[{"content":"Let\u0026rsquo;s get ready to build!\nI installed OpenCode and Pi and was amazed by how smooth the installation goes. Really quick and nice. And after being away from the command line for 10 years, it\u0026rsquo;s so nice to see the modern TUI animations. Wonderful!\nI want to use the Claude model family for production, so I signed up for a Pro account. Works great. Installed the Claude Code cli client, authenticated and we\u0026rsquo;re all good. Then I go to OpenCode to switch the model to Claude Sonnet. It asks me for an API key. No panic, just go to the web dashboard and look around in the settings. Nothing. Checked the documentation and found Anthropic has 2 kinds of subscriptions: chat / cowork (including code) and developer (platform). And the accounts don\u0026rsquo;t seem to be linked. AND THAT PISSES ME OFF. I\u0026rsquo;m okay with paying for stuff. But I am not okay with this intransparant bullshit. I haven\u0026rsquo;t figured out what to do yet. I don\u0026rsquo;t feel like also paying for an API key.\nAnyway, I was able to have Hermes use the OAuth authentication for my Claude models so that was all good. Until I had to re-authenticate the next session because the token expired. And found the same in the Claude Code cli. I have to re-authenticate often. Why?\nDrank 4 coffees to try to calm down.\nIn the bootstrap note, I wrote about my struggle to setup the full integration with Linear. After switching to Claude Sonnet 6, I asked it to look at the Linear skill and fix possible issues; it quickly found the bug and fixed it fairly quickly. Full linear integration is running smoothly now. And it\u0026rsquo;s super cool. I can manage all product related tasks from Linear and assign it to either a human (i.e. me), or an agent (Donny, my agent is called Donny).\nAdded a bunch of other polyhedrons to Hugo static folder because polyhedrons are cool.\nTODO:\nCalm down Find out how to use Clause with a n00b subscription in alternative harnesses; Find out how to not to authenticate every time I start a new Claude code session. ","permalink":"https://thishollow.earth/posts/tech-notes/tn-claude-opencode-pi/","summary":"\u003cp\u003eLet\u0026rsquo;s get ready to build!\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cp\u003eI installed \u003ca href=\"https://opencode.ai\"\u003eOpenCode\u003c/a\u003e and \u003ca href=\"https://pi.dev\"\u003ePi\u003c/a\u003e and was amazed by how smooth the installation goes. Really quick and nice. And after being away from the command line for 10 years, it\u0026rsquo;s so nice to see the modern TUI animations. Wonderful!\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eI want to use the Claude model family for production, so I signed up for a Pro account. Works great. Installed the Claude Code cli client, authenticated and we\u0026rsquo;re all good. Then I go to OpenCode to switch the model to Claude Sonnet. It asks me for an API key. No panic, just go to the web dashboard and look around in the settings. Nothing. Checked the documentation and found Anthropic has 2 kinds of subscriptions: \u003ca href=\"https://claude.com\"\u003echat\u003c/a\u003e / cowork (including code) and developer (\u003ca href=\"https://platform.claude.com\"\u003eplatform\u003c/a\u003e). And the accounts don\u0026rsquo;t seem to be linked. AND THAT PISSES ME OFF. I\u0026rsquo;m okay with paying for stuff. But I am not okay with this intransparant bullshit. I haven\u0026rsquo;t figured out what to do yet. I don\u0026rsquo;t feel like \u003cem\u003ealso\u003c/em\u003e paying for an API key.\u003c/p\u003e","title":"Claude, OpenCode, Pi"},{"content":"I worked for and with many companies and organisations.They are all successful and they are all struggling. Each successful in its own way. And each struggling in its own way. Although, that is how they like to think about themselves: we are unique. Well, yes, you are unique, just like everybody else. And just like everybody else, you are not as unique as you think you are. I started to notice patterns, especially in the struggling. This essay is about those patterns.\nIn this post I distill common patterns of companies stumbling towards success into a maturity model of the consultancy versus Saas struggle. How to deal with the tension between project and product. Or, implied in “maturity”, how to get from project to product.\nI start with a section on definitions, to avoid confusion. Then I will give a short description of each of the 5 maturity stages:\nConsultancy Automated consultancy Hybrid Big Head SaaS Long Tail SaaS Then, we’ll have a look at how this model is especially relevant for service providers in the environmental supply chain risk sector. We wrap it up with a short reflection, a use case and a future outlook.\nDefinitions Consultancy: organisations that sell the value of their time on a project basis\nSaaS: organisations that sell the value of a service or product on a subscription basis\nMaturity: a description, not a judgement. A higher maturity stage is not inherently better. The goal is to know which stage you are in, and to make a conscious choice about whether to move.\nConsultancy builds trust, SaaS builds leverage The tension between them is not a problem to solve. It is the business.\nMaturity Stages 1. Consultancy Every delivery is custom. Every project starts from scratch. Your value is in the people who show up and do the work. You get paid for time and expertise, and you deliver something the client calls theirs, even though it sits on your laptop.\nTrust is built person by person, project by project.This is fun.\nThe problem is, it doesn’t scale. Every new client requires the same investment. Revenue grows linearly with headcount. Your most valuable knowledge lives in the heads of your best people, and when they leave, it walks out the door with them.\nTypical signals: long sales cycles, high margins per project, hard to forecast revenue, dependence on senior people, the same problem solved for different clients every six months. No incentive to extract and distill common patterns into revenue engines.\n2. Automated Consultancy You spent some time solving the same problem for different clients. You know what the answer looks like before you start. So you build tools, scripts, templates, models, internal platforms that do the repeatable parts faster. The last mile is still custom, still human, still consultancy. But the middle is increasingly automated.\nThis stage has two shapes, which look similar from the outside but lead in very different directions.\n2a. Vertically scaling automation (projects) The automation makes your projects faster and cheaper to deliver. You can do more projects with the same team, or the same projects with a smaller team. Revenue still comes from projects, but your margin per project improves. You are selling time more efficiently.\nSide note: this is where most AI efforts go in most companies, doing the same thing but faster or more (Jevons paradox).\n2b. Horizontally scaling automation (subscriptions) The automation becomes something you can deliver repeatedly to multiple clients simultaneously. Instead of one project at a time, you can run the same process for ten clients at once. The business model starts to shift: recurring revenue, subscriptions, ARR, MRR, churn. You are no longer selling time — you are selling access to a system.\nThe fork between 2a and 2b is one of the most consequential decisions a consultancy makes, usually without realising it is making it. 2a makes you a more efficient consultancy. 2b is the beginning of a product company.\nSide note: “scaling” is often interpreted as “technical scaling”: can you process the world? Can you handle 1 million HTTP requests? “Scaling” here means scaling across the organisation: HR, legal, marketing, front office, back office, MT. Often, only 1 or 2 business functions are scaled and the others are left as an afterthought.\nTypical signals: outcomes from internal tools that clients keep asking for repeatedly; the same analyst running five client accounts simultaneously; your 25 year old consulting colleague who did a Python course in university asking for a company CodeRabbit subscription.\n3. Hybrid You have a product, or something that looks like one. But it does not run without your people behind it. Configuration is done by your front office delivery team (operations or account management or forward deployed engineers). Exceptions are handled on the fly by your backoffice teams. The client sees a clean interface; your team sees a rat\u0026rsquo;s nest of bespoke parameters, manual overrides and client-specific edge cases held together with goodwill,institutional memory, rubber bands and duck tape.\nThis is the stage where companies spend the longest time, and the stage they most often mistake for something else. You will hear: \u0026ldquo;we are basically SaaS.\u0026rdquo; You are not. You are a consultancy with a good interface.\nThat is not a failure. The hybrid model can be highly profitable, especially in (new) markets where trust and interpretability matter more than scalability. Many companies should stay here. The mistake is not being in the hybrid stage; the mistake is not knowing you are in it, and building a sales motion or an org structure optimised for a stage you have not reached yet.\nThe transition out of hybrid requires hard decisions: what gets built into the product, and what stays consultancy? The configuration that your ops team does manually, can it become self-service? If not, stop doing it. The exceptions your engineers handle, are they bugs or features? If it’s a feature, your PM should own it and decide if it goes on the backlog or not. The institutional memory, can it be documented, automated, encoded into the system? If not, it doesn’t exist.\nWhat is hard is that is the point where decision making has to be delegated to the organisation. The role of the top management changes from making decisions to leading by creating principles, guidelines and a strategy that allows the organisation to make decisions.\nTypical signals: clients think they are using software; your team knows they are running a service; onboarding takes weeks; the product manager and the delivery manager are fighting about the same backlog; front office teams (support, FDE, account management, customer success) are ingesting client feedback and dumping it in the organisation; engineering teams get overloaded with each additional client.\n4. Big Head SaaS The product is real. Clients can configure it themselves. Your ops team no longer runs the system for each client; they run the system once, clients use it. A lot of your margin is still eaten by front office teams: onboarding still requires effort, relationship, trust-building, this is the \u0026ldquo;big head\u0026rdquo; of the distribution: high-value clients who need hand-holding, onboarding support, account management, bespoke reporting. But the core delivery is automated.\nPricing is value-based. You charge for the outcome the product enables, not for the hours it took to deliver or operate. Revenue scales with clients, not with headcount but is not yet independent of it. Adding a new market or a new vertical requires meaningful investment, but it creates a multiplier rather than a linear addition.\nThe growth model here is typically a mix of referral and direct sales. Your clients talk to prospects. Your sales team closes deals that your account managers retain. The pipeline is manageable and foreseeable for the first time.\nThe risk at this stage is product debt: the pressure to close deals leads to promises, workarounds and special configurations that slowly pull the product back toward hybrid. Every client-specific customisation is a step backward. Discipline here is one of the hardest organisational skills to maintain.\nTypical signals: ARR becomes forecastable; sales cycles are getting stable; the same demo closes most deals; your biggest clients are also where you spend most front office costs; engine of growth is a mix between paid and referral.\nSide note: “Adding a new market or a new vertical requires meaningful investment, but it creates a multiplier rather than a linear addition.” is true as long as you realise that this new vertical is probably less mature than its parent, a step backward.\n5. Long Tail SaaS The product sells itself, or nearly. A new client can sign up, configure and get value without speaking to anyone at your company. The long tail of smaller clients can be served at near-zero marginal cost. Growth is driven by the product itself: through virality, through content, through a freemium model, through integrations that put the product in front of the right people at the right moment.\nPricing is cost-based or usage-based. The engine of growth is paid, viral, or a mix. Margins are lower per client than in Big Tail SaaS, but there are a lot of clients. Scaling within a market is exponential, not linear.\nThis is the stage most startup pitches describe. It is also the stage that almost no B2B company in a new, trust-dependent market reaches quickly. The road to long tail SaaS (especially for new products in new markets) runs through every preceding stage. Trust is built by humans before it is maintained by software. The product that eventually scales to ten thousand clients was first tested on ten clients who needed a lot of handholding.\nThe companies that try to skip to this stage typically end up with a very polished, hollow, interface for a problem they have not fully understood yet. The consultancy stage is not a failure state to escape. It is the R\u0026amp;D process, it is product discovery.\nTypical signals: self-serve sign-up; low or zero marginal cost per new client; product-led growth metrics; the support team handles volume, not complexity; engine of growth is a mix between paid and viral.\nThis is what going wrong looks like A common cause of unaware struggling: a company built successful service X up to Big Head SaaS level, introduces service Y for upselling and x-selling without realising that service is not at SaaS level yet.Sales sells service Y like service X. The company can’t deliver Y like X and has to rely on R\u0026amp;D and sales engineers to fill in the non-automated gaps. Sales and support get frustrated because they are flooded with client feedback about overpromised service Y. Management fixes the gaps with people (hiring more customer support engineers, offloading operational tasks to R\u0026amp;D and product development), leading to higher marginal costs, slower R\u0026amp;D and product development, increased churn, higher pressure, burnouts, misery, doom, apocalypse. Well, ok, you get the gist.\nThe key is that probably everybody is right, but right on a different level. “We are SaaS”. Yes, service X is SaaS, but that doesn’t mean your whole company is SaaS. Every new product or service you introduce starts as consultancy and needs to go through the maturity stages. Winning companies understand this and set up an organisation structure and development flow that accelerates this.\nEnvironmental Supply Chain Risk Sector Sure, but why is this relevant? What does it matter? Who cares? Jeff started Amazon as a Long Tail SaaS from the start and look at him now! That is correct. This is not a recipe or model for every business in every sector. But if you are building a new product in a new market, this is what you go through. Because the most important thing, the quod sine qua non, the thing you are doing that will make or break your product is: trust. The market won’t buy your product if it doesn’t trust it. So what you have to do is build trust. You might call it sales, but it’s not sales in the classic term, it’s relationship building, building trust and product discovery at the same time. Rename it to consultative selling, sign the deal after 18 months lead time, have another front office team take over part of the trust building (Account Management, Sales Engineering, Customer Success, Delivery, R\u0026amp;D), it’s still not sales, it’s consulting. You are building trust and discovering what your product actually looks like. And that is good. But that’s not the goal.