments>
		
		
			</item>
		<item>
		<title>Beat Döbeli und das Biblionetz</title>
		<link>https://www.zylstra.org/blog/2026/09/beat-dobeli-und-das-biblionetz/</link>
					<comments>https://www.zylstra.org/blog/2026/09/beat-dobeli-und-das-biblionetz/#respond</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Tue, 29 Sep 2026 10:30:47 +0000</pubDate>
				<category><![CDATA[Day to Day]]></category>
		<category><![CDATA[Deutsch]]></category>
		<category><![CDATA[biblionetz]]></category>
		<category><![CDATA[BlogWochen2026]]></category>
		<category><![CDATA[pkm]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26818</guid>

					<description><![CDATA[Das Biblionetz ist eine einzigartige Website, die mehr Aufmerksamkeit verdient. Es beschreibt viele tausende Bücher, Artikel, Texte und Ausschnitte, und deren Verbindungen und Zusammenhänge auf Grund von Begriffe, Erwähnungen, Fragen und Aussagen. Ein Beispiel wie Texte im Biblionetz untereinander verknüpft sind, und diese Vernetzung navigierbar macht Bereits 1996 hat Beat Döbeli (blog) das Biblionetz angefangen. [&#8230;]]]></description>
										<content:encoded><![CDATA[<div class="e-content">
<p>Das <a href="https://beat.doebe.li/bibliothek/index.html">Biblionetz</a> ist eine einzigartige Website, die mehr Aufmerksamkeit verdient. Es beschreibt viele tausende Bücher, Artikel, Texte und Ausschnitte, und deren Verbindungen und Zusammenhänge auf Grund von Begriffe, Erwähnungen, Fragen und Aussagen.</p>
<p><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/09/rverbindungenbiblionetz.png" alt="" width="600" height="397" class="aligncenter size-full wp-image-26817" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/09/rverbindungenbiblionetz.png 600w, https://www.zylstra.org/wp/wp-content/uploads/2026/09/rverbindungenbiblionetz-300x199.png 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/09/rverbindungenbiblionetz-18x12.png 18w" sizes="auto, (max-width: 600px) 100vw, 600px" /><br /><em>Ein Beispiel wie Texte im Biblionetz untereinander verknüpft sind, und diese Vernetzung navigierbar macht</em></p>
<p>Bereits 1996 hat <a href="https://beat.doebe.li/">Beat Döbeli</a> (<a href="https://blog.doebe.li/">blog</a>) das Biblionetz angefangen. Seitdem wird jedes Sachbuch und Artikel das ihm interessiert darin aufgenommen. Wie bei allem was man über Jahre hinweg konsistent macht kommt etwas dabei &#8216;raus das durch das erreichte Volumen eine ganz andere Qualität bekommt. </p>
<p>Vor einigen Jahren bin ich über das Biblionetz gestolpert. Seitdem verwende ich es immer wenn ich eine Frage habe über deutschsprachige Bücher. Als Niederländer habe ich den leichtesten Zugang zu Sachbüchern auf Niederländisch und Englisch. Aber, wie bei Französisch und Spanisch, ist das deutsche Sprachgebiet groß genug um eine ganze andere Bücherwelt zu bieten die für andere Sprachgebiete unsichtbar bleibt. Döbeli hat zwar andere Interessen als ich, aber wir teilen auch Interessen, und das heißt das das Biblionetz sowohl genügend Bücher enthält die ich noch nicht kenne, als auch Bücher die ich bereits kenne. Durch die Verbindungen zwischen Bücher und Konzepte kann ich, startend mit was ich kenne navigieren und (wenigstens für mich) interessantes Neues finden. </p>
<p>Selbst mache ich seit ich schreiben kann Notizen, und seit 1991 auch digital. Mit der Zeit hat sich das zu einer persönlichen Wissensmanagement Methodik entwickelt.  Personal Knowledge Management (PKM) ist deshalb auch ein Interesse von mir seit vielen Jahren. In 2024 wurde in Utrecht (NL) die erste <a href="https://pkmsummit.com/">Europäische PKM Konferenz</a> organisiert. Um zu helfen Sprecher und Teilnehmer zu finden aus unterschiedlichen Europäischen Ländern habe ich mich beim Orga-Team gemeldet. Und dann auch Beat Döbeli eingeladen um sein Biblionetz zu zeigen und über seine Erfahrungen zu erzählen. Leider konnte er bis jetzt nicht, da die Konferenz immer zusammenfiel mit einer wichtigen Fachtagung für ihn. Im nächsten Jahr, 2027, wird es dann klappen. Wer an PKM interessiert ist sollte die Konferenz schon mal im Terminkalender vermerken (12. und 13. März 2027).</p>
<p>Darüber und mehr haben wir gesprochen als wir uns im Juni zum ersten Mal trafen, in Zürich. Wir sprachen über die Anfänge von Biblionetz und was es jetzt für ihn in seiner Arbeit als Professor an der Pedagogische Hochschule Schwyz bedeutet. 1996 war er als Informatiker an die Entwicklung des WWW interessiert, und betrachtete was er met Biblionetz anfing nicht direkt als Akademiker. Jetzt verwendet er das Biblionetz als ein akademisches Werkzeug, indem er zB. Studenten und Kollegen leicht mit Leselisten versehen kann. Bereits am Anfang war ihm klar das es mal eine Zeit geben könnte wo Computer das was er sich mit Biblionetz gebaut hat als persönliches Werkzeug, auch von sich aus machen könnten. Jetzt mit KI ist die Zeit vielleicht angebrochen. </p>
<p>Das <a href="https://beat.doebe.li/bibliothek/rss.xml">Biblionetz ist über RSS zu folgen</a> (wie natürlich auch Beat Döbeli&#8217;s Blog)</p>
<p><em>Dieser Blogartikel ist ein Beitrag an den Blogwochen 2026, im Thema &#8216;Mensch/Projekt der/das mehr Aufmerksamkeit verdient&#8217;.<br />
Die <a href="https://nureinblog.at/69172-blogwochen-2026-wir-fangen-an/">Blogwochen</a>, von Mitte September bis Mitte Dezember, sind dazu gedacht mehr deutschsprachige Blogger sichtbar zu machen, und neue interessante Blogs für den eigenen Feedreader zu finden. <a href="https://www.zylstra.org/blog/2026/09/die-deutschsprachigen-blogwochen-2026/">Ich nehme Teil</a> weil ich mehr auf Deutsch schreiben möchte, und auch die Anzahl der deutschsprachigen Blogs die ich in meinem Feedreader folge vergrößern möchte.</em></p>
</div>
<br/><a href="https://translate.google.com/translate?sl=de&tl=en&u=https%3A%2F%2Fwww.zylstra.org%2Fblog%2F2026%2F09%2Fbeat-dobeli-und-das-biblionetz%2F">machine translation into English</a><br/>]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/09/beat-dobeli-und-das-biblionetz/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title>Blog Archiving to Notes and Embeddings</title>
		<link>https://www.zylstra.org/blog/2026/09/blog-archiving-to-notes-and-embeddings/</link>
					<comments>https://www.zylstra.org/blog/2026/09/blog-archiving-to-notes-and-embeddings/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Sun, 27 Sep 2026 19:52:00 +0000</pubDate>
				<category><![CDATA[Day to Day]]></category>
		<category><![CDATA[metablogging]]></category>
		<category><![CDATA[pkm]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26809</guid>

