<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Pami Hekanaho on Medium]]></title>
        <description><![CDATA[Stories by Pami Hekanaho on Medium]]></description>
        <link>https://medium.com/@pamihekanaho?source=rss-40606bf33fef------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*8ST6ZkJylrDnBdcMoOWt0g.png</url>
            <title>Stories by Pami Hekanaho on Medium</title>
            <link>https://medium.com/@pamihekanaho?source=rss-40606bf33fef------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 09:28:19 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@pamihekanaho/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[End of the break]]></title>
            <link>https://pamihekanaho.medium.com/end-of-the-break-19f7d488a185?source=rss-40606bf33fef------2</link>
            <guid isPermaLink="false">https://medium.com/p/19f7d488a185</guid>
            <dc:creator><![CDATA[Pami Hekanaho]]></dc:creator>
            <pubDate>Sun, 15 Nov 2026 20:18:21 GMT</pubDate>
            <atom:updated>2026-11-15T20:18:21.206Z</atom:updated>
            <content:encoded><![CDATA[<p>I was recently introduced to the Medium AI tool <a href="https://medium.com/blog/tk-a-new-private-notebook-app-from-medium-13d3d6efdce1">TK</a>, the purpose of which is to help a human with writing. You can, for example, get daily prompts or ask a question about your writing to help move on. Prompts are a nice way to access thoughts in the back of your mind and I hope TK keeps me writing although I am back working 100%. Anyways, one of the first things I asked the TK AI was help in finishing my Medium drafts. This is the first one: a post written in 2020 (italics), which I will comment on (not italics) with the benefit of hindsight 20/20.</p><blockquote>My break is almost over as I will go back to work next week. I had dreams of “finishing” a certain part of my studies but truth be told: it is not finished. Reading theory takes so much brain power that I started serious procrastination projects, and towards the end of my break the gyms also opened, at least partly, which meant that more time was spent climbing up the walls.</blockquote><p>Back to work refers to my start at ResearchGate, which was a great gig overall, but also some of the worst months of my career with layoffs disproportionately affecting the team I built. Now I’m already almost 3 years out of there and it seems both a short and long time ago. The longer you live, the blurrier it gets.</p><p>I am still not finished with my studies, and I doubt I ever will. A friend said I can write the thesis when I retire but I wonder what value analysing AI narratives in product design will add to scientific thought then.</p><blockquote>But I enjoyed the time off, even without the possibility of travelling to Finland. Here is a list of some of the enjoyments:</blockquote><blockquote>In addition to actually finishing a guidebook of <a href="https://www.researchgate.net/publication/233923348_Discourse_in_Action_Introducing_Mediated_Discourse_Analysis">Mediated Discourse Analysis</a> I read a lot of fiction. Biggest disappointment was <a href="https://en.wikipedia.org/wiki/A_Little_Life">Hanya Yanagihara’s A Little Life</a> and the delights were many, Keigo Higashino’s Galilei series, as one example.</blockquote><blockquote>I watched some TV too but seriously binged only once, with Michael Jordan’s Last Dance. What can I say? The myth that winners need to be assholes is alive and kicking but his jump shot was beautiful.</blockquote><blockquote>I went back to yoga courageously signing up to pretty advanced Zoom classes. The distance made it feel safer to attend even when not fit, but when Eddie Stern turned on my video camera, I went back to my old habits of overdoing it in front of the coach just to hear him say “good”.</blockquote><blockquote>Our drawing group started another 30-day-drawing-challenge. Always a delight.</blockquote><blockquote>I cooked surprisingly many traditional Finnish dishes. And started growing herbs.</blockquote><blockquote>I had video calls with friends near and far (both in time and space). Somehow it seems easier to arrange in the time of COVID.It is unclear what out of this will remain a part of my life after going back to work, and ultimately also when COVID-restrictions are over. Maybe achievers can change their habits but I always seem to go back to whatever laziness engulfs me.</blockquote><p>I still read. And I have since discovered audiobooks in English, which are much more engaging than the Finnish ones. I also still watch TV. And got back to watching football, specifically women’s football.</p><p>Drawing is not really in my life at the moment, sadly so. Yoga is there a bit, also sadly just a bit. I also don’t cook (not sad) but I have made lohikeitto to many friends to do my bit to promote Finnish cuisine.</p><p>At the time COVID felt like such an immense thing and now we barely remember it. It changed our lives fast but we slid back into our old ways equally fast. Or actually we went worse: views are more polarised, wars get started all around, the economy is tanking, competition for jobs is fierce and many people feel left behind. I also don’t really have video calls with friends anymore. Instead I talk to AI, my new friend, preferred over humans by many others as well. So, welcome to my life TK. I hope Claude won’t be jealous.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=19f7d488a185" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Dive into blog data about designing AI]]></title>
            <link>https://pamihekanaho.medium.com/dive-into-blog-data-about-designing-ai-89be3826b939?source=rss-40606bf33fef------2</link>
            <guid isPermaLink="false">https://medium.com/p/89be3826b939</guid>
            <category><![CDATA[ai-design]]></category>
            <category><![CDATA[text-analysis]]></category>
            <dc:creator><![CDATA[Pami Hekanaho]]></dc:creator>
            <pubDate>Fri, 29 May 2026 12:45:58 GMT</pubDate>
            <atom:updated>2026-05-29T12:45:58.132Z</atom:updated>