\nThis is especially true in the environmental supply chain risk sector. A new market, without standards, benchmarks or decades of best practices to refer to. In this new market there are ton of tiny, niche service providers each with their own new product. Some striving to be a SaaS company, some happy with some hybrid form, some … .\nAnd that’s why models like this can be helpful:\n1. The market is too young for shortcuts. There are no established benchmarks, no dominant platforms, no decade of best practices to copy. Every company is figuring out what the product actually is while simultaneously trying to sell it. The maturity model gives you a map for that process. A way to talk about it internally without it feeling like an admission of failure.\n2. Trust takes longer here than almost anywhere else. Your buyers are sustainability managers, procurement leads and financial controllers making decisions that will end up in annual reports and regulatory filings. They will not buy a self-serve product from a company they do not know. The consulting stage is not a phase to escape, it is the mechanism by which trust gets built. Understanding that makes it a strategic asset rather than an embarrassing legacy.\n3. The consolidation is coming. The ESG data and supply chain risk space is full of niche providers, each covering one piece of the puzzle. At some point, through acquisition, partnership or attrition, the market will consolidate. Knowing where your service sits on the maturity curve is directly relevant to how you will be valued, how you will be integrated and whether you are the acquirer or the acquired.\nYou cannot escape building trust and discovering what your product actually is. You do that through consulting. In early stages, it looks like pure consulting, later via other business functions as well. The trick is:\nTo be aware of this process To make sure you don’t get stuck in consulting and promote your service or product to a higher maturity level when it’s ready When to move: transition triggers No company wakes up one morning and decides to advance a maturity stage. The transitions are not planned. They are recognised, usually after the current stage has already started causing pain.\n1 → 2 Consultancy to Automated Consultancy You are ready when\u0026hellip;: You have solved the same problem for the third client; Your best analyst spends more than 30% of their time on tasks a script could do; Clients are asking why delivery takes so long.\nThe risk of waiting: Your best people leave because the work stopped being interesting. Margins erode as headcount scales with revenue.\n2a → 2b Vertical to Horizontal automation You are ready when\u0026hellip;: Clients keep asking for the same outcomes of your internal tooling; You are running the same process for more than five clients; A new client does not require a fundamentally different setup.\nThe risk of waiting: You keep optimising delivery efficiency instead of building leverage. The business stays linear.\n2 → 3 Automated Consultancy to Hybrid You are ready when\u0026hellip;: You have a thing that looks like a product. Clients refer to it by name. Sales is starting to demo it. But your ops team is still running it manually for every client.\nThe risk of waiting: Sales overpromises what the product can do. Delivery scrambles to catch up. The gap between what is sold and what is delivered becomes a management problem.\n3 → 4 Hybrid to Big Head SaaS You are ready when\u0026hellip;: Clients can configure the core use case without your team\u0026rsquo;s involvement. Onboarding still requires effort, but delivery does not. Your ops team is managing exceptions, not running the system.\nThe risk of waiting: Client-specific workarounds accumulate. The product slowly becomes hybrid again. Technical debt disguised as customer success.\n4 → 5 Big Head to Long Tail SaaS You are ready when\u0026hellip;: Small clients are signing up without talking to anyone on your team. Support tickets are about volume, not complexity. A new market\nsegment could be reached without a new sales team.\nThe risk of waiting: You over-invest in enterprise clients at the expense of the engine that would scale the business. The long tail never develops.\nWrapup In the intro I wrote about struggling companies. In reality, many companies are struggling but are not aware of their struggling.This maturity model can help companies understand if and why they are struggling.\nBut first, please note that this is a model, not reality. No company will be in exactly one of the maturity stages. Most companies are somewhere on a gradient from consultancy to SaaS. And companies might have several services, each in their own maturity stage.\nThere can also be a difference in intention or ambition. Some companies are laser focused at becoming a Long Tail SaaS. The classic Silicon Valley startup archetype. They know what they want. I call these intentional SaaS companies. Other companies happen to find themselves at one of the stages at some point in their existence. Most companies are of that type. In most cases it’s unplanned. Reasons can be plenty: new enabling or accelerating technologies, a new hire, changing market demand, leadership change. I call these serendipitous SaaS companies.\nThe Consultancy to SaaS maturity model can help both intentional and serendipitous companies:\nassess the current maturity stage of your services and products; explore implications for business functions like GTM, engineering or organisation roles; understand if it makes sense to go to the next stage and how to get there; Understand if and how finance can accelerate service \u0026amp; product development. The next essay in this series explores implications on business functions like Go To Market (GTM).\nLife is great.\n","permalink":"https://thishollow.earth/posts/essays/consultancy_to_saas/","summary":"\u003cp\u003eI worked for and with \u003ca href=\"https://thishollow.earth/CV_Arjen_Vrielink.pdf\"\u003emany companies\u003c/a\u003e and organisations.They are all successful and they are all struggling. Each successful in its own way. And each struggling in its own way. Although, that is how \u003cem\u003ethey\u003c/em\u003e like to think about themselves: \u003cem\u003ewe are unique\u003c/em\u003e. Well, yes, you are unique, just like everybody else. And just like everybody else, you are not as unique as you think you are. I started to notice patterns, especially in the struggling. This essay is about those patterns.\u003c/p\u003e","title":"Consultancy to SaaS; a maturity model"},{"content":"Super Humans and Super AI\nI had an issue with my buttondown newsletter. I wrote to support. Got an automated reply telling me they got my mail, they are looking into the issue but since they only have limited staff, working in a US time zone, it might take a bit for a proper reply. Within 1 hour, Steph is in my inbox, kindly explaining the issue and how they would resolve it. I said thank you for the quick answer Steph and for being a kind human Steph and Steph said no problem arjen, we don\u0026rsquo;t like talking to chatbots, and we don\u0026rsquo;t want to make you all deal with them either. \u0026lt;3 (\u0026lt;- I got that one from the emoji museum).\nOn my site this \u0026#x1f989; renders as an emoji. I wonder how the newsletter engine renders it.\nsnails\nI have a math professor friend. He is in a depression now since the Navier Stokes incident. OpenAI had an LLM solve a difficult math problem. And that\u0026rsquo;s a problem. Because mathematics was never about solving the problem, as Terrence Tao explains elegantly:\n(\u0026hellip;) But there is now a scenario in which an autonomous AI harness, backed by an enormous amount of computational resources, performs this entire iteration internally, and ends up producing the final ansatz, and thence the solution to the Navier-Stokes regularity problem, while the AI company running the harness keeps the process to arrive at that ansatz almost completely out of public view. Technically, one of the most prominent open problems in mathematics would now be solved; but there would be almost no value added to mathematics as a consequence.\n-\u0026gt;\nFundamentally, problems in pure mathematics, such as the global regularity problem, serve a different purpose than immediately practical problems, such as that of finding a cure to a specific disease, or increasing the energy efficiency of some engine. In most cases in pure mathematics, the problems are posed not because we desperately want the solution to these problems in and of themselves, but because we have seen from past experience that human-directed efforts to solve these problems tend to spur further development of the field through the efforts to solve such problems, and then to digest any partial or complete solutions that emerge for further insights. Prematurely solving the problem by purely AI-powered methods - particularly without full transparency into the solution process - can contaminate this process to the point where it actually becomes a net negative for the progress of mathematics as a whole.\n-\u0026gt;\n(\u0026hellip;) The point remains that there is a substantial opportunity cost in converting a historically productive and motivating problem (such as Navier-Stokes regularity) into a mere viral social media post advertising some benchmark progress, rather than actually advancing the field and developing the next generation of both problems to ask, and people to work on them.\nAnd if even Terry get\u0026rsquo;s worried, we should get worried as well.\nMy friend feels like a fish who has been told that, from tomorrow, there won\u0026rsquo;t be any water anymore. Ever.\nthis pennywise album rocks hard\nAn evening stroll\nI was walking my dog the other day when we ran into the ox\n\u0026ldquo;Good morning cow\u0026rdquo;, I said\n\u0026ldquo;I am not a cow\u0026rdquo;, the ox said\n\u0026ldquo;Oh, I\u0026rsquo;m sorry\u0026rdquo;, I said\nMe and my dog, we walked on\n\u0026ldquo;I\u0026rsquo;m on ox.\u0026rdquo;, the ox said, \u0026ldquo;An ox!\u0026rdquo;\n\u0026ldquo;Sure\u0026rdquo;, I thought, \u0026ldquo;An ox, and I\u0026rsquo;m a lumberjack\u0026rdquo;\nI ran into my friend the other day. She left her job about the same time as I did. In similar circumstances as I did. And started as a freelancer about the same time as I did. We sat on a bench in the autumn sun, overlooking the Utrecht canal and found ourselves in a random state of happiness. Happiness is easy, but difficult.\n\u0026ldquo;Do you know the story of Nobunaga and the three birds?\u0026rdquo;, my son asked.\n\u0026ldquo;No\u0026rdquo;, I said\n\u0026ldquo;Nobunaga, Hideyoshi and Tokugawa were watching a cuckoo bird waiting for it to sing, but the bird wouldn\u0026rsquo;t sing. Tokugawa Ieyasu says to the bird \u0026lsquo;Little bird, if you don\u0026rsquo;t sing I will wait for you to sing\u0026rsquo;, Hideyoshi says \u0026lsquo;Little bird, if you don\u0026rsquo;t sing, I\u0026rsquo;ll make you sing\u0026rsquo;. Then, Nobunaga says \u0026lsquo;Little bird, if you don\u0026rsquo;t sing I will kill you\u0026rsquo;\u0026rdquo;.\nNobunaga, Hideyoshi and Tokugawa are the three unifiers of Japan. You have to be patient, active and ruthless to move ahead in life.\nMost people wait for the bird to sing, without being patient. Some people try to make the bird sing, without a mitigation plan. And only very, very few people are prepared, are willing to kill the bird.\nMy friend killed the bird.\nOh, and I finally published a draft of the Consultancy to SaaS essay. I\u0026rsquo;m not yet satisfied. It needs editing and polishing and I need to add figures and diagrams and I have to elaborate the use case. That\u0026rsquo;s why I didn\u0026rsquo;t publish on Medium yet. But it\u0026rsquo;s good enough to publish here. Let me know your feedback. (I had lunch and had another look and I am actually quite excited, read it!)\nAll systems go for GTM next week!\ncustomatisation\nPublishing early is about gettings things done and not chasing perfection. It will block you. Like when I launched my site last week, I knew it was crap. But it was something and good enough. And while walking my dog, we not only run into animals all the time, I also process stuff and cook up new ideas:\nOn a Sunday evening walk, a better structure for my site popped up (consult, build, write) and from there everything became more logical and fell into place in the updated and restructured version; On a Thursday afternoon walk I thought about a collaboration structure for freelancers to stay engaged with all the wonderful people I am talking to these days. One step beyond \u0026ldquo;let\u0026rsquo;s stay in touch\u0026rdquo;, without any formal or binding stuff. Do stuff together, know there are people willing to listen and just be happy in general. dog in basket\n","permalink":"https://thishollow.earth/posts/test_34/","summary":"\u003cp\u003eSuper Humans and Super AI\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003eI had an issue with my \u003ca href=\"https://buttondown.com\"\u003ebuttondown\u003c/a\u003e newsletter. I wrote to support. Got an automated reply telling me they got my mail, they are looking into the issue but since they only have limited staff, working in a US time zone, it might take a bit for a proper reply. Within 1 hour, Steph is in my inbox, kindly explaining the issue and how they would resolve it. I said thank you for the quick answer Steph and for being a kind human Steph and Steph said no problem arjen, we don\u0026rsquo;t like talking to chatbots, and we don\u0026rsquo;t want to make you all deal with them either. \u0026lt;3 (\u0026lt;- I got that one from the emoji museum).\u003c/p\u003e","title":"Test 34"},{"content":"I write in near real time about technology, human systems and environmental risk , the same three pillars I consult on.\nBlog Posts — the main feed. Blogs about life; short updates on what I\u0026rsquo;m building, deciding and thinking about, as it happens.\nTech notes Tech notes — tools, stack decisions and technical opinions. The nuts and bolts behind CYNTHIA, BERT and WALTER.\nEssays Essays — long form pieces on technology, human systems and environmental supply chain risk, in parallel with Medium.\nScalpel Coming soon. you don\u0026rsquo;t want to know .\nTwitter / X @arjenvrielink — shorter takes, links and running commentary.\nHave a look at my raw, unordered and unstructured ideas — let me know which one you want me to write about.\n","permalink":"https://thishollow.earth/write/","summary":"\u003cp\u003eI write in near real time about technology, human systems and environmental risk , the same three pillars I \u003ca href=\"/consult\"\u003econsult\u003c/a\u003e on.\u003c/p\u003e\n\u003ch2 id=\"blog\"\u003eBlog\u003c/h2\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"/posts/\"\u003ePosts\u003c/a\u003e\u003c/strong\u003e — the main feed. Blogs about life; short updates on what I\u0026rsquo;m building, deciding and thinking about, as it happens.