					<description><![CDATA[Last March I made an export from my blog, so that I have all my blogposts since 2002 (some 3800 posts) in my notes locally. This makes it easier to find and link to my own blogposts while I am working, thinking, or writing in my notes. What I did not do in March was [&#8230;]]]></description>
										<content:encoded><![CDATA[<div class="e-content">
<p>Last March I made an export from my blog, so that I have all my blogposts since 2002 (some 3800 posts) in my notes locally. This makes it easier to find and link to my own blogposts while I am working, thinking, or writing in my notes.  What I did not do in March was set up a way to keep syncing my blogposts to my notes from then on. Meaning I am now 6 months and 60 blogposts out of date.</p>
<p>Meanwhile Dries Buytaert wrote a month ago about how <a href="https://dri.es/finding-related-posts-with-embeddings">he used embeddings to list related postings</a> with a blogpost based on their content, not their tags, keywords or categories.<br />
This sounded immediately interesting and useful. What he describes is to run your site content through a model to create embeddings (a range of numbers as vectors representing your text),  store the embeddings, and have a local script that when a posting is rendered calculates which embeddings match your posting closest. Related postings then is when their numerical vectors are closer than others. Only initially when you calculate embeddings for a posting is this dependent on a LLM, and afterwards it&#8217;s all regular calculations. </p>
<p>I thought creating those same embeddings may also be useful for semantic search within my blogposts for myself.<br />
The plan therefore now is to</p>
<ul>
<li>set up ongoing archiving of my blog  and parsing them to my local notes</li>
<li>create embeddings using a model locally through ollama, e.g. with bge-m3, qwen3-embedding:0.6b, or nomic-embed-text</li>
<li>use those embeddings to calculate related postings for each existing posting, and forwards for new postings</li>
<li>use those same embeddings to allow finding related postings in my notes when writing something</li>
</ul>
<p>The archiving is set up now, and updated once every day.</p>
<p>Next step is building and testing different ways of creating embeddings. The model to make those must be able to deal with and compare multiple languages, English, Dutch and German. I need to use the same model across all embedding of old and new posts, and when I want to compare with something I&#8217;m writing in my notes. This as different models create different representations of the same text and are not interchangeable. Also <a href="https://www.zylstra.org/blog/2026/09/big-data-ldn-impressions/">earlier this week in London</a> the topic of model drift came up, that the same model, if you use it through an API from a vendor, will get tweaked and change over time. This results in slowly shifting embeddings.<br />
With the team at Activate Intelligence we talked about how you&#8217;d need to pin down a model, and also log what a good result looks like when you start using it, so that you can detect drift over time. Using a local model (not an online model through an API) reduces drift, as things stay the same as long as I don&#8217;t update a model on my machine.<br />
Solving drift means recreating all embeddings, which is the same effort as creating them now is. So, it&#8217;s not a barrier, but it is maintenance you need to stay on top of.</p>
</div>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/09/blog-archiving-to-notes-and-embeddings/feed/</wfw:commentRss>
			<slash:comments>1</slash:comments>
		