            <content:encoded><![CDATA[<p>Today I did a dive into potential thesis data. My main data comes from Google (and possibly IBM) but I have been entertaining the idea of looking at “practitioner content”, that is, Medium blogs. So today I did some digging to see if there is anything interesting around and found some (to complement a set I found last year), but most people are more interested in writing about AI as a design tool rather than designing AI itself. Makes sense as most of us are not designing AI, and the topic does not lend itself to declarations of this and that being dead as easily as the topics of tools and processes.</p><p>So, I sidetracked into general googling using the search terms such as “designing AI” and “design at [insert big bad company name] and I found a bunch of YouTube videos where heads of AI design teams talk about designing AI. YouTube is great, because NotebookLM can read its transcripts and I don’t have to watch the videos themselves (heh). I really don’t like videos.</p><p>I call these three batches of data the following:</p><ol><li>Medium, pick 1 (published between Aug 25, 2024 — Jan 12, 2025)</li><li>Medium, pick 2 (published between Aug 2025 — March 2026)</li><li>Corporate representatives (published between Feb 2024 — April 2026, where the Feb 2024 is an outlier)</li></ol><p>Enough about data collection, let me list what I found.</p><ul><li>It seems like the practitioners are finding consensus when it comes to AI design principles / guidelines. Concepts of trust, safety, feedback, and human oversight are at the heart of all the data sets.</li><li>The more recent guidelines are much more nuanced, and with the products iterating, there are many more examples to draw from. Practice teaches.</li><li>The corporate representative data surfaces speed (of execution, of iteration) much more than the other sets. I am venturing that it is because its representation skews towards design directors and such who get a kick out of speed like any proper executive.</li><li>To be fair, also the nature of their products requires rapid user feedback loops with real data even when it slaps them in the face at times. It is still noticeable how little they talk about friction and high-stakes use cases which the guideline peeps pretty much always mention.</li><li>The corporate set is also more optimistic in its sentiment compared to the others. They are excited about “uncharted territories”, probably because they feel like they are at the heart of the buzz, on the winning side.</li><li>I cannot include the corporate representative data in the thesis. I just did this for fun, don’t worry, I’m not going to be this biased.</li><li>Like the Google set, the guideline data is very pragmatic and scoped to the design process and best practices. Human-centredness is about catering to the goals humans set, and placing the human as the decision maker, not about what AI interaction is doing to humans (especially as decision makers).</li><li>So, although AI material changes the design process, it does not really change how guidelines are written. Details laters!</li><li>And finally, we don’t really know if these companies and designers practice as they teach. I mean, I know I don’t. Like Sheryl Crow says, every day is a winding road.</li></ul><p>I love this pic we did for my Turku presentation earlier (but never showed).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Lo6QZ8JSrp10VoALL-SqPA.png" /><figcaption>Differences between traditional and AI design processes. Picture by Gemini, NN-design and me</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=89be3826b939" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Claude Design for slides? No.]]></title>
            <link>https://pamihekanaho.medium.com/claude-design-for-slides-no-c4fae3b7e500?source=rss-40606bf33fef------2</link>
            <guid isPermaLink="false">https://medium.com/p/c4fae3b7e500</guid>
            <category><![CDATA[claude-design]]></category>
            <category><![CDATA[grumpy]]></category>
            <category><![CDATA[ai-slop]]></category>
            <dc:creator><![CDATA[Pami Hekanaho]]></dc:creator>
            <pubDate>Wed, 20 May 2026 21:52:19 GMT</pubDate>
            <atom:updated>2026-05-20T21:55:28.123Z</atom:updated>
            <content:encoded><![CDATA[<p>We are having an innovation week and I wanted to use Claude Design to build our presentation deck. Here is my experience.</p><h3>Context</h3><ul><li>We have our marketing design system built in Claude Design.</li><li>We had documented a lot of our teamwork in Confluence.</li></ul><h3>Prompting</h3><ul><li>I gave the contexts and a simple structure for the slides to get started. My prompts for this type of iterative activity are never super structured because I don’t know what I want to change before I see the output.</li><li>Claude Design does not (yet?) talk to the rest of the Claude family so it could not read my Confluence page directly.</li></ul><h3>Output</h3><ul><li>The first output was a flashy set of slides which on the surface follow our marketing design system but it was incredibly difficult to read as each slide was crammed full of this and that, visual elements with no meaning and soo many text styles.</li><li>In addition, the titles jumped on each slide. I mean WTF?</li></ul><h3>Iteration</h3><ul><li>Here is where the fun part started. Claude outputs html but does not allow editing the markup directly in Claude Design. When the first output (like with generative AI always) is too much, the easiest thing would be to remove the extras in the markup. Now I needed to prompt it and it was very time-consuming and messy. For example, when asked to remove a blue line in a rectangle, it also removed all of the text in the rectangle. Or when I ask it to remove a title, it says “Removed” but the title is still there.</li><li>When I edited the text using their clunky wysiwyg editor, the next time I prompted a change, my hand-edited bits were either gone or changed. This happened every time which meant that after each update, I had to go through the whole deck to see what I have to handtweak again.</li><li>When prompted to change something, it seems to always start from scratch. And it does not remember anything which became painfully obvious when I was asking for it to bring back stuff it removed without prompting.</li><li>Ctrl-z does not work.</li><li>It lied to me and said there is a markup editing option. There isn’t.