\u003c/p\u003e\n\u003ch2 id=\"tech-notes\"\u003eTech notes\u003c/h2\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"/posts/tech-notes/\"\u003eTech notes\u003c/a\u003e\u003c/strong\u003e — tools, stack decisions and technical opinions. The nuts and bolts behind \u003ca href=\"/posts/products-concepts/cynthia\"\u003eCYNTHIA\u003c/a\u003e, \u003ca href=\"/posts/products-concepts/bert\"\u003eBERT\u003c/a\u003e and \u003ca href=\"/posts/products-concepts/walter\"\u003eWALTER\u003c/a\u003e.\u003c/p\u003e\n\u003ch2 id=\"essays\"\u003eEssays\u003c/h2\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"/posts/essays/\"\u003eEssays\u003c/a\u003e\u003c/strong\u003e — long form pieces on technology, human systems and environmental supply chain risk, in parallel with \u003ca href=\"https://medium.com/@vrielink\"\u003eMedium\u003c/a\u003e.\u003c/p\u003e","title":"Write"},{"content":"Some leftover chores.\nI setup Calendly as my meeting scheduler, with Google Meet for web conferencing (I contemplated Jitsi but that does not integrate in Calendly as smoothly); I installed the Excalidraw plugin for Obsidian. It\u0026rsquo;s insane. Especially in combination with a tablet and a stylus pen. Madness; Hugo supports emojis short codes: \u0026#x1f98a; Hugo emojis quick reference \u0026#x1f989; ; Prompt engineering to skill engineering to context engineering to loop engineering to graph engineering. And I probably missed some. There\u0026rsquo;s a new MUST HAVE skill every month in the wonderful world of LLMs. Apparently LangChain and LangGraph are the tools to use. So I wachted the LangChain vs LangGraph video from the IBM series on AI. I like their AI videos. (spoiler: LangChain and LangGraph are both graphs. But LangChain is a directed and acyclic graph, a DAG). Update on Calendly:\nCombining Calendly with mixed fastmail and google account for mail and calendar sucks; it messed up communication (spoofing) and people got errors booking meetings. So I caved in and decided to go for a Google Workspace plan. One of my silent goals was (and is) to see if there are ways to bypass big-corp and / or paid options (and there always are). But in this case the trade-off to convenience was too big. I completely forgot, but Google Calendar has it\u0026rsquo;s own scheduling functionality making Calendly functionally obsolete. I will keep using Calendly for the time being (to get to know it and learn why everyone is using it). But ready to ditch it any time for native Google Calendar scheduling. Tmux Tmux is one of my favourite apps or (should I say \u0026ldquo;concepts\u0026rdquo;?). But it\u0026rsquo;s impossible to use. The default UX is terrible. The UI is also terrible. But since it\u0026rsquo;s open source, there\u0026rsquo;s a ton of plugins and configuration options that make it much easier and pleasant to use.\nMy setup is largely based on this tutorial:\nTook me about 20 minutes to watch and setup my tmux. Then I spent 3,5 hours debugging and tweaking the setup (since the video is 2 years old, some settings are out of date).\nWhen I was finished, I found this video, which is much simpler with (almost) the same result:\nBoth the DevOps Toolbox (he sounds like the YT bass dude) and Typecraft channels are cool and informative if you\u0026rsquo;re into Linux and DevOps. Mischa van den Burgs channel is also cool but more focused on home lab and distros.\nI stumbled upon the channel via his 150$ Laptop video:\nSuperLogic I use Ghostty as a terminal emulator. And Ghostty is built by Mitchel Hashimoto, founder of HashiCorp, creator of ao Vagrant (\u0026lt;- tell me if you used Vagrant and I\u0026rsquo;ll send you a sticker) and Terraform. He is now working on SuperLogical, apparently a sane Tmux alternative:\nQuick Superlogical demo to end the week. This demo is just of the basic functionality, but I want to highlight just how FAST it is. There\u0026#39;s a WHOLE lot more functionality already present and coming but I wanted to keep the demo short. But seriously, check out that speed.\nI\u0026#39;m… pic.twitter.com/teBkX0PCKq\n\u0026mdash; Mitchell Hashimoto (@mitchellh) August 28, 2026 Mitchel explaining why he\u0026rsquo;s building SuperLogical:\nI can\u0026#39;t tell you the number of people over my career that have said \u0026quot;I tried learning tmux\u0026quot; First problem: the fact you have to \u0026quot;learn it\u0026quot; at all. The world\u0026#39;s best multiplexer is the one you don\u0026#39;t have to learn. It just works. That\u0026#39;s what we\u0026#39;re building (and then some 😜).\n\u0026mdash; Mitchell Hashimoto (@mitchellh) August 28, 2026 And I love his dedication and commitment to detail, especially speed:\nI love how many people are attaching to the \u0026quot;half-bounch\u0026quot; startup time metric. I used to do paid macOS dev ~20 years ago and most macOS devs I knew then used the bounce metric! Back then multiple bounces was the norm due to HDDs. One bounce really wasn\u0026#39;t possible until SSDs.…\n\u0026mdash; Mitchell Hashimoto (@mitchellh) August 29, 2026 SuperLogical is the reason I\u0026rsquo;m holding out a bit on Herdr.\nTODO:\nLearn how to use Excalidraw properly ","permalink":"https://thishollow.earth/posts/tech-notes/tn-meetings-excalidraw/","summary":"\u003cp\u003eSome leftover chores.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eI setup \u003ca href=\"https://calendly.com/arjen-thishollowearth\"\u003eCalendly\u003c/a\u003e as my meeting scheduler, with Google Meet for web conferencing (I contemplated \u003ca href=\"https://jitsi.org\"\u003eJitsi\u003c/a\u003e but that does not integrate in Calendly as smoothly);\u003c/li\u003e\n\u003cli\u003eI installed the \u003ca href=\"https://community.obsidian.md/plugins/obsidian-excalidraw-plugin\"\u003eExcalidraw\u003c/a\u003e plugin for Obsidian. It\u0026rsquo;s insane. Especially in combination with a tablet and a stylus pen. Madness;\u003c/li\u003e\n\u003cli\u003eHugo supports emojis short codes: \u0026#x1f98a; \u003ca href=\"https://gohugo.io/quick-reference/emojis/#animals--nature\"\u003eHugo emojis quick reference\u003c/a\u003e \u0026#x1f989; ;\u003c/li\u003e\n\u003cli\u003ePrompt engineering to skill engineering to context engineering to loop engineering to graph engineering. And I probably missed some. There\u0026rsquo;s a new MUST HAVE skill every month in the wonderful world of LLMs. Apparently \u003ca href=\"https://www.langchain.com/langchain\"\u003eLangChain\u003c/a\u003e and \u003ca href=\"https://www.langchain.com/langgraph\"\u003eLangGraph\u003c/a\u003e are the tools to use. So I wachted the \u003ca href=\"https://www.youtube.com/watch?v=qAF1NjEVHhY\"\u003eLangChain vs LangGraph\u003c/a\u003e video from the IBM series on AI. I like their AI videos. (spoiler: LangChain and LangGraph are both graphs. But LangChain is a directed and acyclic graph, a DAG).\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cstrong\u003eUpdate on Calendly:\u003c/strong\u003e\u003c/p\u003e","title":"Calendly, Excalidraw, LangChain|Graph, SuperLogic"},{"content":"Bootstrapping This Hollow Earth.\nI want to know if the AI hype is real. To find out, I decided to start using AI in my day to day workflow. The idea is as follows:\nI am consulting about technology, product development, organisations and environmental supply chain risk. Great, but there are many people doing that and claiming to be good. How do you know if I\u0026rsquo;m for real or full of shit? By show, don\u0026rsquo;t tell. I build products and show how I do it. To build (design, engineer, ship, market, sell) the products, I will use AI. Using AI to build products allows me to learn and evaluate. I write about my adventures and publish them to exploit the teachers paradox. So, to get us started, we need 2 things:\nA publishing platform (this website) An AI workflow 1. Website update This site is built with Hugo, originally with the YinYang theme and hosted on Gitlab. The newsletter is powered by Buttondown. I use Obsidian for note taking and Vim.\nThe YinYang theme is nice and minimal and worked for my previous blog series but it\u0026rsquo;s a bit too limited. I looked around for other minimal options and found PaperMod. PaperMod is great. It\u0026rsquo;s simple but has a ton of options, it\u0026rsquo;s fast, it\u0026rsquo;s lean, it\u0026rsquo;s responsive. PaperMod is great.\nPapermod works out of the box but since I use Umami for analytics, I had to set that up by hand. I added layouts/partials/extend_head.html to my Hugo root and added the following snippet:\n{{- /* Head custom content area start */ -}} {{- /* Insert any custom code (web-analytics, resources, etc.) - it will appear in the \u0026lt;head\u0026gt;\u0026lt;/head\u0026gt; section of every page. */ -}} {{- /* Can be overwritten by partial with the same name in the global layouts. */ -}} {{- if site.Params.analytics.umami.SiteVerificationTag }} \u0026lt;script async src=\u0026#34;https://analytics.umami.is/script.js\u0026#34; data-website-id=\u0026#34;{{ site.Params.analytics.umami.SiteVerificationTag }}\u0026#34;\u0026gt;\u0026lt;/script\u0026gt; {{- end }} {{- /* Head custom content area end */ -}} Then I added the token to the hugo.yml configuration file, and it worked:\nanalytics: umami: SiteVerificationTag: \u0026#34;\u0026lt;token\u0026gt;\u0026#34; That way, the PaperMod theme folder doesn\u0026rsquo;t get polluted with custom code and you can run git submodule to keep it up to date.\nFor authoring, I switched to the LazyVim setup. LazyVim is a superfast, batteries included Neovim setup. DevOps Toolbox has a nice video about LazyVim (the channel is one of my sources for DevOps tooling btw; I find new cool tools all the time).\nTODO:\nreview text on syntax, grammar, tone, style, content review information architecture optimisation and tweaking visuals and branding enable multilingual and translate to Dutch migrate medium posts to essays AI workflow To setup a basic AI agent workflow I figured I need:\nA model A harness A memory A task manager A notification channel I kept some engineering principles in mind while selecting candidates: loose coupling, orthogonality, KISS and look at what you already have.\nThis README ticked all boxes. I mostly followed the recipe with the following notes:\nI use Hermes agent instead of OpenClaw since Hermes seems to most hyped one and the second mouse gets the cheese. I use Telegram as a notification channel. It\u0026rsquo;s cool to try new stuff and I never used Telegram, so I went for Telegram. I picked Linear as a task manager since I wanted to experience working with Linear. It\u0026rsquo;s hyped and I want to know why. And it has the coolest swag. I use Obsidian for memory. I reorganised my existing vault according to the PARA method. This setup adheres to all my principles: it\u0026rsquo;s model agnostic, I can switch and add notification channels, the tasks are managed independently and can be picked up by humans or agents, it allows for both HITL and AFK workflows.\nOne caveat: everything went smooth and worked pretty much out of the box following the README above. Until I got to the Connect Linear part. It says:\nThe easiest way to set this up: Ask your OpenClaw agent to do it. It already knows your machine, your setup, and your tools. Just send it this:\nHelp me set up a Linear integration for OpenClaw. I want you to: - Be assignable as a team member in Linear - Respond when I @mention you in comments - Update issue status (In Progress, In Review, Done) - Post comments on issues with work summaries Walk me through creating the OAuth app, setting up webhooks, and wiring everything together. Do as much of it as you can for me. Your agent will walk you through it step by step — and handle most of the heavy lifting itself. That\u0026rsquo;s the whole point of having an agent.\nInstead of OpenClaw, I used Hermes. And the agent did indeed walk me through the steps and set everything up. But it cost me my afternoon. Maybe because I was using a crappy model. But mostly (and this was actually fun) because getting my agent to post to Linear was easy, but @mentioning my agent in a Linear issue opened a can of worms. I run agent on my laptop. So there is no public endpoint or webhook for Linear to post to. Nevermind, Cloudflare fixes this with Cloudflared, a tunnel client. I made a Cloudflare account, installed the tunnel client and setup a tunnel in the cloudflare web UI. Easy. Got it working in no-timne. But a tunnel is a tunnel, you need a route as well, a from (public webhook) and a to (my laptop). Turns out, Cloudflare doesn\u0026rsquo;t supoprt .earth domains, so I couldn\u0026rsquo;t use thishollow.earth. Nevermind, So I registered thishollowearth.net for 10$ / year and used that as the public endpoint. Then my agent was smart enough to setup the to of the route as a new port in the Hermes gateway and Bob\u0026rsquo;s your uncle! It works. Well, almost. I can @mention my agent from Linear issues, it gets picked up by the agent. But when the @mention is not clear, the agent returns with clarification questions, which Linear doesn\u0026rsquo;t respond to and then simply times out the connection. O well, let\u0026rsquo;s figure that out later.\nUntil now, I ran Hermes with free OpenRouter models, which is not ideal. But it works and you learn a lot about managing your token spend. I am planning to run a Claude Pro subscription model for production eventually.\nEverything works. Now I have to start using it to learn and get familiar with the tooling.\nNext Hermes is perfect for day to day work. For hardcore engineering and software development (building products) I will to use Pi and Opencode.\n","permalink":"https://thishollow.earth/posts/tech-notes/tn-bootstrap/","summary":"\u003cp\u003eBootstrapping This Hollow Earth.\u003c/p\u003e\n\u003cp\u003eI want to know if the AI hype is real. To find out, I decided to start using AI in my day to day workflow. The idea is as follows:\u003c/p\u003e\n\u003col\u003e\n\u003cli\u003eI am consulting about technology, product development, organisations and environmental supply chain risk. Great, but there are many people doing that and claiming to be good. How do you know if I\u0026rsquo;m for real or full of shit?\u003c/li\u003e\n\u003cli\u003eBy \u003cem\u003eshow, don\u0026rsquo;t tell\u003c/em\u003e. I build products and show how I do it.\u003c/li\u003e\n\u003cli\u003eTo build (design, engineer, ship, market, sell) the products, I will use AI. Using AI to build products allows me to learn and evaluate.\u003c/li\u003e\n\u003cli\u003eI write about my adventures and publish them to exploit the teachers paradox.\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eSo, to get us started, we need 2 things:\u003c/p\u003e","title":"Bootstrapping This Hollow Earth"},{"content":"Launch time.\nTechnology, Human Systems, Environment. I soft launched the new version of thishollow.earth today. New theme, new content, new streams, same me.\nnew landing page familiar blog stream (this newsletter) new products \u0026amp; concepts stream new solutions page, outlining the 3 pillars (technology, human systems, environment), each with their own product show case CYNTHIA, BERT, WALTER new about page with profile, expertise,experience and how to get in touch new tech notes stream long form essay stream (TBA, in parallel from medium) scalpel stream (TBA) It\u0026rsquo;s far from finished. It\u0026rsquo;s full of errors and mistakes and omissions. But it\u0026rsquo;s something, it\u0026rsquo;s good enough and it\u0026rsquo;s live. This is part of the bootstrapping ThisHollowEarth. After bootstrapping, I will start building the products, setup a Go To Market (GTM) plan, focus on branding, visuals and design, do optimisation / tweaking / security / legal / finance and then rinse and repeat.