		
			</item>
		<item>
		<title>Big Data LDN Impressions</title>
		<link>https://www.zylstra.org/blog/2026/09/big-data-ldn-impressions/</link>
					<comments>https://www.zylstra.org/blog/2026/09/big-data-ldn-impressions/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Sat, 26 Sep 2026 13:27:12 +0000</pubDate>
				<category><![CDATA[conferences]]></category>
		<category><![CDATA[data governance]]></category>
		<category><![CDATA[Digital disruption]]></category>
		<category><![CDATA[Distributed digital transformation]]></category>
		<category><![CDATA[Innovation]]></category>
		<category><![CDATA[technology]]></category>
		<category><![CDATA[agenticai]]></category>
		<category><![CDATA[AI]]></category>
		<category><![CDATA[bigdataldn]]></category>
		<category><![CDATA[london]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26787</guid>

					<description><![CDATA[The past days I was in London to visit the Big Data LDN event and meet with my friend Paolo and his company&#8217;s team at Activate Intelligence. It made for interesting contrasts, and those contrasts made the trip worthwhile where the event on its own would not have. Big Data LDN is a yearly event. [&#8230;]]]></description>
										<content:encoded><![CDATA[<div class="e-content">
<p>The past days I was in London to visit the <a href="bigdataldn.com/">Big Data LDN</a> event and meet with my friend Paolo and his company&#8217;s team at Activate Intelligence. It made for interesting contrasts, and those contrasts made the trip worthwhile where the event on its own would not have.</p>
<p>Big Data LDN is a yearly event. I realised as I explored the venue it&#8217;s been a long time since I was at a trade fair style event where the focus is on enterprises and business for business sake, as opposed to mostly an exchange of experiences and ideas. That still happened somewhat but it was clearly not the main focus, the <a href="https://www.bigdataldn.com/en-gb/conference.html#/sessions">program</a> with 280 or so sessions in two days was mostly to ensure all vendors could share something on stage at some point.<br />
Of course next to big data, or rather enterprise data (most data isn&#8217;t actually big imo, and what counts as big is ever shifting upwards, data is mostly just messy), AI took most of the attention. Enterprise data serves as input for anything AI.</p>
<p><a href="https://flickr.com/photos/tonz/55546659579/in/datetaken/"><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/09/rolympiahallTonZijlstraccbyncsa.jpg" alt="" width="400" height="533" class="aligncenter size-full wp-image-26793" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/09/rolympiahallTonZijlstraccbyncsa.jpg 400w, https://www.zylstra.org/wp/wp-content/uploads/2026/09/rolympiahallTonZijlstraccbyncsa-225x300.jpg 225w, https://www.zylstra.org/wp/wp-content/uploads/2026/09/rolympiahallTonZijlstraccbyncsa-9x12.jpg 9w" sizes="auto, (max-width: 400px) 100vw, 400px" /></a><br /><em>One of the two halls at the Olympia venue</em></p>
<p>Back in the &#8217;00s when I visited corporate knowledge management events there were plenty of slides showing complicated overviews of data sources, stores and flows in a company and how adding name-your-brand-here&#8217;s knowledge management system would make it easy to manage all of it and use it. The promise was that adding one more technology layer in the mix would finally make everything easy and governable, effortless even, if you were doing it right. If it wasn&#8217;t effortless then obviously you weren&#8217;t doing it right, and you needed to buy still one more tool or platform. </p>
<p>Two decades on that pattern is still very much the same.<br />
There were plenty of complicated slides trying to show all the many data related technological building blocks. Where there used to be data warehouses and then data lakes as panacea a few years ago, there are now lakehouses added to combine aspects of both and make their contents available to machine learning, and various calls to go &#8216;beyond&#8217; the lakehouse were heard. You end up with something that still nobody really has a handle on, but don&#8217;t worry, AI will finally make everything effortless, by adding an agentic AI layer. AI agents tend to multiply, and that means you need to start adding &#8216;agent control towers&#8217;, and agents make stuff up so you need to provide the right harnesses for them, which also multiply, and manage those, and do a whole bunch of additional logging for auditing and compliance. All that comes with a range of additional tools for your already very complicated information household, adding just more complicatedness and uncertainty about what it all is for. </p>
<p>What will save you is adding just one more thing, or actually two things, the message was. More context and persistent memory for your AI agents. This according to various speakers was the new focus compared to the event last year. Context and memory so that your AI will understand your business. The underlying assumption being that anyone within enterprise actually understands their business themselves, and that it&#8217;s just a matter of letting your AI know it too. Nobody does really, understand their business that is, and for evidence look at the sheer volume of back and forth and reporting and analysis and meetings going on that is only dealing with internal flows of data, and those same reports, analysis and meeting notes, to &#8216;finally&#8217; make things run &#8216;smoothly&#8217; internally.</p>