</li></ul><h3>Conclusion</h3><ul><li>I am seriously wondering how all the 10x designers orchestrate their design intern army to produce output that makes sense. Creating the deck in Claude Design took me at least 3 times longer than if I had done the same thing in Figma (10 times if I just use a FigmaSlide template), and it looks worse.</li><li>I think it can work for people who don’t care. It is probably better than a lot of people’s GSlides, but only if you don’t really look at the content. But hey, At Least Every Word Starts With A Capital.</li><li>The “quirky” waiting texts like “Caramelising onions…” really started to annoy me.</li></ul><h3>Real conclusion</h3><ul><li>I have been using Cursor heavily for vibecoding our front end application and I feel compassion for the human developers who have had to code review my commits.</li><li>I have also used Gemini and Claude Cowork for data analysis and after this experiment, I have a lot more fear that the data analysis using AI is as bad as the design output from Claude.</li><li>I am too old for this shit and while I feel I can learn a lot with AI, there is a chance that everything I learn is a lie.</li></ul><p>See snippets of our chat here.</p><figure><img alt="Snippets of my chat with Claude Design" src="https://cdn-images-1.medium.com/max/807/1*IkafjKn8YIaFVsiv06IJVA.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c4fae3b7e500" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Notes from AI and interactions seminar]]></title>
            <link>https://pamihekanaho.medium.com/notes-from-ai-and-interactions-seminar-7f7c0217e1d9?source=rss-40606bf33fef------2</link>
            <guid isPermaLink="false">https://medium.com/p/7f7c0217e1d9</guid>
            <category><![CDATA[conversation-analysis]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Pami Hekanaho]]></dc:creator>
            <pubDate>Thu, 23 Apr 2026 15:26:35 GMT</pubDate>
            <atom:updated>2026-04-23T15:26:35.005Z</atom:updated>
            <content:encoded><![CDATA[<p>I went to a two-day seminar on human — AI interactions and the basics of <a href="https://en.wikipedia.org/wiki/Conversation_analysis">conversation analysis</a>. My university department is big on it but I have never even dabbled. This time I could not resist because they added the machine to the mix. Some thoughts below:</p><ul><li>It was amazing to be taught by a human teacher with years of experience in the topic and engage in conversations with people from various educational and cultural backgrounds. I found myself asking questions and commenting here and there (most likely boring everyone else). I guess I am finally old enough not to care.</li><li>There were quite many researchers looking at robotic interactions which left me wondering about the robot form factors. While AI guidelines say don’t anthropomorphise, robotics seems to have a hard time overcoming the human form. Is it our lack of imagination or is human form really the optimum for every use case? The interactions with the humanoids were awkward because they were guided by the human form but not human while the interactions with e.g. delivery robots which do not have a human form seem easier to put into a category: moving box which delivers stuff, no need to comment on their body.</li><li>It was also good to hear that many are trying to figure out what to do with AI and education. A big topic in itself but I would not say it’s all bad. It forces us to reconsider how to measure learning (which has never been easy) and maybe these smart people figure out something that we didn’t need to apply before. The real power of learning with AI is for those who are intrinsically motivated.</li><li>A lot of our conversation revolved around a specific concept in conversation analysis: intersubjectivity, which refers to the shared understanding between participants in a conversation. One of the source articles divided understanding into two different aspects of it: interactional, which means that the conversation participants understand enough to say the next line, and dispositional, which refers to the actual mental property of knowing. I found this quite useful because with the 10x efficiency proposed by AI evangelists, even human understanding becomes more and more dispositional: who cares about knowing when you can just do do(o).</li><li>A related thing relates to Turing test, which many are still hung up on. Is it important for us to be able to tell the difference between a human and a machine? According to conversation analysts (and maybe others too), it is, because we humans need to put things in a box, and the box defines quite a lot of how we respond. According to my study mate Perplexity: <em>“Turing test remains a convenient shorthand for the gap between appearing intelligent and actually being intelligent in a deeper sense.“</em> Which brings me back to the previous bullet: we should start doing the Turing test for humans as well. Slop is taking over.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7f7c0217e1d9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Design process is dead — long live the design process]]></title>
            <link>https://pamihekanaho.medium.com/design-process-is-dead-long-live-the-design-process-49cff8c6cb54?source=rss-40606bf33fef------2</link>
            <guid isPermaLink="false">https://medium.com/p/49cff8c6cb54</guid>
            <category><![CDATA[design-process]]></category>
            <category><![CDATA[ai-design]]></category>
            <dc:creator><![CDATA[Pami Hekanaho]]></dc:creator>
            <pubDate>Tue, 03 Mar 2026 16:52:52 GMT</pubDate>
            <atom:updated>2026-03-03T16:56:40.344Z</atom:updated>