\nYou can follow everything I do, build, decide in NEAR REAL TIME via the tech notes stream for tech, tools \u0026amp; opinions and this blog. It will be great fun.\nLet me know what you think. Any feedback is welcome.\nAn evening stroll\nI was walking my dog the other day when we ran into the fly\n\u0026ldquo;Good morning fly\u0026rdquo;, I said\n\u0026ldquo;Hmmm\u0026rdquo;, the fly said\n\u0026ldquo;Hmmm?\u0026rdquo;, I said, \u0026ldquo;Something wrong?\u0026rdquo;\n\u0026ldquo;I\u0026rsquo;m tired\u0026rdquo;, the fly said, \u0026ldquo;I flew far, fetched fumbling for fakes\u0026rdquo;\n\u0026ldquo;That doesn\u0026rsquo;t make sense fly\u0026rdquo;, I said\n\u0026ldquo;I\u0026rsquo;m tired\u0026rdquo;, the fly said\nCynicism\nI read the \u0026quot;MIT PRESS ON ESSENTIAL KNOWLEDGE SERIES\u0026quot; book on Cynicism which was very disappointing. I like other books in the series but this one is not good. Nevertheless, there was this one passage that I like a lot:\nEveryday existence, as its critics claim, is now relentlessly exploited by capital. Workers and educational subjects are expected to take upon themselves a stylization of life that will render them permanently adaptable to the whims of the marketplace. Individual lives are framed increasingly in this way, folded back on themselves, as lives of self-stylization vis-a-vis the cultivation of [potentially counter-]cultural taste, body language and so forth. When social problems are confronted, uncomfortable political questions are suspended \u0026ldquo;in the name of ethical immediacy.\u0026rdquo; For instance, since the first Live Aid concert in 1985, a specific form of \u0026ldquo;ideological blackmail\u0026rdquo; has taken root, insisting that \u0026ldquo;caring individuals\u0026rdquo; can end famine directly by collectively reaching in their pockets, without the need for systemic reorganization\nIn other words: virtue signalling. I hate virtue signalling.\nWhat\u0026rsquo;s next?\nNext blog will announce a new section: scalpel. I plan to publish the essay on Consultancy -\u0026gt; Product organisation maturity levels before September 15. ","permalink":"https://thishollow.earth/posts/test_33/","summary":"\u003cp\u003eLaunch time.\u003c/p\u003e\n\u003cp\u003eTechnology, Human Systems, Environment. I soft launched the new version of \u003ca href=\"thishollow.earth\"\u003ethishollow.earth\u003c/a\u003e today. New theme, new content, new streams, same me.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003enew \u003ca href=\"/\"\u003elanding page\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003efamiliar \u003ca href=\"/posts/\"\u003eblog stream\u003c/a\u003e (this newsletter)\u003c/li\u003e\n\u003cli\u003enew \u003ca href=\"/posts/products-concepts/\"\u003eproducts \u0026amp; concepts\u003c/a\u003e stream\u003c/li\u003e\n\u003cli\u003enew \u003ca href=\"/solutions\"\u003esolutions\u003c/a\u003e page, outlining the 3 pillars (technology, human systems, environment), each with their own product show case \u003ca href=\"/posts/products-concepts/cynthia\"\u003eCYNTHIA\u003c/a\u003e, \u003ca href=\"/posts/products-concepts/bert\"\u003eBERT\u003c/a\u003e, \u003ca href=\"/posts/products-concepts/walter\"\u003eWALTER\u003c/a\u003e\u003c/li\u003e\n\u003cli\u003enew \u003ca href=\"/about\"\u003eabout\u003c/a\u003e page with profile, expertise,experience and how to get in touch\u003c/li\u003e\n\u003cli\u003enew \u003ca href=\"/posts/tech-notes/\"\u003etech notes\u003c/a\u003e stream\u003c/li\u003e\n\u003cli\u003e\u003ca href=\"/posts/essays/\"\u003elong form essay stream\u003c/a\u003e (TBA, in parallel from \u003ca href=\"https://medium.com/@vrielink\"\u003emedium\u003c/a\u003e)\u003c/li\u003e\n\u003cli\u003escalpel stream (TBA)\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eIt\u0026rsquo;s far from finished. It\u0026rsquo;s full of errors and mistakes and omissions. But it\u0026rsquo;s something, it\u0026rsquo;s good enough and it\u0026rsquo;s live. This is part of the \u003ca href=\"/posts/tech-notes/tn-bootstrap/\"\u003ebootstrapping ThisHollowEarth\u003c/a\u003e. After bootstrapping, I will start building the products, setup a Go To Market (GTM) plan, focus on branding, visuals and design, do optimisation / tweaking / security / legal / finance and then rinse and repeat.\u003c/p\u003e","title":"Test 33"},{"content":"Bert\u0026rsquo;s Enterprise Readiness Templates\nShift Left Enterprise Readiness\n","permalink":"https://thishollow.earth/posts/products-concepts/bert/","summary":"\u003cp\u003eBert\u0026rsquo;s Enterprise Readiness Templates\u003c/p\u003e\n\u003cp\u003eShift Left Enterprise Readiness\u003c/p\u003e","title":"BERT - Shift Left Enterprise Readiness"},{"content":"From environmental landscape risk to product environmental passport, a pipeline.\n","permalink":"https://thishollow.earth/posts/products-concepts/cynthia/","summary":"\u003cp\u003eFrom environmental landscape risk to product environmental passport, a pipeline.\u003c/p\u003e","title":"CYNTHIA - Landscape 2 Product"},{"content":"A quarterly decision radar for navigating environmental risk in global supply chains\nEnvironmental risk in commodity supply chains is hard to understand. Regulations are emerging and changing. Climate and ecological conditions are shifting. New satellite, geospatial and AI capabilities are producing more data. Commodity supply chains increasingly intersect with questions of water, forests, biodiversity, land degradation and ecosystem resilience. At the same time, companies, investors and consultants are being asked to turn all of this complexity into practical decisions.\nYet much of the information available today falls into two categories: bespoke consulting reports that become outdated quickly, or generic risk scores and data products that are difficult to translate into action.\nWalter, The Environmental Supply Chain Risk Radar intends to fill that gap.\nIt is a recurring, evidence-linked, opinionated overview of what is changing in environmental supply-chain risk: what matters now, what is becoming important, what is mature enough to use, and what organisations should do next.\nThe model is inspired by the Thoughtworks Technology Radar: a simple visual language, a recurring publication cadence, and explicit recommendations rather than another passive database.\n","permalink":"https://thishollow.earth/posts/products-concepts/walter/","summary":"\u003cp\u003eA quarterly decision radar for navigating environmental risk in global supply chains\u003c/p\u003e\n\u003cp\u003eEnvironmental risk in commodity supply chains is hard to understand. Regulations are emerging and changing. Climate and ecological conditions are shifting. New satellite, geospatial and AI capabilities are producing more data. Commodity supply chains increasingly intersect with questions of water, forests, biodiversity, land degradation and ecosystem resilience. At the same time, companies, investors and consultants are being asked to turn all of this complexity into practical decisions.\u003c/p\u003e","title":"WALTER - Environmental Supply Chain Risk Radar"},{"content":"Conference time.\nTwo conferences, serendipity, long form essays, the future.\nI didn\u0026rsquo;t go to a lot of conferences the past few years. Let alone speak at conferences. But then I got invited to speak at 2 conferences last spring. Many good things came from both. I made a lot of notes but reviewing them I realised they don\u0026rsquo;t make much of a coherent story. But who needs coherence? Instead I present you observations, impressions and random thoughts.\nAI fun in Stanford\nHow often do you get invited to speak at an AI conference at one of the worlds top ranked universities, more specifically at their consistently ranked 1st business school?\nNever.\nUntil you do. My friend Dan Iancu kindly invited me to his conference: Sustainable and Responsible Operations in the Age of AI. So I said \u0026ldquo;yes\u0026rdquo;.\nImmediately, 2 problems emerged:\nIf I was going to present something, I needed something to present. And I had no clue what to present. I was slightly intimidated by the other speakers in my panel, coming from MIT, Stanford and University of Washington. I would surely be exposed as a fraud and ridiculed and mocked. Luckily, I got an idea while walking my dog the other day. Last year, I did a presentation to an elementary school class of 12 year olds about satellite monitoring. I decided to take that presentation as a basis; if it\u0026rsquo;s good enough for kids, it\u0026rsquo;s good enough for Stanford. Next, I used the conference title as presentation title and used each noun from the title as a section title. Voila, my presentation, hello success.\nAll went well, I survived the panel discussion and made my way back to my seat. But whereas my little walk up to stage before the panel was a 10 second formality, the way back felt like an Odyssey where I had to overcome challenges, riddles and temptations. For example:\nJames Leape, WWF endboss, comes up and starts lecturing me on how it is unethical to work with companies like Cargill, Pepsico or Nestle. I get the reflex, but I don\u0026rsquo;t believe Dutch uncling is helping the world move forward. It\u0026rsquo;s great to have an activist stance to raise awareness, put topics on the agenda. But we should realise a new generation is working for those companies now. An motivated, aware and engaged generation who will be the leaders of those companies in 10 or 20 years. I think looking for ways to collaborate yields much more for the planet than only raising your finger. A Chinese lady came up and stared at me for a long time. Then she said \u0026ldquo;Zhang Ziyi is my family\u0026rdquo;. I said \u0026ldquo;great\u0026rdquo;. I never heard of Zhang Ziyi. She said \u0026ldquo;I never saw a presentation with stills from a Chinese director\u0026rdquo;. We talked about Chinese cinema, family and startups. Later, I found Zhang Ziyi played in 2046. Thank you serendipity. appreciation\nObservations\nWhen the average is the noise (eg crowd sourcing ): when a bipolar distribution is polluted by noise and shows up as a Gaussian distribution, the average is actually the worst answer.\nWalking around the campus with I found everything is named and sponsored. I get the philantropy but why the naming? Be humble.\nThe breadth of the conference was really nice, with people from academia and business presenting about AI. There seemed to be a general optimistic, positivist attitude towards AI. Later I thought that must be partially because most people were from a business / economics background where there are many problems that are perfectly suited for AI based solutions (like optimisation problems)\nParadoxes\nHow some of the smartest minds are so easily impressed (by AI hype and marketing); where is your critical mind, where is your skeptical perspective? Challenge me!\nPeople asked me what I thought about the conference and the presenters and the attendees. I said the conference might as well be subtitled \u0026ldquo;impostor syndrome gathering\u0026rdquo;. It was basically really smart people trying to convince other really smart people that they are really not that smart.\nMe\nI really like presenting (at conferences, offices, funerals) I really like being around thoughtful, curious, open-minded people who are willing to listen seriously without taking themselves too seriously. People that are comfortable with their intelligence. Serendipity Epilogue\nWalking back to the car with Dan, we came by his office in the Knight Management Center.\nMe: \u0026ldquo;Knight. Michael Knight?\u0026rdquo;\nDan: \u0026ldquo;Yeah, we have his shoes here.\u0026rdquo;\n\u0026ldquo;\u0026hellip;\u0026rdquo;\nchair\nAn evening stroll\nI was walking my dog the other day when we ran into the cat\n\u0026ldquo;Good morning cat\u0026rdquo;, I said\n\u0026ldquo;Good morning\u0026rdquo;, the cat said\n\u0026ldquo;Do you like chicken?\u0026rdquo;, I asked\n\u0026ldquo;I like chicken\u0026rdquo;, the cat replied\n\u0026ldquo;Do you like salmon?\u0026rdquo;, I continued\n\u0026ldquo;I do like salmon\u0026rdquo;, the cat answered\n\u0026ldquo;Do you like mice?\u0026rdquo;\n\u0026ldquo;I like mice a lot!\u0026rdquo;, the cat exclaimed\n\u0026ldquo;Then, why don\u0026rsquo;t they sell cat food with mice?\u0026rdquo;\nWe left the cat puzzled and hungry\nOpenGeospatialConsortium fun in Helsinki The night I arrived in Finland, the whole city (and country) was celebrating their jr. World Cup title in ice hockey. Those were not the quiet and calm Fins I remembered from the few months I lived in Helsinki, about 17 years ago.\nI came to Helsinki to talk about the operational aspects of implementing TNFD at the OpenGeospatialConsortium technical meeting.\nI always thought that these meetings would be too techie, too nerdy. And they are. But since Peter Rably leads the executive team as a CEO, OGC is much more business and commercially oriented. Which is a good thing. The OGC meetings also have an applied track now for example, making it interesting for a much wider audience. I enjoyed the conference a lot.\nWhen I boarded my plane in Amsterdam, I found the dude in the seat next to me working on a GeoBIM presentation. I casually inquired if he was attending the conference. He was.\nI picked the conference hotel, assuming the venue would be in the hotel or else, close to the hotel. It wasn\u0026rsquo;t. The venue was a 30mins car ride away. Fine, there\u0026rsquo;s Uber. While waiting for my Uber, I saw a nervous Spanish guy in the lobby asking questions to the hotel staff. I guessed he had the same problem as me: venue is far away. I offered to share the Uber. Turns out, the guy was a modeler (\u0026ldquo;I model the world\u0026rdquo;), had some bad luck with storms over Europe and only arrived at the hotel at 3:00 in the morning. In his attempt to get a quick shower and some sleep, he accidentally pulled the emergency cord in the hotel bathroom setting off some alarm sound denying him sleep.\nWhile he was explaining all this to me, a big Jordanian living in San Diego happy dude stepped out of the elevator and greeted the Spaniard in recognition. He joined the Uber as well. The next time I saw him was the next nigh at 12:30 in the hotel lobby when he rushed in urging everyone to visit \u0026ldquo;the island\u0026rdquo;, only 30 minutes away by ferry. He was contemplating going to Lofoten since he was in the Nordics anyway (\u0026ldquo;I visited The Philippines and Japan when I was in Korea\u0026rdquo;). The man really loves islands.\nDuring the taxi ride to the venue, I started to recognise streets and buildings from my time in Finland. I asked the driver if the address was at the research campus. He said yes. The research campus is where my office was located. I started to remember people and names from that time, especially Kristian, a Norwegian I spent many lynch breaks with. I searched my old emails and found our latest exchange from 17 years before. I wrote him a short message to the old address. He replied within 10 minutes. 1 hour later we were having coffee in our old office. 17 years had past but he was still exactly the same person. My faith in life was restored just a little bit.\nThe conference venue was in \u0026ldquo;the wooden building\u0026rdquo;, just across the street of my old research building. A chain of serendipity. Thank you, life.\nCoffee is really strong in Finland. I was reminded. Really strong. Good.