<p>Still I learned a bunch of things. How data that used to be for analysis after the fact is now being used for decision making upfront and for actions right now (but without mentioning whether that may require different sorts of data, and silently assuming you can do either with the same data). How data governance is becoming a continuous and real time thing, and observability is of increasing importance (I agree that is highly interesting). About what context and memory for AI is, and that despite calls to not just look at token cost it is often touted as a great reduction in token cost. (Not using AI all over the place also greatly reduces token cost of course.) How everyone builds tools to make other tools by vendors in the booth next to them work better, or coordinate between that and the tools of vendors several booths over. It&#8217;s tools and agents all the way down, where there used to be <a href="https://en.wikipedia.org/wiki/Turtles_all_the_way_down">turtles</a>, to make sense of the world. Much of it sounds like coping strategies made of python and markdown to me. </p>
<p>Hearing the lingo, and listening to what was said and what wasn&#8217;t was genuinely useful too. The way that terms like (un)structured data, intelligence and knowledge, but never information, are bandied about as if they are precisely defined, meaningfully quantifiable, and all in the same category. That ethics and regulation is always treated as a compliance issue, adding a burden to your logging and auditing needs, and that no one tends to view ethics and compliance as upfront instruments to improve the overall quality of what you&#8217;re designing or doing, thus reducing your externalities and their attached risks. Which would greatly reduce internal friction, actually.</p>
<p>That impact is of course the only measure of succes, but hardly anyone goes on to explain what the impact actually has been for them or their clients. Where the externally visible acceleration is, where the non-lineair jump in actual business outcomes, remains opaque. Impacts that do get mentioned are again often about increased volumes of AI deployment against lowered token costs, about the reduced time needed to pump around those internal reports, meeting notes, and analysis so you can now pump around more of them. That is consistent though, because that was true for the KM conferences of old too, and just as much for any big data, data warehousing, data lake event since then until now. Because none of it is designed for real world impact, it is being designed and deployed for internal efficiency. Not to change or innovate your external outputs and how they have influence in the world. It was tiring to see how much of it all comes across as just busywork. </p>
<p>That we are moving from &#8216;data driven&#8217; to &#8216;decision centric&#8217; as one panelist remarked. Which sounds good and proper, but didn&#8217;t go into how messy our decisions mostly really are, and how it might be difficult to externalise how we come to decisions, or reverse engineer them so that AI agents might do it too.</p>
<p>Some speakers clearly felt something&#8217;s off and launched their own small philosophies in response. Some were told by others something&#8217;s off, and where it was their children who told them, they were jolted into making slides about it, but not changing their company meaningfully (judging by those same slides) even if it had been years since their kid stated the obvious. Those philosophies tended to stay within the same unchallenged domain assumptions and usually suggested the answer was another technology component or layer for the complicated overview sheet we started out with at the outset. That, or a clever acronym. Any other remaining issues were left as an exercise for &#8216;leadership&#8217;. </p>
<p>Not that I know how all this should be done, but over the years I have come to the conclusion that the problem with doing anything at scale that isn&#8217;t physical like assembly lines but informational is, well, scale itself. Regular large scale, as well as the hyped and hyper varieties. This as it is usually approached or assumed as a type of centralisation, which is fruitless as it crumbles in the face of the next change in the status quo in 3, 2, 1&#8230; And it tends to ignore the people involved at every step, and the intention or agency they have, which guarantees the type of complexity that reality can roll with but not your lakehouse or agent control tower.</p>
<p>Back to the slides, before I get to the contrast that made it a valuable trip.<br />
With AI front and center, it came as no surprise that many slides were obviously AI generated. With all the tells that Claude has, you know them, phrases like &quot;This is not coincidence. This is a pattern.&quot; and &quot;All four are sound. One is different.&quot; and &quot;Subheaders of Overstated Importance&quot; followed by a sentence that doesn&#8217;t bear it out. It made for overly verbose slides, because Claude likes to spend them tokens. It became funny when speakers hadn&#8217;t fully read their generated slides beforehand but had trusted the output blindly. Trust was another big topic of course, but trust is more than ever better harnessing up to the logical endpoint where you&#8217;ve made something fully deterministic and fit for regular automation, or, like, just do it yourself. Talking about trusting AI outputs and presenting slides where there are generated sentences that break the logic of your narrative, or drop in additional terms midway that came out of nowhere, I assume from another AI conversation,  was funny. Or a random insertion like &#8216;The other ten are not&#8217; while recounting a list of four things. Kudos to all those speakers, even when they were clearly spooked a bit by what they found on their own slides, for making nonsensical bridges, or jumping over the misaligned bits as if the slide didn&#8217;t exist, and generally getting away with it. Good enough, they&#8217;ll have thought.</p>