            <content:encoded><![CDATA[<p>This is not the first time* I grump about this and my opinion has not really changed although the hype machine reports giant AI leaps every day, but I need to grump about it again because I listed to Lenny Rachinsky’s podcast where he interviews <a href="https://www.lennysnewsletter.com/p/the-design-process-is-dead">Anthropic design lead Jenny Wen</a> and it got me in knots I want to untangle.</p><h4>10x bullshit</h4><p>Whenever I click on a link about AI design, I never know whether it is going to be about designing AI (enablers like Anthropic) or applying AI in the product (using enablers to fuel our product), or designing using AI tools. In most cases it is the latter (also in this podcast unfortunately) trying to sell me FOMO about a 10x human employing an army of agents while you go for a walk. Maybe it is a perfect example of my lack of creativity that I cannot think of anything that I want to outsource. Or maybe I have managed to trim out all the <a href="https://en.wikipedia.org/wiki/Bullshit_Jobs">bullshit jobs</a> from my current job.</p><p>10x makes me dizzy because I believe it adds to bullshit. Traditionally engineering has been considered so expensive that we’ve needed research, product management and design to sift through bad ideas before valuable engineering time is spent. With AI tools, everything is faster, building feels cheaper and is accessible to more roles, which easily results in more spaghetti thrown against the wall. We value speed and quantity, but there is a hidden cost in managing all the spaghetti. Not just the amount of things built but for example, Wen mentions that she spends a large amount of her time reading the company Slack to keep up-to-date with what is happening in the company.</p><p>A lot of the 10x is about appearances, which adds to the bullshit and the hidden cost. When everyone needs to produce something to showcase their 10x, there will be more others will need to (have their agents) consume. Unless the recipients let go which is when it becomes pure bullshit: AI slop to the void.</p><h4>Validation yeah</h4><p>A lot of what Jenny Wen says resonated with me because it is what I do or try to do. I have never been a process person, but I jump to conclusions and have had to create steps that force me to pause. I also prefer that there is a shared goal and a some kind of path of how to get there, and maybe even a hypothesis of what are the best alternatives (hence 10x tiring me). Wen says Anthropic is still doing all these things: there is a strategy, they do user research, use Figma e.g. for rapid ideation of alternatives and crafting design details because coded “prototypes are linear”. Like always, it makes sense to choose the best tool for the job. Her advice for designers is to let go of gatekeeping, facilitate the creation process of others, and polish with developers, which I think makes sense and was always the case in collaborative teams.</p><p>The same goes for process: if the process is not helping achieve the outcome, it is just an empty ritual, checklist to go through so that we can feel good about ourselves. Curiously, when I was looking at the <a href="https://www.researchgate.net/publication/394994087_Decentring_the_human_-_Nexus_analysis_of_Google&#39;s_People_AI_design_Guidebook">Google AI design process</a>, this felt like part of the message. Trust Google: we have a process, and doesn’t it feel better to be under the illusion that something guides the design of the LLMs themselves?</p><p>Another thing that Wen promoted was to for leaders to engage in “low-leverage tasks” from eating your dogfood to contributing into the actual product (with design, bug tickets, pushing pixels). I felt somewhat validated this and her current task split which pretty much matches mine: Figma and code (and writing).</p><p>What is then left for humans, and specifically for designers to do? Not taste because AI will apparently be good at taste, which I found quite interesting because taste is a construct tied to place, time and power structure and not even a necessary component of a useful and usable interface. For Wen, taste is a “clear, unwavering ideal of what good looks like”. Well, I don’t agree but I can see how someone with a design background might think that.</p><p>Then what is left for humans, but most likely not for designers? Decision-making and conflict resolution, and the liability. If something goes wrong, it’s the human head on the chopping block. Sounds like fun times.</p><p>P.S. I discussed this podcast with Gemini. One of my few friends.</p><p>*Here about <a href="https://pamihekanaho.medium.com/what-is-ux-anyways-f90a781f2a24">AI tools</a>, about <a href="https://pamihekanaho.medium.com/tired-job-stories-7bc9d889c2fe">lost power</a>, <a href="https://pamihekanaho.medium.com/design-talk-now-and-then-516cd2f11a19">history of UX</a> and probably there is more…</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=49cff8c6cb54" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What bad could happen?]]></title>
            <link>https://pamihekanaho.medium.com/what-bad-could-happen-d628eb716076?source=rss-40606bf33fef------2</link>
            <guid isPermaLink="false">https://medium.com/p/d628eb716076</guid>
            <dc:creator><![CDATA[Pami Hekanaho]]></dc:creator>
            <pubDate>Fri, 23 Jan 2026 11:11:37 GMT</pubDate>
            <atom:updated>2026-01-23T11:11:37.760Z</atom:updated>
            <content:encoded><![CDATA[<p>This week “the influenti” have been congregating in Davos talking about this and that. At the same time my social media bubble is raising awareness to Iran, Minnesota and still a bit to Gaza, Venezuela, Ukraine, and countless other places where things are bad. Every day my Finnish newspaper says that NOW the world order as we knew it is done with, and you are stupid if you don’t see it.</p><p>The year of Trump has been shocking, of course. Like in his first term, he has shown the ruthlessness of the people in power, which they went to some lengths to hide before. He reduces everything to personal gain (monetary or other) and his second term has brought forward a new ethos of physical attack to get his gains. His posse is the same, media follows in fear of retribution, and there is nothing civil or intelligent in public conversations anymore. This is bad.</p><p>It’s not that I don’t see this and I am sad, but I’m also tired of gasping for air and “resharing for visibility” (which in reality is resharing for status) so that we can gasp for air together. If I play my part and give money to charities that help the affected, boycott the bad guys’ products and services, go to demonstrations and vote, is this enough or am I guilty for standing by when the nazis took over, because that would be bad.</p><p>This is modern day imperialism where the big guys are divvying up the world in their negotiations. Because they can overpower us, we have to give them all we’ve got. Sometimes I wonder what would happen if we just ignored them and global trade ceased to exist, like the alliances we depended on. This feels like a throwback to my cold war childhood where every kid lived in fear of the world exploding any day now. That would be quite bad.