\ncoffee, a beverage liked by Fins\nDigital World Models\nThe most popular topic was digital twins or digital world models. The confusing thing is there are a lot of digital flavours: digital twin. bim (the professor from Wales next to me in the airplane), OGC, world models (the Spanish guy I shared an Uber with) or, supply chain models which, in the end, all point to the same phsyical reality. Some of these models only recently discovered there is a (lacking) geospatial component, so they wrench it in. At some point, there will be convergence: but will there real convergence or simply connectors?\nkukko\nBuilding Blocks\nVery interesting: OGC Building blocks: if you can break the problem down to human meaningful components, tooling can build it back up again; define and advocate principles and the complexity will evolve from that.\nGeospatial AI, where it hurts\nJevons paradox, AI is the new excel It\u0026rsquo;s an amplifier, more noise Hallucinations Probabilistic nature Semantics The ontology problem in geospatial is orders of magnitudes harder than for organisations Things I learned\nRacing Santander made it back to the Primera Division Have probabilistic systems generate deterministic systems (LLM produces code, code is deterministic) From file to service. From central to distributed. From binary to granular. From monolith to modular. Familiar patterns. AI will favour people and organisations who have their shit in order trust: no one buys your product if they don\u0026rsquo;t trust it Open data can’t be open by default: we need curated exposure. Else we are just adding noise, which doesn\u0026rsquo;t help in building trust Earth Observation should be designed to be compatible, not retrofitted The amount of data is irrelevant. Start with the problem since we already have a graveyard of solutions Trust as multimodal grounding A Digital Twin is only a twin if it interacts with the real world. Else it’s a digital shadow. Me\nI don\u0026rsquo;t like group pictures, I\u0026rsquo;m allergic to groups applauding themselves I want to be a conference anthropologist No one talks about the hard and boring work Commercial confs, nerd confs\nI started this post with the observation that it had been while for me to go to a conference. That\u0026rsquo;s not exactly true. I did go to some commercial conferences. Which I liked, since I like almost anything new. But attending the Stanford AI nerd gathering and the OGC geospatial nerd meeting, the difference in atmosphere struck me.\nAt commercial conferences, it\u0026rsquo;s really easy to connect to people. Most people are outgoing, open and friendly. But all talk remains superficial blabla. There is no real substance.\nAt tech conferences, the connections happen rather serendipitiously. But the moment you talk to someone, there is IMMEDIATE, crazy depth. No small talk, no uneasy introductions, no cautious looking for common ground. Just sincere interest and discussions about niche topics.\nRandom person: \u0026ldquo;I don’t know you but I have to talk to you, you are interesting\u0026rdquo;\nMe: \u0026ldquo;\u0026hellip; sure!\u0026rdquo;\nThe commercial confs feel like real life LinkedIn. The tech confs are where people leave their costume at home. I enjoy both. I\u0026rsquo;m the same person at both. The interaction and atmosphere are just different.\nGTM, Engineering, Product and AI\nThank you people who forwarded their comments and requests and suggestions. GTM and its discontents is the most popular topic.I will write about it. But I will first write the conceptual essay on Consultancy 2 SaaS maturity levels. That gives us a common framework and mental model which we can apply then apply to the GTM and second brain use cases.\nInnovation is fast, adoption is slow.\nSneak Peak\nEnvironmental Supply Chain Risk Radar Shift Left Enterprise Landscape \u0026raquo; Supply Chain \u0026raquo; Product \u0026ldquo;When is the last time your dashboard made a decision?\u0026rdquo;\n?\nWhatever man.\nNote to self: please review this post\n","permalink":"https://thishollow.earth/posts/test_32/","summary":"\u003cp\u003eConference time.\u003c/p\u003e\n\u003cp\u003eTwo conferences, serendipity, long form essays, the future.\u003c/p\u003e\n\u003cp\u003eI didn\u0026rsquo;t go to a lot of conferences the past few years. Let alone speak at conferences. But then I got invited to speak at 2 conferences last spring. Many good things came from both. I made a lot of notes but reviewing them I realised they don\u0026rsquo;t make much of a coherent story. But who needs coherence? Instead I present you observations, impressions and random thoughts.\u003c/p\u003e","title":"Test 32"},{"content":"I am Arjen Vrielink.\nI write, build and consult at the intersection of geospatial technology, organisation \u0026amp; communication and environmental landscape risk.\n\u0026ldquo;Arjen has the ability to bring structure to chaos and chaos to structure; he brings people together.\u0026rdquo;\nMy talent is bridging the tech and commercial sides of organisations. I can quickly grasp complex issues and translate them into simple, technically feasible, and commercially viable terms.\nSkills, expertise and experience in my CV\nWork with me Book a meeting, send me an email at arjen@thishollow.earth or find me on Twitter to get in touch. I\u0026rsquo;m happy to discuss ideas, suggestions, areas of collaboration and life.\nI am also happy to collaborate with other freelancers!\n","permalink":"https://thishollow.earth/about/","summary":"\u003cp\u003eI am Arjen Vrielink.\u003c/p\u003e\n\u003cp\u003eI \u003ca href=\"/write/\"\u003ewrite\u003c/a\u003e, \u003ca href=\"/posts/products-concepts/\"\u003ebuild\u003c/a\u003e and \u003ca href=\"/consult\"\u003econsult\u003c/a\u003e at the intersection of geospatial technology, organisation \u0026amp; communication and environmental landscape risk.\u003c/p\u003e\n\u003cblockquote\u003e\n\u003cp\u003e\u003cem\u003e\u0026ldquo;Arjen has the ability to bring structure to chaos and chaos to structure; he brings people together.\u0026rdquo;\u003c/em\u003e\u003c/p\u003e\n\u003c/blockquote\u003e\n\u003cp\u003eMy talent is bridging the tech and commercial sides of organisations. I can quickly grasp complex issues and translate them into simple, technically feasible, and commercially viable terms.\u003c/p\u003e\n\u003cp\u003eSkills, expertise and experience in my \u003ca href=\"/CV_Arjen_Vrielink.pdf\"\u003eCV\u003c/a\u003e\u003c/p\u003e","title":"About"},{"content":"You read the books and blogs, you listen to the gurus podcasts, you followed the instructions. But it doesn\u0026rsquo;t work quite like the books and the blogs and instructions say. Because the tech you need doesn\u0026rsquo;t exist yet, companies cannot be run from paper or a slide deck and the environment doesn\u0026rsquo;t follow accounting rules.\nBased on more than 25 years of failures, misunderstandings and disappointments I know your case is different from the books and blogs and podcasts. But not that different.\nI advise at the intersection of Technology, Human Systems and Environment.\nI develop products and concepts to show what I do. You can follow the development in near real time via my tech notes, blog and essays.\nTechnology Technology is only useful when it changes what an organisation can do. I work across geospatial technology, data and product development. From making new concepts tangible to turning complex technology into products people actually use.\n\u0026ldquo;Your prospects don\u0026rsquo;t give a shit about your technology, they want their problem solved\u0026rdquo;\nCYNTHIA is my brainstorming space to map and model environmental landscape risk (growing \u0026amp; sourcing) to commodity supply chains (trading. manufacturing) to products (retail). Producer to consumer, farm to table.\nHuman Systems Organisations are human systems, messy systems. I work on the structures, culture, communication and ways of working that allow organisations to move from startup to scaleup, from consultancy to product, and from good intentions to effective execution.\n\u0026ldquo;That\u0026rsquo;s what I said, but that\u0026rsquo;s not what I meant\u0026rdquo;\nBERT is shift left enterprise readiness. You are a small company, a startup, 3 founders in a garage and suddenly you hit bullseye.\nEnvironment Environmental risk doesn\u0026rsquo;t care about organisational boundaries. I work on how geospatial technology can help translate environmental landscape risk to commodity supply chains, connecting what happens in the landscape to the companies, products and decisions downstream.\n\u0026ldquo;When is the last time you saw a dashboard make a decision?\u0026rdquo;\nWALTER is a Environmental Supply Chain Risk Radar, in the spirit of Thoughtworks Tech Radar.\nBusiness Type Startup You have a product, some customers and perhaps some funding. Now you need to turn a good idea into an organisation that can actually deliver.\nI can help with the uncomfortable questions: what to build, who to hire, how to organise the team, what to automate, what not to build and when you need to stop behaving like three people in a garage.\nScaleup The processes that worked at 8 people break at 20. The architecture that worked at 10 customers breaks at 100. The founders who did everything themselves become the bottleneck.\nI help scale the technology, product and organisation without losing the things that made the company work in the first place.\nConsultancy You sell expertise. But the whizzkid discovered github, Python and AI and built a tool. Your fellow consultants discovered the tool and are using it in their project proposals. The tool creates happy clients.\nWhat to do? Invest in the tool? Hire more engineers? How to manage development and maintenance cost? What\u0026rsquo;s the return on investment?\nI help consultancy businesses make the transition from project → platform → product: identifying what can be standardised, what should remain bespoke, and how software, data and repeatable processes can lower cost and increase margins.\nEnterprise You are a sustainability manager, an innovation manager or responsible for reporting on environmental KPIs at the financial department. You need reliable data for reporting, decision making and planning. But the data isn\u0026rsquo;t there, the data is unclear, on a consultant\u0026rsquo;s laptop or on again another platform or screen from one of your service provider\nWhat to do?\nI can help with your data and supplier strategy in the environmental commodity supply chain risk field.\nInvestors Considering a stake, an acquisition or a partnership in a company in this space? I can help you look past the pitch deck.\nTech due diligence Security scan Certification readiness Organisation review Culture assessment Bullshit assessment M\u0026amp;A: horizontal \u0026amp; vertical consolidation strategy for ESG service providers Business Need Growing Growth is fun until suddenly everything becomes someone else\u0026rsquo;s problem.\nWho should we hire next: an engineer, salesperson, marketeer, product manager or consultant? When does a startup need management? When do you need HR, finance, legal or security expertise? Which processes should be introduced — and which should be left alone? How do you keep communication and culture from disappearing as the organisation grows? I\u0026rsquo;ve been there. At Satelligence, for example, I helped grow the organisation from 6 to 50+ people while leading technology, product, operations and cross-functional teams.\nConsultancy → product Your consultants keep solving the same problem for different customers. Or your tech colleagues built cool tools and scripts. Or your flagship product is a sophisticated carbon emissions model. But it lives in Excel.\nMaybe it\u0026rsquo;s time to start thinking like a product organisation.\nI\u0026rsquo;ll help you work out what the product actually is, what should remain consultancy, how to structure the team around it, and how to move from selling hours towards building something repeatable.\nESG strategy In-house, platform, service provider, or open data layers?\n\u0026ldquo;We\u0026rsquo;ll build it ourselves, how hard can it be?\u0026rdquo;\nEveryone selling you one of those four has an interest in the answer. The real trade-offs are cost, control, maintenance burden, lock-in, and whether the thing survives the person who built it walking out.\nI help you decide what to build, what to buy, what to outsource and what is already free, and how to stitch it into something consistent and coherent you can actually report on.\nM\u0026amp;A | Consolidation The ESG market is full of interesting companies that do one part of the puzzle.\nData, satellite imagery, (Process Based) Modeling, Risk Analysis,. Supply-chain intelligence, consultancies, aggregators, \u0026ldquo;user friendly platforms with actionable insights\u0026rdquo;.\nThere is a lot of niche, a lot of overlap, a lot of confusion, a lot of waste but mostly: a lot of opportunity.\nWhat would happen when you put them together?\nAt some point, vertical and horizontal consolidation will happen.\nI can help investors and management teams assess technology, product, organisation and strategic fit.\nWork with me How that usually looks:\nSparring session a single conversation. You talk, I ask uncomfortable questions; Diagnostic two to five days looking at the technology, the organisation or both, ending in a written answer or presentation; Interim role one to three days a week for a few months. C-level, advisor or product lead while you work out who to actually hire; Advisory retainer a standing weekly or monthly slot. Someone to call before the decision, not after; Due diligence fixed scope, fixed fee, for investors and management teams. Not sure which one fits? That is what the first conversation is for.\nInterested? Book a meeting to explore areas of collaboration or send me an email at arjen@thishollow.earth\nAbout me\n","permalink":"https://thishollow.earth/consult/","summary":"\u003cp\u003eYou read the books and blogs, you listen to the gurus podcasts, you followed the instructions. But it doesn\u0026rsquo;t work quite like the books and the blogs and instructions say. Because the tech \u003cem\u003eyou\u003c/em\u003e need doesn\u0026rsquo;t exist yet, companies cannot be run from paper or a slide deck and the environment doesn\u0026rsquo;t follow accounting rules.\u003c/p\u003e\n\u003cp\u003eBased on more than 25 years of failures, misunderstandings and disappointments I know your case is different from the books and blogs and podcasts. But not \u003cem\u003ethat\u003c/em\u003e different.\u003c/p\u003e","title":"Consult"},{"content":"You\u0026rsquo;re an independent freelancer and happy and confident and successful and life is great!\nBut, but, but.\nSometimes, your project is too big to handle alone or needs additional skills and expertise. Sometimes, you want to build something with other people. And sometimes, you just want to have a informal chat, a coffee, to brainstorm ideas, get feedback on your latest pitch deck or simply let off steam.\nIf you are a freelancer in the ESG, organisation \u0026amp; communication or technology field, I am happy to collaborate in one or more of these fields:\nCreative: a joint build or write or other creative project; it\u0026rsquo;s more fun to build together. We show the world what we are capable of. Our idiot ideas die before they ship to production.\nCommercial: if your project needs more capacity or additional skills, we can complement each other. Increase our network. Joint acquisition if project demands. Share prospects and opportunities.\nReview: you need someone to review your article, give feedback on your MVP, the communication plan you drafted for your client. Or you want to brainstorm ideas, explore opportunities, do a reality check. Or you simply want to have a coffee.\nPlease note, all collaboration is lightweight and low-key. Nothing binding, no rules, no stress.