<p>I stayed on through the entire first day, and over lunch met someone who had already arrived at similar conclusions as above, just much faster, but on the second day I acutely felt the diminishing returns of staying longer and walked out well before lunch time. I strolled the ten minutes or so down the road to <a href="https://en.wikipedia.org/wiki/Holland_Park#The_public_park">Holland Park</a>, with a short detour through the adjacent <a href="https://designmuseum.org/">Design Museum</a>, and sat under a tree in fresh air and gorgeous weather, being stared at by both a crow and a squirrel waiting to see if I had anything edible in my backpack that I cared to share or leave crumbs of.</p>
<p><a href="https://flickr.com/photos/tonz/55548368640/"><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/09/rcrowTonZijlstraccbyncsa.jpg" alt="" width="533" height="400" class="aligncenter size-full wp-image-26796" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/09/rcrowTonZijlstraccbyncsa.jpg 533w, https://www.zylstra.org/wp/wp-content/uploads/2026/09/rcrowTonZijlstraccbyncsa-300x225.jpg 300w, https://www.zylstra.org/wp/wp-content/uploads/2026/09/rcrowTonZijlstraccbyncsa-16x12.jpg 16w" sizes="auto, (max-width: 533px) 100vw, 533px" /></a><br /><em>Watchful eyes while I&#8217;m in the park</em></p>
<p>The trip was valuable, despite my 1500 word rant you just waded through, because of context and memory. They definitely had the importance of context and memory right at the event.<br />
I spent the late afternoon and early evening of the second day with the <a href="https://activateintelligence.co.uk/team/">team</a> of <a href="https://activateintelligence.co.uk/">Activate Intelligence</a>, the company of my friend <a href="https://val.demar.in/">Paolo</a>, placing it in context with my Big Data LDN visit. His team consists of very bright mostly young people that like my own <a href="https://thegreenland.eu/team/">team</a> seem so much more capable and driven than I ever was at their age. Paolo and I go back a very long time (2003), and as I was in knowledge management from a change perspective then, he was into vector graphs and ontologies. Remembering that early work and especially what wasn&#8217;t possible then, because it depended on people being extremely consistent in e.g. providing metadata (tags etc.), made him see early on what LLMs might mean. The focus of the tools they build, and how they deploy AI agents is both aimed internally and externally. For both it is about doing two things. The first is that anything they build is geared towards making a professional and their team as a whole more effective, by supporting them. People aren&#8217;t there to support the process, the process is to support the professional. The second is that AI is being used to do things that would otherwise be too cumbersome or costly to be useful to do by hand internally or do at all. In a previous conversation with Paolo, we talked about this in terms of tasks underneath the <a href="https://en.wikipedia.org/wiki/The_Nature_of_the_Firm">Coasean</a> floor. Here AI is not a substitute for existing human tasks, but used for new types of tasks that are not being done yet because they would otherwise require too much of a coordination burden. This makes it possible to do things with small teams that previously would require a bigger size to be able to cover the associated overhead costs, as well as things that would require too much coordination for any use case. This contrast, enterprises aiming to reduce internal size and friction, albeit by adding new layers of it, for the same external outcome, versus small teams working towards the largest external outcome while avoiding or obviating the need of becoming a sizeable enterprise stands out. Also because its respective directions and pressures ultimately point to the same thing, I think, in bringing the Coasean ceiling, i.e. the maximum expected useful size of an enterprise, down. Paolo wondered how many floors of London&#8217;s high rise office towers would go dark in the coming time because of it. That seems promising in a way too, that where trust busting does not currently succeed much in the informational realm, it may well become ineffective to be that big with swarms of smaller efforts delivering the same or better. Why build an enterprise above a certain size, a 19th century <a href="https://www.zylstra.org/blog/2019/06/our-ai-overlords-are-already-here-they-likely-employ-you/">slow AI</a>, if you can have the faster 21st century version at team or individual scale without all the hassle?</p>
</div>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/09/big-data-ldn-impressions/feed/</wfw:commentRss>
			<slash:comments>1</slash:comments>
		