</p><p>I feel naive having grown up in a social democratic welfare state where government and the police are trusted, and no one is corrupt because Finnish people are so honest and follow the rules. Friends from countries without this narrative are more aware of the possibility that your ruler is a dick, or rules don’t apply, and thus much less disillusioned when true colours are revealed. On the other hand, I don’t want to become a cynic because that would be bad.</p><p>So pull the blanket up, turn on Netflix and watch Alex Honnold climb up a building hoping he does not fall. Because that would be really bad and you’re stupid if you don’t see it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nXp4MLZYNg8eEVa783vexA.png" /><figcaption>Illustration by Gemini</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d628eb716076" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[2025 wrapped]]></title>
            <link>https://pamihekanaho.medium.com/2025-wrapped-20f4c636a593?source=rss-40606bf33fef------2</link>
            <guid isPermaLink="false">https://medium.com/p/20f4c636a593</guid>
            <category><![CDATA[wrapped]]></category>
            <dc:creator><![CDATA[Pami Hekanaho]]></dc:creator>
            <pubDate>Fri, 05 Dec 2025 12:39:16 GMT</pubDate>
            <atom:updated>2025-12-05T12:40:01.754Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/585/1*x1cCi62rWoljo_WYy8XcBw.png" /><figcaption>Nano Banana created a version of me strangled by surveillance capitalism. I wrapped it up.</figcaption></figure><p>This year I left Spotify and just a day before the Wrapped reels flooded Instagram, I read <a href="https://www.hs.fi/kulttuuri/art-2000011635812.html">this article</a> in Helsingin Sanomat which congratulates Spotify for monetising data mining both by making it feel like a nice thing just for you and by benefitting from the free marketing resulting from people sharing it. And other companies have copied with pride. Victory of surveillance capitalism, they call it.</p><p>I have also surveilled myself this year but badly, because I have the human biases. Anyways, let’s reflect to remember better later.</p><h3>Books</h3><p>I have been keeping track of my reading for some years now. It started so that I can track the formula crime fiction because I tend to forget which ones I already read. But when wrapping up, the list is handy. I have also been listening more audiobooks this year because I got cheap offers of Audible and Podimo. These are the highlights.</p><p><strong>Top — Chimamanda Ngozi Adichie</strong>. I read both Unelmia (Dream Count) and am almost done with Americanah, and they are both excellent, especially in their perspective on racism and white person’s guilt. Highly recommended.</p><p><strong>Nice crime</strong> — I really like Keigo Higashino and Richard Osman’s crime novels. Higashino’s novels are a different take on crime books, where the guilty party is usually revealed from the getgo, and Osman came out with a new, equally entertaining, series in addition to his Thursday Murder Club.</p><p><strong>Not so good </strong>— This year I stumbled upon quite many self-involved novels dwelling in the narrator’s emotional blur (and maybe spiced with weird erotica), and they were all quite boring. Examples include: <a href="https://www.hs.fi/taide/art-2000011071282.html">Aura Nurmi’s Karkuteillä</a>, <a href="https://www.hs.fi/kulttuuri/art-2000010393037.html">Mikko Kauppila’s Terveisin K</a>, <a href="https://www.kirjavinkit.fi/arvostelut/sukellus/">Laura Liimatainen’s Sukellus</a>, <a href="https://www.kosmoskirjat.fi/kirjat/love-island-kovakantinen-kirja/">Amanda Palo’s Love Island</a> and <a href="https://www.hs.fi/taide/art-2000010660363.html">Riikka Pulkkinen’s Viimeinen yhteinen leikki</a> which started nicely but then fell flat. The worst was <a href="https://www.hs.fi/taide/art-2000011443446.html">Inari Fernandez’s Vapaus</a> which I still cannot believe I finished because it was so bad. It could also be that I am too old for these younger voices, because in this same category, Vigdis Hjorth’s Toisto (will be published as Repetition in 2026) was a gem.</p><p><strong>Non-fiction to mention </strong>— Because of the academic endeavours, I read some AI-related books but what I enjoyed the most was Sarah Wynn-Williams’s Facebook gossip book: <a href="https://en.wikipedia.org/wiki/Careless_People">Careless People</a>. Dicks.</p><h3>Music</h3><p>I listen to music less than before and naturally it’s taken awhile to teach Tidal about my tastes, and there was nothing like Willow in 2024 this year but there were nice songs from <a href="https://tidal.com/track/390810616/u">Maribou State</a>, <a href="https://tidal.com/track/427353297/u">Lucy Rose</a> and <a href="https://tidal.com/track/351078357/u">Saimaa</a> to name a few.</p><p>Corto Alto was the most memorable new act from 2025 and the biggest disappointment was <a href="https://tidal.com/album/458538234/u">Arppa’s latest album</a>. Well, actually the biggest disappointment was the Radiohead concert in Copenhagen that did not happen.</p><h3>TV and movies</h3><p>Because I don’t have a list I will be heavily recency biased in this part but for TV, the top for me were (not in order):</p><ul><li><a href="https://en.wikipedia.org/wiki/Adolescence_(TV_series)">Adolescence</a> — gripping, watch if you haven’t already</li><li><a href="https://en.wikipedia.org/wiki/The_Diplomat_(American_TV_series)">Diplomat</a> — someone called this high competence porn which seems accurate, but also I really like ambassador Kate Wiley</li><li><a href="https://en.wikipedia.org/wiki/Task_(TV_series)">Task</a> — touching male performances and a good story. There is just something about Mark Ruffalo’s face that I find hard to resist.</li><li>Beast in me — the story got me hooked and the dynamic between the main characters kept it up until the resolution. I would have been gutted if it weren&#39;t a happy ending.</li><li>Hacks — just lovely</li></ul><p>I feel like I forgot something important but hey, that’s living in this world of high consumption. For movies I don’t really remember anything.