\nMail me at arjen@thishollow.earth or book a meeting and let\u0026rsquo;s get to know each other.\nWe are a loosely organised collective of happy people doing things they like with other happy people doing things they like.\n","permalink":"https://thishollow.earth/freelancers/","summary":"\u003cp\u003eYou\u0026rsquo;re an independent freelancer and happy and confident and successful and life is great!\u003c/p\u003e\n\u003cp\u003eBut, but, but.\u003c/p\u003e\n\u003cp\u003eSometimes, your project is too big to handle alone or needs additional skills and expertise. Sometimes, you want to build something with other people. And sometimes, you just want to have a informal chat, a coffee, to brainstorm ideas, get feedback on your latest pitch deck or simply let off steam.\u003c/p\u003e\n\u003cp\u003eIf you are a freelancer in the ESG, organisation \u0026amp; communication or technology field, I am happy to collaborate in one or more of these fields:\u003c/p\u003e","title":"Freelancers"},{"content":"I have notes, so many.\nI have raw notes. I have useless notes. I have notes about notes.\nI have too many notes. And there is no time.\nBut there is time!\nThe years go fast, the hours go slow.\nThere is time.\nWe\u0026rsquo;ve got snackbars, AI, books, Den Helder, more AI, essays, and a stroll in the evening.\nSnackbar Maandag\nWhen walking my dog in the evening I regularly pass by Snackbar Maandag (Snackbar Monday).\nSnackbar Maandag\nI always found the name amusing. What an odd name for a snackbar. Or any restaurant really.\nUntil I felt like a snack on a Monday. Found my neighbourhood snackbars all closed. Searched the web and found an open one: Snackbar Maandag!\nHow wonderful. Idiot me. I should have been smarter. This is true SEO.\nHighway Karaoke\nSaves the Day and Face to Face and H2O and Madball are dominating the car playlist on a Tuesday morning drive to nowhere.\nMy 10 yo daughter: can we play Type O Negative next?\nSure.\nLOVING YOU IS LIKE LOVING THE DEATH\n❌ Bruno Mars\n✅ Peter Steele\nAn evening stroll\nI was walking my dog the other day when we ran into the pheasant\n\u0026ldquo;Good morning pheasant\u0026rdquo;, I said\n\u0026ldquo;Good morning\u0026rdquo;, the pheasant said\n\u0026ldquo;Where are you going?\u0026rdquo;, I said\n\u0026ldquo;Taxidermist\u0026rdquo;, the pheasant said\n\u0026ldquo;Taxidermist\u0026rdquo;, I said, \u0026ldquo;Short on cash?\u0026rdquo;\n\u0026ldquo;Short on cash? Short on cash? HA!\u0026rdquo;, the pheasant replied\n\u0026ldquo;Skin in the game bro\u0026rdquo;, the pheasant sneered, \u0026ldquo;Skin in the game. You wouldn\u0026rsquo;t know.\u0026rdquo;\nMy dog looked up.\nHe was nervous.\nDen Helder\nGTM, Engineering, Product and AI\nI have so many ideas and notes, I decided to write some long(er) form essays on Medium (it\u0026rsquo;s been 3 years!). To structure my thoughts and to get rid of bad ideas.\nPreliminary topics:\nConsultancy to SaaS maturity levels. Every startup aspires to be a SaaS (as in: scale, scale, scale; exponential growth; zero marginal cost). Almost every startup gets stuck into some hybrid form of Consultancy - SaaS. Some consciously, most unconsciously. Some on purpose, most not on purpose. Most become aware but have no clue how to get to the next maturity level. I will analyse and model Consultancy - SaaS maturity phases, how companies get stuck and how to get unstuck.\nSecond brain! One of the AI hypes. How to use AI as your second brain. Or your companies second brain. Well, it won\u0026rsquo;t. There is no second brain because the brain doesn\u0026rsquo;t work like that. It will fail for the same reason companies struggle with \u0026ldquo;knowledge bases\u0026rdquo; or \u0026ldquo;knowledge management\u0026rdquo;. I will analyse and explain how \u0026ldquo;second brains\u0026rdquo; and \u0026ldquo;knowledge bases\u0026rdquo; are not persistence (files on disk, wikis, databases) but behaviour rooted in principles.\nGTM, the misunderstandings, the misaligment or the pure lack. Your company fails (read: gets stuck) because of your lack of a Go To Market strategy. I will analyse and explain why GTM is the core reason your company isn\u0026rsquo;t functioning as you think it could. There are 2 basic failure modes:\nLack of GTM: there are a ton of company\u0026rsquo;s that simply don\u0026rsquo;t have a GTM. Misaligned GTM: conmpanies that do have a GTM but one that is (completely) misaligned with Product and / or Engineering. This essay boils down to: every organisation problem is a communication problem.\nIt will be great, it will be fun.\nLet me know if you have a preference for one of the topics, suggestions or requests.\nBooks\nWhat greater joy than compiling a long list for your holiday reading?\nlonglist\nWhat greater pain than culling the longlist to a short and final list?\nnote: so, yes, that means I bring analogue, physical books on holiday, the bulk of my luggage (in weight). Because:\nI should be able to smell the book. I should be able to paper cut myself. Reading should be dangerous. Else, it\u0026rsquo;s not a book, it\u0026rsquo;s not reading\nThe Singularity\nMy X timeline is dominated lately by AI. AI pilled zealots and AI skeptics and cynics. I do have my own opinion about AI based on what I read and my past experience with similar tech hype waves but I thought it would be nicer and more nuanced if I\u0026rsquo;d write about AI from my own experience. So I decided to start using and building AI and write about it. Exploit the teachers paradox.\nYou will find snippets in this newsletter. I am using AI to build a company where AI does the product management, GTM, CFO \u0026amp; accounting, engineering and still it won\u0026rsquo;t feel like AI slop. Not to build a company but to learn about AI and how to use it and how n00bs would use it and experience it. And I will use AI as a builder, as an engineer, as a software developer. Not to develop software, but to learn how and if I, and other engineeers, can be more productive.\nOf two things I am already convinced:\nThere is a gap, a huge gap, between the AI users and builders. See also my comments above about GTM, Product and Engineering. Evals will determine your (you, your company, the sector) success in using / building AI. And evals are difficult. As a start, I installed and played around with Hermes. The hype is silly, and the tutorials are hard to digest. But to be honest, it\u0026rsquo;s fun! It really is. Apart from that, I noticed you will have a ridiculous advantage if you know and worked with Linux, CLI, vi, tmux, zsh and the likes. Every tool, every concept, every principle, every foundation boils down to simple Linux core philosophies and design principles.\n:wqa!\nhappiness is easy\nInnovation is fast, adoption is slow.\nI have many people asking / telling me the reply to this newsletter doesn\u0026rsquo;t work. Sorry about that. I haven\u0026rsquo;t figured out how to fix that in buttondown yet. If you have questions or comments, drop me a line at arjen@thishollow.earth. I will reply.\n(and of course this one is in, it\u0026rsquo;s great, it\u0026rsquo;s about the love for food, each chapter is named after a Taiwanese dish.)\nREAD IT!!\n","permalink":"https://thishollow.earth/posts/test_31/","summary":"\u003cp\u003eI have notes, so many.\u003c/p\u003e\n\u003cp\u003eI have raw notes. I have useless notes. I have notes about notes.\u003c/p\u003e\n\u003cp\u003eI have too many notes. And there is no time.\u003c/p\u003e\n\u003cp\u003eBut there is time!\u003c/p\u003e\n\u003cp\u003eThe years go fast, the hours go slow.\u003c/p\u003e\n\u003cp\u003eThere is time.\u003c/p\u003e\n\u003cp\u003eWe\u0026rsquo;ve got snackbars, AI, books, Den Helder, more AI, essays, and a stroll in the evening.\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003e\u003cstrong\u003eSnackbar Maandag\u003c/strong\u003e\u003cbr\u003e\nWhen walking my dog in the evening I regularly pass by \u003cem\u003eSnackbar Maandag\u003c/em\u003e (Snackbar Monday).\u003c/p\u003e","title":"Test 31"},{"content":"Nothing is important\nstahlhartes Gehäuse\nLast time I mentioned Webers [https://en.wikipedia.org/wiki/Iron_cage](Iron Cage) and I do have to apologise.\nSorry.\nI should have said stahlhartes Gehäuse of course. It\u0026rsquo;s a much more accurate and telling description. Because it\u0026rsquo;s in German. German is one of my favourite languages. Since German has poetry baked into the language. Every German is a poet.\nWir haben uns lange, lange nicht geseh\u0026rsquo;n\nund es gibt so viel, so viel zu erzähl\u0026rsquo;n\nAber wir schweigen, wir schweigen fast schon laut\nDas nicht gesagte beschreibt alles ganz genau\nIt\u0026rsquo;s not a coincidence, I referenced Kafka and Marx as well. They wrote in German.\nIrgendwann ist es zu spät\nUm zu früh draufzugeh\u0026rsquo;n\nIrgendwann bin ich zu alt\nZu alt, um jung zu sterben\nDoch ich hab\u0026rsquo;s versucht\nIch hab\u0026rsquo; es wirklich versucht\nWirklich versucht\nGlaub mir, ich, ich hab\u0026rsquo;s versucht\nDoch vielleicht nicht gut genug\nThe sad thing is I never brought up the discipline to properly learn German and be comfortable reading German texts or having conversations. Maybe I should. But then again, there\u0026rsquo;s maybes everywhere and there\u0026rsquo;sa maybes all the time and maybes about everything and nothing. Maybe I shouldn\u0026rsquo;t.\nDie Hoffnung stirbt zuletzt\nAn evening stroll\nI was walking my dog the other day when we ran into the coot\n\u0026ldquo;Good morning coot\u0026rdquo;, I said\n\u0026ldquo;Good morning\u0026rdquo;, the coot said\n\u0026ldquo;You are not a fish, though you can swim\u0026rdquo;, I said\nThe coot stared at me\n\u0026ldquo;You are not a plane, though you can fly\u0026rdquo;, I continued\nThe coot lit a cigarette\n\u0026ldquo;You are not a human, though you can walk\u0026rdquo;, I tried\nThe coot sighed\n\u0026ldquo;You are not, \u0026hellip;\u0026rdquo;\n\u0026ldquo;Dude, would you just take it easy? I\u0026rsquo;m a bird, ok? Relax\u0026rdquo;\n\u0026ldquo;A bird\u0026rdquo;, I thought, \u0026ldquo;a bird\u0026rdquo;\nSmiles \u0026amp; Joy \u0026amp; Laughter\nI work(ed) with so many nice people. They all score very high on the \u0026ldquo;#-of-smiles-generated\u0026rdquo; KPI. The KPI that is not in job ads, not in function profiles, not in competence matrices. But it\u0026rsquo;s there and it\u0026rsquo;s important. It\u0026rsquo;s the glue and lubricant of your culture.\n___\nSomeone finally did it, the translation engine we have all been waiting for.\nTurns out, scientists found out you\u0026rsquo;re in shits creek if you are impressed by this sort of language:\n\u0026ldquo;To test the impact of corporate bullshit on workers, Littrell developed a “corporate bullshit generator” that generates statements such as “we will actualize a renewed level of cradle-to-grave credentialing”, creating “a hyper-connected, frictionless, and impact-minded global enterprise” all while “getting our friends in the tent with our best practices, we will pressure-test a renewed level of adaptive coherence”.\u0026rdquo;\n","permalink":"https://thishollow.earth/posts/test_30/","summary":"\u003cp\u003eNothing is important\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003estahlhartes Gehäuse\u003c/strong\u003e\u003cbr\u003e\nLast time I mentioned Webers [https://en.wikipedia.org/wiki/Iron_cage](Iron Cage) and I do have to apologise.\u003c/p\u003e\n\u003cp\u003eSorry.\u003c/p\u003e\n\u003cp\u003eI should have said \u003cem\u003estahlhartes Gehäuse\u003c/em\u003e of course. It\u0026rsquo;s a much more accurate and telling description. Because it\u0026rsquo;s in German. German is one of my favourite languages. Since German has poetry baked into the language. Every German is a poet.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eWir haben uns lange, lange nicht geseh\u0026rsquo;n\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003eund es gibt so viel, so viel zu erzähl\u0026rsquo;n\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003eAber wir schweigen, wir schweigen fast schon laut\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003eDas nicht gesagte beschreibt alles ganz genau\u003c/em\u003e\u003c/p\u003e","title":"Test 30"},{"content":"Today, I wrote nothing.\n¯(°_o)/¯\nThe asymetric nature of change:\nHumans change slow The machine changes fast \u0026ldquo;The machine\u0026rdquo; as systems of power (politics, technology, institutions, finance). They are self-reinforcing, self-interested, and largely autonomous from the individuals inside them. A banker, a politician, a programmer; each feels like they are making choices. But the system they inhabit has its own logic, its own momentum. The machine moves, whether or not any individual wills it.\nThis is what Kafka understood. What Marx was pointing at with capital. What Weber meant by the iron cage. The machine isn\u0026rsquo;t a conspiracy, it\u0026rsquo;s an emergent property of coordination at scale.\nHumans are structurally slow against all of these. Humans update through experience, grief, conversation, and storytelling. These are slow, lossy, and local processes. Systems of power update through incentives, feedback loops, and compounding, which are fast, efficient, and global. The information asymmetry alone is disabling.\nIf the machine is any fast-moving system of power, then the asymmetry isn\u0026rsquo;t a modern problem, it\u0026rsquo;s a permanent feature of civilization. Every era has its machine. What changes is the speed differential. And right now multiple machines (financial, technological, political) are accelerating simultaneously. That convergence is what makes this moment novel. Not that the asymmetry exists, but that there\u0026rsquo;s nowhere left where it doesn\u0026rsquo;t.\nThe machine produces, but without meaning. It has no interior. It cannot grieve, hesitate, or doubt. It\u0026rsquo;s hollow, empty. Unintentional.\nTherefore, we need story tellers, we need writers, we need poets. Not to describe the gap, not to close the gap, not to fight the asymmetry. But to widen the gap, to deepen the gap, to create space for meaning. Stare into the abyss, preserve interiority, ambiguity, and slowness. Don\u0026rsquo;t escape it. Occupy the territory the machine cannot enter.\nAn evening stroll\nI was walking my dog the other day when we ran into the dog\n\u0026ldquo;Good morning dog\u0026rdquo;, I said\n\u0026ldquo;Good morning\u0026rdquo;, my dog said\nData Rot\nAI causes brain rot. If brain rot is what happens to human cognition when fed a diet of low-quality content, data rot is what happens to machine intelligence when fed itself, the slow epistemic decay of AI systems trained on AI-generated outputs, compounding across iterations until the signal is mostly echo. The Jevons paradox drives it: as AI makes data generation cheaper and faster, we produce vastly more of it, but abundance masks degradation. Datasets age, go uncurated, calcify around the assumptions of the moment they were built, and get recycled into the next generation of models as if they were fresh.\nThe rot compounds because static datasets don\u0026rsquo;t decay visibly; they just quietly diverge from a world that kept moving, while the models trained on them grow more confident but less accurate, mistaking familiarity for truth. The fix isn\u0026rsquo;t more data (that accelerates the paradox) it\u0026rsquo;s curation as a discipline, or curation as a service: living datasets that are actively pruned, dated, contested, and annotated for provenance, combined with models that can express uncertainty about the age and origin of what they know.