		
			</item>
		<item>
		<title>Transparant Lies</title>
		<link>https://www.zylstra.org/blog/2026/09/transparant-lies/</link>
					<comments>https://www.zylstra.org/blog/2026/09/transparant-lies/#respond</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Sat, 26 Sep 2026 13:18:12 +0000</pubDate>
				<category><![CDATA[Flashes]]></category>
		<category><![CDATA[ethics]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26788</guid>

					<description><![CDATA[During lunch in Holland Park, London this week, the water bottle had the text &#8216;ethical glass&#8217;.]]></description>
										<content:encoded><![CDATA[<p><a href="https://flickr.com/photos/tonz/55548264319/in/datetaken/"><img loading="lazy" decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2026/09/rethicalglassTonZijlstraccbyncsa.jpg" alt="" width="400" height="533" class="aligncenter size-full wp-image-26789" srcset="https://www.zylstra.org/wp/wp-content/uploads/2026/09/rethicalglassTonZijlstraccbyncsa.jpg 400w, https://www.zylstra.org/wp/wp-content/uploads/2026/09/rethicalglassTonZijlstraccbyncsa-225x300.jpg 225w, https://www.zylstra.org/wp/wp-content/uploads/2026/09/rethicalglassTonZijlstraccbyncsa-9x12.jpg 9w" sizes="auto, (max-width: 400px) 100vw, 400px" /></a><br /><em>During lunch in Holland Park, London this week, the water bottle had the text &#8216;ethical glass&#8217;.</em></p>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/09/transparant-lies/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title>Die deutschsprachigen Blogwochen 2026</title>
		<link>https://www.zylstra.org/blog/2026/09/die-deutschsprachigen-blogwochen-2026/</link>
					<comments>https://www.zylstra.org/blog/2026/09/die-deutschsprachigen-blogwochen-2026/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Mon, 21 Sep 2026 16:53:23 +0000</pubDate>
				<category><![CDATA[Deutsch]]></category>
		<category><![CDATA[metablogging]]></category>
		<category><![CDATA[blogwochen]]></category>
		<category><![CDATA[BlogWochen2026]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26780</guid>

					<description><![CDATA[Die deutsprachigen Blogwochen sind wieder gestartet. Eine Initiative von Robert Lender, Benedikt Müller und Dirk Deimeke, deren Blogs sich schon lange in meinem Feedreader befinden. Die Idee ist: Blogger schreiben während 3 Monate über eine kleine Liste von Themen, und werden so gegenseitig sichtbarer. So verknüpfen sich die Blogger, entstehen neue Gespräche, und bereichert man [&#8230;]]]></description>
										<content:encoded><![CDATA[<div class="e-content">
<p>Die deutsprachigen Blogwochen sind wieder gestartet.<br />
Eine Initiative von <a href="https://nureinblog.at/69172-blogwochen-2026-wir-fangen-an/">Robert Lender</a>, <a href="https://benedikt.io/2026/09/blogwochen-2026-los-gehts/">Benedikt Müller</a> und <a href="https://www.deimeke.net/dirk/blog/index.php?/archives/4522-BlogWochen-2026.html">Dirk Deimeke</a>, deren Blogs sich schon lange in meinem Feedreader befinden. </p>
<p>Die Idee ist: Blogger schreiben während 3 Monate über eine kleine Liste von Themen, und werden so gegenseitig sichtbarer. So verknüpfen sich die Blogger, entstehen neue Gespräche, und bereichert man seinen Feedreader. Robert baut eine öffentliche <a href="https://data.roblen.at/apps/bookmarks/public/dF8IRUM">Liste von Blogwochen-Beiträgen</a> auf, damit man sehen kann wer da so mitmacht.</p>
<p>Obwohl mir bei den vorgeschlagenen Themen nur sehr wenig einfällt werde ich versuchen trotzdem mit zu machen. Das hat zwei Gründe. Erstens sind für mich Blogs wichtig als distribuiertes Gespräch: unterschiedliche Blogger reagieren in eigenen Beiträgen auf was Andere geschrieben haben, und man kan über die verlinkten Artikel die verschiedene Gespräche nicht nur mitlesen sondern man kann einfach mitmachen indem man selbst auch was schreibt worin man das Vorhergehende per Hyperlink erwähnt.<br />
Zweitens habe ich schon viel länger den Wunsch mehr auf Deutsch zu bloggen. Meistens blogge ich auf English, und ab und zu <a href="https://www.zylstra.org/blog/category/nederlands/">auf Niederländisch</a>, und nur sehr selten <a href="https://www.zylstra.org/blog/category/deutsch/">auf Deutsch</a>. Obwohl ich sehr daran interessiert bin mich auch im deutschsprachigen Raum aus zu tauschen mit anderen Bloggern, vor allem über die Themen die mich interessieren: Europäische Gesetze und Initiative rundum Daten und Digitalisierung, persönliches Wissensmanagement, persönliche Tools (Software, AI usw.), und Digital-Ethik. Um Blogger zu finden die bei Gelegenheit auch über diese Themen schreiben hilft nicht nur suchen sondern auch mich sichtbar zu machen damit ich selbst leichter findbar bin. Sichtbarkeit ensteht nur wenn man tatsächlich bloggt, auch wenn man noch keine Ahnung hat ob irgendwer das auch liest. Wâhrend die Blogwochen gibt es eine etwas höhere Chance darauf das es jemand liest.<br />
Genügend Gründe also um bei den Blogwochen ein wenig mitzumachen. Wer weiss welche andere Blogger ich dabei auf die Spur komme!</p>
</div>
<br/><a href="https://translate.google.com/translate?sl=de&tl=en&u=https%3A%2F%2Fwww.zylstra.org%2Fblog%2F2026%2F09%2Fdie-deutschsprachigen-blogwochen-2026%2F">machine translation into English</a><br/>]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/09/die-deutschsprachigen-blogwochen-2026/feed/</wfw:commentRss>
			<slash:comments>7</slash:comments>
		