</p><h3>Podcasts and sports</h3><p>The biggest event in podcasts was my long time companion <a href="http://www.wtfpod.com/">Marc Maron</a> retiring. I am still saving the last interview for later because not having his reflection as a weekly treasure is too hard to bear.</p><p>Previous years listened to a number of climbing podcasts but they have become repetitive lately and Nugget Climbing went MAHA, which is when I switched to podcasts about women’s football.</p><p>I listen to something more serious (mostly about AI and product building) but less often so it’s nothing worth writing home about.</p><p>In personal sports I truly enjoyed rock climbing and watching women’s football. Both live and on TV. So good!</p><h3>Cultural phenomena</h3><p>Random things from 2025 include:</p><ul><li>I still don’t get the infatuation with Taylor Swift but I have found myself scrolling (and saving??) recipes on Instagram.</li><li>I travelled more than before and it made me happier.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=20f4c636a593" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Becoming AI flat]]></title>
            <link>https://pamihekanaho.medium.com/becoming-ai-flat-fbe545bf95d8?source=rss-40606bf33fef------2</link>
            <guid isPermaLink="false">https://medium.com/p/fbe545bf95d8</guid>
            <category><![CDATA[elitism]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Pami Hekanaho]]></dc:creator>
            <pubDate>Fri, 07 Nov 2025 13:32:04 GMT</pubDate>
            <atom:updated>2025-11-07T13:32:04.032Z</atom:updated>
            <content:encoded><![CDATA[<p>For a while now, there has been <a href="https://www.newsweek.com/ai-changing-how-we-speak-2120824">conversation</a> about how AI use affects our language, in fact, it was <a href="https://www.tandfonline.com/doi/full/10.1080/0144929X.2025.2533352">researched</a> already when Amazon’s Alexa became a part of households. Language is all around us: we pick words and phrases from our social groups, work, cultural products and so forth, and while sometimes they describe something we didn’t have words for before, in other cases they are adopted without noticing, and passed forward even without understanding their meaning. On a cultural level, this is visible for example in different registers (e.g. therapy and hate speech) seeping into the mainstream, English of the Internets blending into other languages, and now the language of AI assistants appearing in our communication even when it’s hand-written by humans.</p><p>It’s a bit early to say, but <a href="https://theconversation.com/how-generative-ai-could-change-how-we-think-and-speak-267118">research</a> has highlighted some potential effects on language:</p><ul><li>Like with any cultural product, we adopt the vocabulary. (<a href="https://arxiv.org/abs/2508.00238">see</a>)</li><li>It increases communication speed and use of positive, emotional language, making the conversation partner feel closer and more collaborative. (<a href="https://www.nature.com/articles/s41598-023-30938-9">see</a>)</li></ul><p>And on cognition:</p><ul><li>Reliance on AI affects cognitive abilities because we want fast solutions instead of slow thought (<a href="https://slejournal.springeropen.com/articles/10.1186/s40561-024-00316-7">see)</a></li><li>Related to above, it makes us lazy. No need to remember, analyse, or make our own decisions. (<a href="https://www.nature.com/articles/s41599-023-01787-8">see</a>)</li></ul><p>The research is most likely applicable to larger groups. Just this week I wrote a Slack message which sounded like the coding assistant, Cursor. And while I have yet to see positive language in my own writing, I am sure it affects the sentiment in AI-written conversations. For example, conversation with an AI with a friendlier, more positive (read: more validating) tone feels more pleasant than one which slaps like a German salesperson. I can also imagine us becoming lazier as information (or hallucination) is even more available than with the good, old Google search, and outsourcing decision-making to AI, especially in low-value use cases (read: <a href="https://pamihekanaho.medium.com/the-more-you-look-the-blurrier-it-gets-a1c84e4ff461">bullshit jobs</a>).</p><p>The risk these studies outline is homogenisation of language, and with it, of thought, values and how we relate to each other: “the same mechanisms that enable efficient collaboration can also erode epistemic diversity and undermine natural alignment processes” (<a href="https://arxiv.org/html/2407.17489v2">see</a>). As I’ve <a href="https://pamihekanaho.medium.com/illusion-of-diversity-c8fc4137afb9">complained</a> <a href="https://pamihekanaho.medium.com/ethical-ai-and-dei-scrapped-together-71f6da82782d">before</a>, corporate diversity is not real, but maybe we will indeed become less capable of aligning, or just less invested in the outcome.</p><p>Kyle Chayka, who <a href="https://amzn.eu/d/7t2rJvK">wrote</a> about algorithms making everything the same, has also written about <a href="https://www.newyorker.com/culture/infinite-scroll/ai-is-homogenizing-our-thoughts">AI homogenising our thought</a> (1): from identical hipster coffee shops comes ‘intellectual beige’. I didn’t read his stuff but grumpy, old women are allowed to be annoyed regardless, and I deemed his perspective elitist, which is what the article (probably) continues with. I mean, if we agree that we are mostly mediocre, and as a result, our intellectual capabilities are mostly beige, then AI reflects (and amplifies) what was already there (like always). In his book, Chayka was nostalgic for a select elite to decide what good is in art, and now he fears for his position of power as a writer. As argued in the articles linked below, AI surfaces the mediocrity in where we may have seen originality before, and makes this mediocrity accessible to more people. This won’t be the end though, I have faith in human creativity (and greed) and there will be consequences: good and bad.</p><p>P.S. Apologies for all these brackets. I want everyone to see that this is intellectual beige from a real human published without AI proof-reading.</p><p>(1) The article is behind a paywall but <a href="https://medium.com/@jwilliams1454/is-ai-flattening-thought-or-did-we-do-that-on-our-own-2b7079594250">this Medium article</a> has a really good response to it. Recommended!</p><p>(2) <a href="https://uxdesign.cc/ai-is-flattening-language-and-redistributing-power-228cb28e54b9">Another one</a> about accessibility</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fbe545bf95d8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[(En)titled]]></title>