\nThe algorithm won\u0026rsquo;t be the USP (AI will be a commodity) but the quality and quantity of your training data (defining your ontologies) and feature data (dynamic, curated data as the fuel for your AI engine). And someone has to do the curating. This is the wide open gap that is waiting to be filled by companies that can develop a \u0026ldquo;curation as a service\u0026rdquo; engine on top of data lakes, repositories, private databases and enterprise file and knowledge systems.\n\u0026ldquo;Ik houd mijn publiek graag (met een beetje humor) een spiegel voor!\u0026rdquo;\nJos Burgers is our hero. Jos doesn\u0026rsquo;t sell drills, Jos sells holes.\n1-9-5-3-7\nWhen you enter the male toilet area at eg a highway gas station, you see this:\nThe insight that \u0026ldquo;privacy is not the same as secrecy\u0026rdquo; can be felt strongly here as the typical filling pattern is: #1 \u0026gt; #9 \u0026gt; #5 \u0026gt; #3, etc; maximising your private space, in exchange for relief.\nWhen I enter the men\u0026rsquo;s bathroom at a gas station and I see #1 is occupied, I take #2.\nMaximum awareness. Maximum cringe. Maximum uncannyness.\n.\n","permalink":"https://thishollow.earth/posts/test_29/","summary":"\u003cp\u003eToday, I wrote nothing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e¯(°_o)/¯\u003c/strong\u003e\u003cbr\u003e\nThe asymetric nature of change:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eHumans change slow\u003c/li\u003e\n\u003cli\u003eThe machine changes fast\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u0026ldquo;The machine\u0026rdquo; as systems of power (politics, technology, institutions, finance). They are self-reinforcing, self-interested, and largely autonomous from the individuals inside them. A banker, a politician, a programmer; each feels like they are making choices. But the system they inhabit has its own logic, its own momentum. The machine moves, whether or not any individual wills it.\u003c/p\u003e","title":"Test 29"},{"content":"Scanning the surface.\nA preview edition.\nMore to come.\nAlways.\nThere\u0026rsquo;s always more to come.\nAnd always later.\nInteresting stuff 1\nI got an email from one of our IT service suppliers about how they joined another company for scale and efficiency reasons and how they made some changes in management. So far so good. So far all generic text and reassuring language. \u0026ldquo;Nothing will change for you\u0026rdquo;, \u0026ldquo;you can be sure you receive the quality you deserve\u0026rdquo; or something along those lines.\nI almost archived the mail when my eyes landed on the final sentence: \u0026ldquo;So, that\u0026rsquo;s why I wanted to schedule a call to get to know you better\u0026rdquo;, signed, the new manager. I thought that was surprising and strange and a bit weird (\u0026ldquo;is he going to call +200 clients?\u0026rdquo;). But I suggested some dates and 1 week later we were having the call.\nAnd it wasn\u0026rsquo;t awkward or weird at all. It was great. It was really nice to see behind the new \u0026ldquo;manager\u0026rdquo; there was a real human being. With a family and life and a work history and passion and enthousiasm. I was sold. I felt guilty for thinking it weird when humanity comes peaking through the AI dominated biz-speak and corporate blurps.\nAn evening stroll\nI was walking my dog the other day when we ran into the butterfly hanging the laundry\n\u0026ldquo;Good morning butterfly\u0026rdquo;, I said\n\u0026ldquo;Good morning\u0026rdquo;, the butterfly said\n\u0026ldquo;I quit smoking\u0026rdquo;, the butterfly said\n\u0026ldquo;You quit smoking\u0026rdquo;\n\u0026ldquo;I got married\u0026rdquo;, the butterfly continued\n\u0026ldquo;You got married\u0026rdquo;\n\u0026ldquo;I started drinking\u0026rdquo;, the butterfly added\n\u0026ldquo;You started drinking\u0026rdquo;\nMy dog was licking his balls\n\u0026ldquo;I underestimated the butterfly\u0026rdquo;, I thought while walking away\n\u0026hellip;\nInteresting stuff 2\nWe are in the age of AI. Or so the hype goes. Everyone and everybody is jumping on AI. And I think rightfully so. It is another abstraction level up in programming skills. It does leverage creative and skillful brains. It will make certain types of work irrelevant. But it will also create a lot of noise and create more work through the Jevons paradox. There will be a shift but maybe not in the way people think.\nOne of the things that will happen (is happening) is governments and companies and organisation building and training their own AI models. With Claude and Mistral and Gemini everyone has access to world class lawyers, big 4 consultancy advice or big 3 strategic analysis and board room plans. But everyone has access to the same lawyer, the same consultant, the same strategist.\nYou can get a competitive advantage by prompt engineering and context engineering. But those are hard to persist and get consistent. What organisations want is unique but predictable outcomes and output. So they build and train their own models, their own contexts and their own guard rails. To get good models, you need 3 things:\nfeature data training data a model As this is happening, most focus is on the modeling side. In the geospatial world (earth intelligence), the focus is on geofoundational models like Clay. And they are promising and will be great. But to make them competitive for companies, companies should build their own models on top of those foundational models. And they will.\nBut what I think is overlooked structurally is the need for good feature data. In geospatial that is consistent and coherent satellite composites and mosaics. And everything on top of that. The quality of your model is irrelevant when you have inconsistent feature data. And in geospatial, most feature data is still inconsistent and incoherent. Because of lack of standards and benchmarks. This is the reason the META global biomass is impressive, but unusable. The model, the math is on a different level from everything that came before. But because of crappy feature data, the output is crappy as well.\nThe feature data overall quality will improve over time. But for now, it\u0026rsquo;s a competitive advantage if your organisation uses consistent and coherent geospatial datacubes.\nInteresting stuff 3\nThis is so interesting it\u0026rsquo;s secret.\nInteresting stuff 4\nThis one is not secret but really not that interesting.\n.\n","permalink":"https://thishollow.earth/posts/test_28/","summary":"\u003cp\u003eScanning the surface.\u003c/p\u003e\n\u003cp\u003eA preview edition.\u003c/p\u003e\n\u003cp\u003eMore to come.\u003c/p\u003e\n\u003cp\u003eAlways.\u003c/p\u003e\n\u003cp\u003eThere\u0026rsquo;s always more to come.\u003c/p\u003e\n\u003cp\u003eAnd always later.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInteresting stuff 1\u003c/strong\u003e\u003cbr\u003e\nI got an email from one of our IT service suppliers about how they joined another company for scale and efficiency reasons and how they made some changes in management. So far so good. So far all generic text and reassuring language. \u0026ldquo;Nothing will change for you\u0026rdquo;, \u0026ldquo;you can be sure you receive the quality you deserve\u0026rdquo; or something along those lines.\u003c/p\u003e","title":"Test 28"},{"content":"Short one.\nGary doesn\u0026rsquo;t like Yann\nAI flame war, Gary Marcus puts Yann Lecun\u0026rsquo;s fame in perspective.\nAn evening stroll\nI was walking my dog the other day when we ran into the sphinx\n\u0026ldquo;Good morning sphinx\u0026rdquo;, I said\nThe sphinx said nothing\n\u0026ldquo;How was your weekend sphinx?\u0026rdquo;, I tried\nThe sphinx stared into the distance\nMy dog came up carrying a note in his mouth\nI deciphered the note:\n\u0026ldquo;The net serves to catch the fish; once the fish is caught, the net must be forgotten.\nThe snare serves to catch the rabbit. Once the rabbit is caught, the snare is forgotten.\nWords serve to catch meaning; once the meaning has been caught, the words can be forgotten.\nIf only I could find someone who has forgotten the words, someone I could talk to for a while.\u0026rdquo;[1]\nWe walked on\nMy dog looked around and glanced back at the sphinx one more time\n[1] Zhuang Zi, 3rd century BC\nEben Moglen\nA forgotten hero. I saw Eben Moglen live once, 14 years ago, in Bruxelles as a keynote speaker for FOSDEM. He has a very inspiring and engaging and charismatic personality. I completely forgot about him until a friend notified me of this recent speech in Berlin. He\u0026rsquo;s still great, though YouTube doesn\u0026rsquo;t do him justice.\nAI is an amplifier\nMadianou writes in Technocolonialism: \u0026ldquo;In other words, because it reflects ways of seeing the world, design ‘creates ways of being’. Such practices explain why automation and AI systems reproduce and heighten inequality (Eubanks, 2017) and racial and gendered discrimination (Benjamin, 2019; Noble, 2018).\u0026rdquo;\n\u0026ndash;---__\nChappie\nEnd\nO nee, toch niet, een stukje kantoorbeleving, dat kan dus gewoon.\n.\n","permalink":"https://thishollow.earth/posts/test_27/","summary":"\u003cp\u003eShort one.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eGary doesn\u0026rsquo;t like Yann\u003c/strong\u003e\u003cbr\u003e\nAI flame war, Gary Marcus puts Yann Lecun\u0026rsquo;s fame in \u003ca href=\"https://garymarcus.substack.com/p/the-false-glorification-of-yann-lecun\"\u003eperspective\u003c/a\u003e.\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://aneveningstroll.com\"\u003eAn evening stroll\u003c/a\u003e\u003c/strong\u003e\u003cbr\u003e\n\u003cem\u003eI was walking my dog the other day when we ran into the sphinx\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026ldquo;Good morning sphinx\u0026rdquo;, I said\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003eThe sphinx said nothing\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003e\u0026ldquo;How was your weekend sphinx?\u0026rdquo;, I tried\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003eThe sphinx stared into the distance\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eMy dog came up carrying a note in his mouth\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003eI deciphered the note:\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026ldquo;The net serves to catch the fish; once the fish is caught, the net must be forgotten.\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003eThe snare serves to catch the rabbit. Once the rabbit is caught, the snare is forgotten.\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003eWords serve to catch meaning; once the meaning has been caught, the words can be forgotten.\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003eIf only I could find someone who has forgotten the words, someone I could talk to for a while.\u0026rdquo;\u003c/em\u003e[1]\u003c/p\u003e","title":"Test 27"},{"content":"This one is about reading. And about life. La condition humaine.\nAnd a bit about AI and IT.\nAnd also a bit about the EO adoption cycle-curve-graph.\nAn evening stroll\nI was walking my dog the other day when we ran into the bear\n\u0026ldquo;Good morning bear\u0026rdquo;, I said\n\u0026ldquo;I\u0026rsquo;m not a bear\u0026rdquo;, the panda said\n\u0026ldquo;Alles blijft\u0026rdquo;\n\u0026ldquo;Alles gaat voorbij\u0026rdquo;\n\u0026ldquo;Alles blijft voorbijgaan\u0026rdquo;, the panda said\nThe panda stared into the horizon\nReading\nI like reading. I like reading a lot. It\u0026rsquo;s simultaneously a struggle, rewarding, humbling, mesmerising, depressing, uplifting, meditative, contemplative, isolating, bonding and all the other things take make humans human. I mostly read novels but I also like to read essays, non-fiction or poetry. A lot of writing is about writing itself. Or about being, or rather becoming, an author. That\u0026rsquo;s where the writing about reading mostly manfifests itself. But I find it often superficial: the author as a child or adolescent loves to read and reads and reads and reads. As if it\u0026rsquo;s a static, repetitive, automated experience that you can activate and de-activate at will.\nThat is not how I experience reading. I like it but it\u0026rsquo;s often a struggle. Especially when starting with a new, unfamiliar author or genre. You have to get \u0026lsquo;into\u0026rsquo; their style, their poetica, their themes, their hearts. But once it clicks, it\u0026rsquo;s mostly very rewarding and you find yourself with a new problem: \u0026ldquo;I need to read all books of this author!\u0026rdquo;.\nOr I simply \u0026lsquo;forget\u0026rsquo; about a book. I read multiple (4 - 6) books simultaneously, all at different paces. It often so happens that I start a book, find another interesting book, start that as well, get hooked, forget about the first book which then might temporarily get lost only to serendipically resurface weeks or months or years later.\nI don\u0026rsquo;t have any set reading hours or pace or structure but I do try to adhere to the rule that once started, a book should be finished. When I\u0026rsquo;m busy, I read slow. When I\u0026rsquo;m on holiday I also read slow, but a lot.\nMy love for reading is also one of the reasons why I love work. Work, or, going to the office. Because that is where other humans are, your colleagues. Some of whom also read. And they always read more interesting stuff than you do so you need to talk to them and ask them what they are reading. Ok, not all of them read interesting stuff. Not even all of them read. But that doesn\u0026rsquo;t matter. You will find that new colleague that introduces you to that new book by that one new author you never heard of.\nSo you buy the book.\nAnd so you start reading. Excitedly. And then it doesn\u0026rsquo;t click. Of course.Of course it doesn\u0026rsquo;t click. New authors never click for at least the first 20 - 30 pages. You read on, slower and slower, struggling and struggling, feeling embarassed when your colleague kindly informs about the progress (\u0026rsquo;not yet\u0026rsquo;, \u0026lsquo;just started\u0026rsquo;). You want to like it. But you don\u0026rsquo;t. And your colleague politely stops asking. And you\u0026rsquo;re wrestling with page 43. Page 43, with it\u0026rsquo;s mission to never let you arrive at page 44. Page 43 being very successfull in it\u0026rsquo;s mission.\nThere\u0026rsquo;s the book, giving you the finger every night from it\u0026rsquo;s spot halfway the pile of novels on your bedside table. Day after day, week after week, month after month.\nBut my bedside pile is patient. And holiday happened and I read Never let me go by Kazuo Ishiguro. I read the final 40 pages sitting on the couch in my living room on a grey Monday morning. I finished the last sentence at 11:34. And I kept sitting on my couch, staring into the grey sky for more than 40 minutes. Thinking about the book, the story, about how it captures being human perfectly.\nIn the evening, my kid asked if I had a fun book about math. I said, \u0026ldquo;no, all my books are about life and death and misery and being happy nevertheless\u0026rdquo;. He rolled his eyes. \u0026ldquo;But I should have Gödel Escher Bach somewhere\u0026rdquo;. I dug up the book and gave it to hime. Three minutes later, he gave it back. Too boring.\nAt night, feeling very motivated for finishing a book and feeling very inspired by Never let me go, I looked at the pile. And Playground smiled at me. I picked it up and (re)started reading. Page 43 was a pushover, page 44 and 45 went down on TKO and before I knew it I was calmly ploughing through page 50, 60, 70, 80, \u0026hellip;\nAND THEN IT HAPPENED. IT HAPPENED. IT CLICKED.