		
			</item>
		<item>
		<title>Treffen wir uns auf der Swiss Data Alliance Konferenz in Lausanne?</title>
		<link>https://www.zylstra.org/blog/2026/09/treffen-wir-uns-auf-der-swiss-data-alliance-konferenz-in-lausanne/</link>
					<comments>https://www.zylstra.org/blog/2026/09/treffen-wir-uns-auf-der-swiss-data-alliance-konferenz-in-lausanne/#respond</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Sat, 19 Sep 2026 16:53:06 +0000</pubDate>
				<category><![CDATA[conferences]]></category>
		<category><![CDATA[Deutsch]]></category>
		<category><![CDATA[opendata]]></category>
		<category><![CDATA[travel]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26772</guid>

					<description><![CDATA[Am 28. Oktober findet in Lausanne in der Schweiz die jährliche Konferenz der Swiss Data Alliance (SDA) statt. Die Swiss Data Alliance ist ein unabhängiger Think-Tank, der den Dialog über Daten und Digitalisierung zwischen Wirtschaft, Gesellschaft und Verwaltung fördern will. Die Konferenz findet statt unter dem Titel Swiss Data Space Forum 2026: Unlocking Data for [&#8230;]]]></description>
										<content:encoded><![CDATA[<div class="e-content">
<p>Am 28. Oktober findet in Lausanne in der Schweiz die jährliche Konferenz der <a href="https://swissdataalliance.ch/">Swiss Data Alliance</a> (SDA) statt. Die Swiss Data Alliance ist ein unabhängiger Think-Tank, der den Dialog über Daten und Digitalisierung zwischen Wirtschaft, Gesellschaft und Verwaltung fördern will. Die Konferenz findet statt unter dem Titel <a href="https://forum.swissdataalliance.ch/">Swiss Data Space Forum 2026: Unlocking Data for AI</a>. Am 27. Oktober, gibt es zusätzlich einen Masterclass für Data Stewards mit <a href="https://en.wikipedia.org/wiki/Stefaan_Verhulst">Stefaan Verhulst</a> vom Governance Lab der New York University. Als Mitglied der SDA werde ich an beiden Tagen <a href="https://www.zylstra.org/blog/2026/09/26769/">mit dabei</a> sein in Lausanne.</p>
<p>Meine eigenen Kontakte mit der Swiss Data Alliance sind schon viele Jahre alt. Angefangen hat es mit Kontakten zu der <a href="https://opendata.ch/">Schweizer Open Data Community</a>, als ich <a href="https://opendataconsulting.ch/en/">André Golliez</a> auf einer Tagung in <a href="https://www.zylstra.org/blog/2011/10/making-local-open-data-work-for-local-government/">Warschau 2011</a> zum ersten Mal getroffen habe. Über einige Jahre gab es dann monatliche deutschsprachige Onlinetreffen mit Open Data-Aktivisten aus den DACHLI Ländern, und mit mir aus den Niederlanden. </p>
<p>Ich habe ein paar Mal auf der Schweizer Open Data Konferenz gesprochen und auch vor zehn Jahren beigetragen an einer Studie nach einer nationalen Dateninfrastruktur für die Schweiz.<br />
In meiner Arbeit setze ich mich sehr viel auseinander mit der europäischen digitalen und Datenstrategie, mit europäischen Datenregeln, mit internationalen Standarten dazu, und wie die europäischen Data Spaces implementiert werden. </p>
<p>Obwohl die Tagung in Oktober auf Englisch stattfindet und im französischsprachigen Teil der Schweiz, schreibe ich hier über das Event auf Deutsch, weil für mich die Kontakte mit der Schweiz rund um offene Daten und Data Spaces vor allem mit der deutschsprachigen Community verknüpft sind. </p>
<p>Ich hoffe in Lausanne meine Beziehungen zu dieser Community zu reaktivieren, nach einigen Jahren meiner Abwesenheit, und vielleicht kommt auch ein potentielles Projekt daraus hervor. Eine meiner Kolleginnen in meinem Team bei <a href="https://thegreenland.eu">The Green Land</a> ist in der Schweiz aufgewachsen und vielleicht können wir da was zusammen machen. </p>
<p>Drei Wochen vor der Konferenz kommt André Golliez nach Holland und wir werden uns zusammen mit Leuten des TNO Instituts treffen, u.a. um über <a href="https://coe-dsc.nl/">Data Space Initiative</a> zu sprechen, und auch meinem Kunden <a href="https://www.geonovum.nl/index.php/about-us">Geonovum</a> besuchen.</p>
<p>Lasst es mich wissen, falls ihr auch vorhabt, in Lausanne mit dabei zu sein, und ich freue mich auf ein Treffen im nächsten Monat. </p>
</div>
<br/><a href="https://translate.google.com/translate?sl=de&tl=en&u=https%3A%2F%2Fwww.zylstra.org%2Fblog%2F2026%2F09%2Ftreffen-wir-uns-auf-der-swiss-data-alliance-konferenz-in-lausanne%2F">machine translation into English</a><br/>]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/09/treffen-wir-uns-auf-der-swiss-data-alliance-konferenz-in-lausanne/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
		<item>
		<title></title>
		<link>https://www.zylstra.org/blog/2026/09/26769/</link>
					<comments>https://www.zylstra.org/blog/2026/09/26769/#comments</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Fri, 18 Sep 2026 15:58:53 +0000</pubDate>
				<category><![CDATA[Dopplr]]></category>
		<category><![CDATA[geneva]]></category>
		<category><![CDATA[lausanne]]></category>
		<category><![CDATA[swissdataalliance]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26769</guid>