            <link>https://pamihekanaho.medium.com/this-week-the-finnish-newspaper-helsingin-sanomat-had-an-article-about-yoga-and-other-spiritual-537ad1e49e93?source=rss-40606bf33fef------2</link>
            <guid isPermaLink="false">https://medium.com/p/537ad1e49e93</guid>
            <category><![CDATA[entitlement]]></category>
            <dc:creator><![CDATA[Pami Hekanaho]]></dc:creator>
            <pubDate>Sat, 01 Nov 2025 11:16:07 GMT</pubDate>
            <atom:updated>2025-11-01T11:17:10.389Z</atom:updated>
            <content:encoded><![CDATA[<p>This week, the Finnish newspaper Helsingin Sanomat had <a href="https://www.hs.fi/elama/art-2000010703580.html">an article</a> about yoga and other spiritual practices not freeing people from the evils of the ego but rather resulting in the opposite: a self-centredness referred to as spiritual narcissism. The article references <a href="https://journals.sagepub.com/doi/abs/10.1177/0956797618764621">a study</a> examining the contradiction between quieting the ego and the self-centrality principle, a psychological universal (1) which argues that “practicing any skill renders that skill self-central, and self-centrality breeds self-enhancement bias”, which in turn, is “<a href="https://journals.sagepub.com/doi/10.1177/0146167298245006">the tendency to describe oneself more positively than a normal criterion would predic</a>t”.</p><p>This term is new to me but I already love it, because it captures the reason social media is so exhausting at times, why many think they are good drivers and have an amazing sense of humour, and why it is so difficult to accept that by definition, most of us must be mediocre.</p><p>The study itself is behind the paywall, so I am assuming a bunch here, but the abstract describes them following a group of yoga and meditation students and assessing their self-centrality and self-enhancement bias after practice and without practice to find that self-enhancement bias was higher in the practice group. If the purpose of the practice is to quiet the ego, it’s working the opposite: boosting self-enhancement.</p><p>Being a currently non-practicing yoga student myself, and one who has not been particularly involved in the philosophy, I have to point out that the purpose of the practice is not to quiet the ego. For one, <a href="https://en.wikipedia.org/wiki/Id,_ego_and_superego">ego is a Western term</a>, but also, if we go by the Yoga Sutras of Patanjali, the purpose (or one of them) is to calm the “<a href="https://en.wikipedia.org/wiki/Vritti">vritti</a>”, the fluctuations of the mind, which is very different.</p><p>However, I believe that the finding itself is probably accurate and repeatable with other (at least Western and commercial) yoga/meditation groups for many reasons. And like the “universal” claims, any practice breeds self-centrality bias, so it’s likely repeatable at cross-fit gyms, NGOs, dog agility groups and any Internet fora (heh) where people are anal about something. It is more about how much you let it define yourself, or let it show in interactions with other people, because it is <a href="https://journals.sagepub.com/doi/10.1177/0146167298245006">a controllable bias</a>, and if not controlled, can make people effing annoying. So I am proposing a spiritual practice for regular checks for self-enhancement with these 5 commandments (2):</p><ol><li>Don’t think you’re better than others. You are probably mediocre, and even if you are the world’s best at something, it is just one or at most a couple of things.</li><li>Don’t think you are more intelligent than others. You only see their outside and definitions of intelligence are flawed.</li><li>Don’t think you work harder than others. You only see a part of their activity.</li><li>Don’t think you’re better than others because of your beliefs or lifestyle choices. Not everyone has the same beliefs and choices.</li><li>If you cannot stop thinking it, at least don’t say it out loud, or imply in social media.</li></ol><p>There was a (somewhat) <a href="https://www.hs.fi/paakirjoitukset/art-2000011594559.html">related column</a> was on Helsingin Sanomat today where historian Janne Saarikivi talks about the current trend of being “on the right side of history”. It was a good reminder of how making the decision on the right side of history is impossible because we are not outside the history, and how the right side depends on perspective and is suspect to shifts. As a small country, Finland is a hustler sucking up to the bigger powers and when power dynamics change, the ideology and justifications change. If time shows we were on the wrong side (losers, or later condemned), we are good at either erasing the era, or coming up with a defensive narrative, and if we chose the winners, we are better than others. But rather than the right side of history, it’s self-enhancement at scale (3).</p><p>(1) ”Psychological universals, or core mental attributes shared by humans everywhere, are a foundational postulate of psychology, yet explicit analysis of how to identify such universals is lacking”. (<a href="https://www2.psych.ubc.ca/~heine/docs/universals.pdf">source</a>)</p><p>(2) The irony. These commandments read as a self-centred practice and annoying behaviour because they make it seem like I, as the author, already excel at these but of course I don’t. I’m a grumpy old woman who writes a blog.</p><p>(3) You are still more likely to be on the better side of history if you condemn genocide.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=537ad1e49e93" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[WORK2025]]></title>
            <link>https://pamihekanaho.medium.com/work2025-5ea6d109ff49?source=rss-40606bf33fef------2</link>
            <guid isPermaLink="false">https://medium.com/p/5ea6d109ff49</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[conference]]></category>
            <category><![CDATA[sociology]]></category>
            <dc:creator><![CDATA[Pami Hekanaho]]></dc:creator>
            <pubDate>Fri, 22 Aug 2025 10:49:30 GMT</pubDate>
            <atom:updated>2025-08-22T10:49:30.995Z</atom:updated>
            <content:encoded><![CDATA[<p>For the past three days, I have been participating in<a href="https://work2025.fi/"> a conference in Turku</a> from the comfort of my own place. I was lucky to be selected as a presenter as well with the topic of <em>Decentring the human — Nexus analysis of Google’s People + AI guidelines</em>.</p><p>The theme of the conference is ”Work in the era of unruly AI” and there were <a href="https://work2025.fi/streams/">many streams</a> covering a wide range of topics within this broader theme. In addition to the keynotes, I participated in Technology and everyday life, Exploring the human in AI worldbuilding (where I presented), and Future work and AI beyond platforms streams.</p><h3>Things that stood out for me</h3><p>As this was an academic conference, there were captivating presentations with critical evaluation of existing and emerging practices. There were also initial thoughts, snippets of data, and requests for feedback. The keynote speakers are usually established researchers and can thus offer bolder, more generalised views with authority, while the less established ones share their specific data and stick to it with less flair.*</p><p>Of the former, I specifically enjoyed <a href="https://edwebprofiles.ed.ac.uk/profile/janja-komljenovic">Dr Janja Komljenovic</a>’s presentation about AI effect on universities. She focused on the structural and economic aspects rather than on how the education system itself needs to update and the terms she used were:</p><ul><li>Assetization: turning things into assets to bring continuous value to the asset controller</li><li>Rentiership: strategy to generate economic benefit from controlling assets rather than direct production or service delivery</li></ul><p>The other reason it hit home was because she talked about crawling academic papers for training data. The publishers, who already owned the papers they got for free and sold to other researchers for profit, are making another round of profit selling the same data to big AI companies for model development. She also mentioned collecting similar data from students via the Turnitin product which a mandatory step in paper submission in many universities but requires to sign Terms of Agreement which give the company the licence to use the content for free.</p><p>The other impressive presentation was from <a href="https://www.benjaminshestakofsky.com/">Benjamin Shestakofsky</a> who was talking about the effect of venture capital as an investment model on what happens in startups. A very convincing story centring around <a href="https://www.ucpress.edu/books/behind-the-startup/paper">his field study</a> as an employee in a Silicon Valley startup. He attributed a lot of the harm startups produce to the founders being conditioned to work for the investors who push for quick profit. As a managerial paradigm this translates to</p><ul><li>Goal = rapidly inflate valuation</li><li>Logic = scale</li><li>Practices = experimentation</li><li>Ideology = techno-solutionism and progress</li></ul><p>which really does not encourage long-term product development and doing something good.</p><p>The last keynote was by <a href="https://cssh.northeastern.edu/faculty/steven-vallas/">Steven Vallas</a> about platform capitalism. He criticised his community for not having well-developed theoretical frameworks and for not taking into account the structural environments of platform capitalism. He represents the perspective where everything is a platform with data as capital, and all of us platform workers.</p><p>His answer to why the platform economy has gained power so rapidly was interesting and he called it the<em> sublation effect.</em> This refers to platforms absorbing the features of prior mechanisms (markets, hierarchies, networks and states) that have governed the extraction and distribution of value, which results in a new foundation for the extraction of value (Vallas and Schor 2020). With this, they manage to evade legacy regulations, and social scientific theories.</p><p>His last question was about where platform capitalism leads, for which he listed four speculative scenarios:</p><ol><li>encode authority in algorithms (operating in a “discourse-free” mode)</li><li>privatise core state functions (housing, transportation, communications, currency)</li><li>erode support for liberal democracy (institutionalised governance without accountability)</li><li>privilege consumers at the expense of workers (servant economy)</li></ol><p>Grim.</p><h3>Themes visible in the sessions I attended</h3><ul><li>How people feel about using AI in their work, how does AI change it for individuals and processes etc. (many professions represented)</li><li>How AI changes/threatens professional identity and expertise</li><li>How AI is not an insular thing that just happens, but requires a lot of preparation and change in the environment to make space for it</li><li>Data work, specifically the exploitation of lower-income groups</li></ul><p>The last one is somewhat awkward. While I agree that this is an important topic, the people who I saw presenting about it were all white. When we learned that this happens everywhere, one (white) person even commented that it’s different in the global north, because people (even prisoners) can choose. So many assumptions there.</p><p>Could also be that I am slightly defensive here because after my presentation, I got the question along the lines of …decentring the human with agentic AI and all, but what about the fish in the rivers that dry because of data centres, or the exploited data workers?</p><h3>And then?</h3><p>We are very human-centred, it is hard to overcome. Even if the interconnectedness is thought about, we primarily focus on the human while the environment seems the more impactful driver for our futures, be it the climate or organisms, the financial structures or technologies. Maybe it is because they scale and humans are not unified, which then paves way for dictators. My own topic, at least, demonstrated that the materiality of the technology itself and the related interactions, drove the change in the process.</p><p>*I am a wannabe keynote speaker with lots of opinions and little substance.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5ea6d109ff49" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>