					<description><![CDATA[I will be in Lausanne from Monday October 26th late afternoon until Thursday October 29th early morning. I will be in Lausanne for an event by the Swiss Data Alliance. If you care to meet-up let me know. I will fly in and out of Geneva, so on the day of arrival a meet-up in [&#8230;]]]></description>
										<content:encoded><![CDATA[<p><img decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2022/10/dopplrblocks.png" style="float:left; margin-right: 5px; padding-top:4px;"> <span style="display: block;"></br></p>
<p><em> I will be in <span class="p-locality" style="display: inline-block;">Lausanne</span>  <span style="display: inline-block;">from <span style="display: inline-block;" class="dtstart" title="26-10-2026">Monday October 26th late afternoon</span> until <span style="display: inline-block;" class="dtend" title="29-10-2026">Thursday October 29th early morning.</span></span></em></span></span></p>
<p></p>
<p>I will be in Lausanne for an event by the Swiss Data Alliance. If you care to meet-up let me know.</p>
</p>
<p>I will fly in and out of <span class="p-locality" style="display: inline-block;">Geneva</span>, so on the day of arrival a meet-up in Geneva is possible too.</p>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/09/26769/feed/</wfw:commentRss>
			<slash:comments>1</slash:comments>
		
		
			</item>
		<item>
		<title></title>
		<link>https://www.zylstra.org/blog/2026/09/26767/</link>
					<comments>https://www.zylstra.org/blog/2026/09/26767/#respond</comments>
		
		<dc:creator><![CDATA[Ton Zijlstra]]></dc:creator>
		<pubDate>Fri, 18 Sep 2026 15:54:21 +0000</pubDate>
				<category><![CDATA[Dopplr]]></category>
		<category><![CDATA[jtc25]]></category>
		<category><![CDATA[oslo]]></category>
		<guid isPermaLink="false">https://www.zylstra.org/blog/?p=26767</guid>

					<description><![CDATA[I will be in Oslo from Tuesday October 13th early afternoon until Friday October 16th early evening. I will be in Oslo for a CEN/CENELEC JTC25 plenary meeting (European data space standardisation work). I may have time on at least one half day, and two of the evenings if you want to meet up.]]></description>
										<content:encoded><![CDATA[<p><img decoding="async" src="https://www.zylstra.org/wp/wp-content/uploads/2022/10/dopplrblocks.png" style="float:left; margin-right: 5px; padding-top:4px;"> <span style="display: block;"></br></p>
<p><em> I will be in <span class="p-locality" style="display: inline-block;">Oslo</span>  <span style="display: inline-block;">from <span style="display: inline-block;" class="dtstart" title="13-10-2026">Tuesday October 13th early afternoon</span> until <span style="display: inline-block;" class="dtend" title="16-10-2026">Friday October 16th early evening.</span></span></em></span></span></p>
<p></p>
<p>I will be in Oslo for a CEN/CENELEC JTC25 plenary meeting (European data space standardisation work). I may have time on at least one half day, and two of the evenings if you want to meet up.</p>
]]></content:encoded>
					
					<wfw:commentRss>https://www.zylstra.org/blog/2026/09/26767/feed/</wfw:commentRss>
			<slash:comments>0</slash:comments>
		
		
			</item>
	</channel>
</rss>