\nThe chess match is where it clicked. Page 85. The first time the 2 protagonists play a chess match is where it suddenly all made sense. The contradictions, the engineer and the poet. The engineer being obsessed with the mechanics of the game, the poet with the humanity. Both playing with the same fierce intensity. Of course, on page 87, the engineer finds the poet reading Gödel Escher Bach. The book I dug up for my kid just 2 hours before. Of course, they discover the game of Go, a game with simple rules but endless complexity (where did we hear that before?). Of course the book explores themes of capitalism, libertarianism, autonomy, democracy, freedom, information, knowledge, environmentalism via beautifully interwoven story lines.\nAnd of course on page 139, the main character reflects on his dementia: My doctor told me about a thing that dementia patients do called \u0026ldquo;showtiming\u0026rdquo;. In denial, embarassment, or terror, they peform themselves in front of other people as if they have no symptoms at all. Perfectly describing the earlier stages of my own fathers 2 year dance with Alzheimers. Dance? Yes, dance. Not struggle, not battle, not fight: dance. He dances with the disease. First he led the disease in an elegant tango of showtiming, charisma and optimism. Fooling us that all was well. Until about a year ago, barely noticeable, the disease took over and started leading him (without him realising) into the inevitable danse macabre, slowly but certainly.\nI devoured the rest of the book.\nOn page 246 the poet turned the engineer on to a thin little book of gnomic aphorisms called Finite and Infite Games by James P. Carse. Finite and Infite games. Where did we hear that before?\nReading is beautiful. You should read as well.\nThe AI effect\nMy friend succesfully defended her PhD thesis in anthropology the other day in Leiden University about the tech startup culture in Singapore in the 2010s. I had never been to a PhD defense in Leiden before. It was beautiful, exhilarating and humbling.\nWhile standing in line for congratulating the candidate, I found one of the committee members standing before me. We came to talk about AI and I mentioned my observation about the semantic narrowing of the use of the term AI. She smiled and said, well, what you just described is called the AI effect, you can look it up on WikiPedia.\nI did that for you:\nThe AI effect is the discounting of the behavior of an artificial intelligence program as not \u0026ldquo;real\u0026rdquo; intelligence.\nThe author Pamela McCorduck writes: \u0026ldquo;It\u0026rsquo;s part of the history of the field of artificial intelligence that every time somebody figured out how to make a computer do something—play good checkers, solve simple but relatively informal problems—there was a chorus of critics to say, \u0026rsquo;that\u0026rsquo;s not thinking\u0026rsquo;.\u0026rdquo;\nResearcher Rodney Brooks complains: \u0026ldquo;Every time we figure out a piece of it, it stops being magical; we say, \u0026lsquo;Oh, that\u0026rsquo;s just a computation.\u0026rsquo;\u0026rdquo;\nsmalltalk\nI can\u0026rsquo;t do smalltalk. I just can\u0026rsquo;t. Death and decay are always 1 step away, patiently waiting. I cannot not talk about life.\nTech radar\nMy favourite twice-yearly opinionated guide to today\u0026rsquo;s technology landscape recently published their 2025 H2 edition. I didn\u0026rsquo;t count the blips but it feels like more than half are about AI. Some thoughts:\nI like reading \u0026ldquo;old\u0026rdquo; IT books and blogs. Partly because it reminds us of how fast developments go. How fast things are absolete. But mostly because there\u0026rsquo;s also fundamental laws, truths, structures that withstand the wheel of time. Like Joel Spolsky\u0026rsquo;s blog on leaky abstractions from 2002. Context engineering is a leaky abstraction.\nAI is the new Excel, another version of the Jevons Paradox.\nRelease It!\nNot as old as Spolsky\u0026rsquo;s 00s rambling but 2018 is still already old in tech. Release It! talks about stability being your friend in complex IT systems. Stability is a necessary condition for performance. There\u0026rsquo;s no point in optimising for performance when your system is not stable. Like consistency and accuracy. Accuracy is a vanity metric when not consistent.\nThe EO adoption cycle\nAravind dropped another brilliant mental model: the EO Adoption Curve:\nEarth observation doesn\u0026rsquo;t win when it\u0026rsquo;s visible - it wins when it disappears.. Reminiscent of Arthur C. Clarke\u0026rsquo;s 3rd law: Any sufficiently advanced technology is indistinguishable from magic. Arthur C. Clarke\u0026rsquo;s 3rd law, or, The inverse AI effect.\nOr, my own \u0026ldquo;controversial take on EO\u0026rdquo;: (y)our clients don\u0026rsquo;t care about satellites, they want their problems solved.\nEinde.\n","permalink":"https://thishollow.earth/posts/test_26/","summary":"\u003cp\u003eThis one is about reading. And about life. La condition humaine.\u003c/p\u003e\n\u003cp\u003eAnd a bit about AI and IT.\u003c/p\u003e\n\u003cp\u003eAnd also a bit about the EO adoption cycle-curve-graph.\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://aneveningstroll.com\"\u003eAn evening stroll\u003c/a\u003e\u003c/strong\u003e\u003cbr\u003e\n\u003cem\u003eI was walking my dog the other day when we ran into the bear\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026ldquo;Good morning bear\u0026rdquo;, I said\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003e\u0026ldquo;I\u0026rsquo;m not a bear\u0026rdquo;, the panda said\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026ldquo;Alles blijft\u0026rdquo;\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003e\u0026ldquo;Alles gaat voorbij\u0026rdquo;\u003c/em\u003e\u003cbr\u003e\n\u003cem\u003e\u0026ldquo;Alles blijft voorbijgaan\u0026rdquo;, the panda said\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eThe panda stared into the horizon\u003c/em\u003e\u003c/p\u003e","title":"Test 26"},{"content":"Why not have another edition of This Hollow Earth?\nHere\u0026rsquo;s some random ramblings and thoughts that circled my brain while doing the dishes.\nget rid of your dishwasher\nDishwashers are a source of stress. Do the dishes yourself instead: 30 minutes meditative work per day keeps the doctor away. Alternatively: buy a dog.\nAn evening stroll\nI was walking my dog the other day when we ran into the kingfisher\n\u0026ldquo;Good afternoon kingfisher\u0026rdquo;, I said\n\u0026ldquo;Good afternoon\u0026rdquo;, the kingfisher said\n\u0026ldquo;Why are you listening to music?\u0026rdquo;, I asked\n\u0026ldquo;I\u0026rsquo;m not listening to music\u0026rdquo;, the kingfisher answered\n\u0026ldquo;Why are you eating?\u0026rdquo;, I continued\n\u0026ldquo;I\u0026rsquo;m not eating\u0026rdquo;, the kingfisher replied\n\u0026ldquo;Why are you drinking beer?\u0026rdquo;, I asked\n\u0026ldquo;I\u0026rsquo;m not drinking beer\u0026rdquo;, the kingfisher insisted\nI smiled at the kingfisher and the kingfisher smiled back\nOh Big head, where is thy Long tail?\nOne of the charms (and why investors like them) of SaaS companies is their promise to address the long tail. In that sense, \u0026ldquo;we\u0026rsquo;re a SaaS company\u0026rdquo; (implying long tail growth opportunities) must be the most mis-used term for labeling your start-, excuse me, scale-up. Not in terms of what SaaS means semantically, but what it means emotionally; Saas implies \u0026ldquo;Scale\u0026rdquo;. Then, you\u0026rsquo;re not a SaaS if you do consulting for your launching customer. You\u0026rsquo;re not a SaaS when your growth + operating margin is \u0026lt; 40%. And you\u0026rsquo;re not a SaaS when your servicing the big head without a viable Go To Market for addressing the long tail.\n\u0026ldquo;Bridging the gap\u0026rdquo; is the classic jump from early adopters to early majority. That is indeed the most challenging hurdle to overcome when building a company. But I would like to argue there are many more gaps or local optimums where companies get stuck. Getting stuck in the Big head is one of them.\nThere are many reasons for that is but one of the most overlooked ones is that companies start playing the finite game instead of the infinite game. I didn\u0026rsquo;t know about the infite game concept until recently but it provides a great lens for analysing why companies get stuck:\nPlanet Labs optimised for winning DoD contracts instead of building a commercial platofrm Cleantech 1.0 Startups (2006 - 21013): solving for subsidies, not markets Blockchain supply chain traceability: technology in search of a problem We can dig deeper in these cases if you like, and I will. But let\u0026rsquo;s first have a look at what I think is the #1 finite game mistake of many, many (EO based) service providers that deliver environmental landscape indicators: competing on accuracy. That is why they get stuck in consulting for their launching customers(s), that is why they get stuck in the Big\nAccuracy is a red herring, accuracy is bike shedding\nEssay coming up! I will write a blog on why focusing on accuracy will yield mild wins in the short term but will kill your business in the long term. In short, how is competing on accuracy playing the finite game?\nWhen you benchmark \u0026ldquo;95% accuracy!\u0026rdquo; against yourselves (or the clients ask!), you\u0026rsquo;re playing a defined game with clear rules and an endpoint: \u0026ldquo;We win when our accuracy number is higher than it was last quarter.\u0026rdquo;\nThis creates finite game behaviors:\nMeasurable endpoint: 95% → 96% → 97% → 99%\u0026hellip; then what? You \u0026ldquo;won\u0026rdquo;? Zero-sum thinking: Resources spent chasing 2% accuracy improvement = resources NOT spent building adaptive capabilities Easy to communicate: Executives love numbers. Boards love KPIs. \u0026ldquo;We improved accuracy by 3%!\u0026rdquo; sounds like progress. Bikeshedding: Everyone can have an opinion about accuracy. It\u0026rsquo;s like debating the color of the bike shed instead of whether you need a bike shed at all. But here\u0026rsquo;s the trap: Your customers don\u0026rsquo;t actually buy \u0026ldquo;accuracy.\u0026rdquo; They buy:\nRisk reduction Compliance confidence Decision-making speed Supply chain visibility Business model enablement You should focus on these points, the things that make customers tick. What is the superpower customers gain with your service?\nStay tuned for more.\nNo images this time but there\u0026rsquo;s an image of cake in The staff ate it later\nSysiphus is sweating sweet sweat\nNot all tears cried are bitter. Not all sweat is salty. The tears of painting yourself into an ever smaller corner are bitter. Sweating while biking uphill through the autumn forest is sweet sweat.\ntrust and confidence\nLast time we argued that trust is the currency of productivity. Jokes are an expression of trust. I only insult people that I trust.\nFin.\n","permalink":"https://thishollow.earth/posts/test_25/","summary":"\u003cp\u003eWhy not have another edition of This Hollow Earth?\u003c/p\u003e\n\u003cp\u003eHere\u0026rsquo;s some random ramblings and thoughts that circled my brain while doing the dishes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eget rid of your dishwasher\u003c/strong\u003e\u003cbr\u003e\nDishwashers are a source of stress. Do the dishes yourself instead: 30 minutes meditative work per day keeps the doctor away. Alternatively: buy a dog.\u003c/p\u003e\n\u003chr\u003e\n\u003cp\u003e\u003cstrong\u003e\u003ca href=\"https://aneveningstroll.com\"\u003eAn evening stroll\u003c/a\u003e\u003c/strong\u003e\u003cbr\u003e\n\u003cem\u003eI was walking my dog the other day when we ran into the kingfisher\u003c/em\u003e\u003c/p\u003e","title":"Test 25"},{"content":"Raw, unordered and unstructured. Let me know which one you want me to write about.\n1 step forward 2 steps back The illusion of progress\nSaving the forest and Black Friday The cloud doesn’t exist, there’s only other peoples computers\nNL and EU leading digitisation Digitalisation, NL, EU, analogous processes\nIn NL we are already used to the Digital only paradigm while probably a lot of our customers are have many analogue operational processes or parts of processes.\nThat is an important thing to realise when developing\nJust enough structure Team topologies Implicit leadership wegwerpstructuur\nOptimise throughput To not get blocked in decision making when sorting piles of stuff / ideas / crap, make 3 piles:\nThrow away Keep Doubt Don’t let lack of choice block throughput, a tool to do that is above piles. Describe problem \u0026lt;\u0026gt; solution process\nBut that doesn’t always solve the problem\nSure and not so sure type people, how do they approach this?\nTypes of people vs choice matrix:\nThe tool comes with a mindset or approach, if that is lacking, the tool is useless\nTools provide most leverage to people with already proper mindset\nLadders of abstraction Semantic hijacking\nDevlid stickers Each sticker is a story\nHacker camp Kid prodigy with chaperone for social interaction CCC boycotting OHM SHA with kids Hacker Mindset\nMy mother was a feminist But I didn’t know\nI see persons, all of them idiots that I despise\nSurprises We don\u0026rsquo;t expect any surprises\nNotes on the synthesis of form\nSkills, experts \u0026amp; visionaries Someone who made all mistakes in a certain field\nShow the toolbox:\nskill = a filled toolbox expert = knowing which tool to pick visionary = knowing the tool is not there (yet) but will be some day Heuristics table Decision tree flowchart\nVision is listening to your surroundings and be open for change Most visionaries are sending\nHuman creativity is the import and export of ideas via communication. Communication is imperfect evolution of ideas.\nThe perils of financialisation https://twitter.com/mikesington/status/1608932204784287746?s=46\u0026amp;t=LQTAbXfaP18t78vrO4IM3w\nFinancialisation of HR, see HBR article\nJapanese service level JR refund story Matsumoto\nEO services are foreign food Looks tempting, definitely interesting, sometimes tasty, sometimes horrific but you only know if it\u0026rsquo;s for you when you taste it.\nEarly adopter paradigm trap Don’t let early adopters catch you in the incumbent paradigm: they will forward requirements based on the context of their current paradigm; for example: desktop thinking where each file is local, discrete and a snapshot in time versus cloud where there is a remote, constantly updated data stream\nNobody wants insights Blogify the Mollema presentation “No one wants pixels, they want insights” is wrong\nArchitecture elevator A. Pushing buttons is lost knowledge\nB. Writing (code) is potential knowledge\nC. Architectured code is shared knowledge\nPotential knowledge is useless\nFrom potential knowledge to shared knowledge\nMonitoring versus surveillance Monitoring: continuously assess state of the system and its components\nSurveillance (tracking): assess state of objects or subjects (persons) whithin or relative to a system or it\u0026rsquo;s surroundings.\nHidden work Spillover The company on paper, the company in life The map and the territory\nThe role of companies in society Neoliberal reductionism responsibility Long term thinking (Roman Krnaric)\ntime rebels regenarative economy burger beraden new political institutions Schmitt: partijpolitiek is slecht (party democracy, nation state, consumerism culture, capitalism) Pace layers Doughnut economy Capitalism explained with wanking horses: https://twitter.com/peterkwint/status/1634304719555313664?s=46\u0026amp;t=9MvwxenI3xquqBGmiWFYLA \u0026hellip; \u0026hellip; State slaves (Orwell) vs Corporate slaves (2020)\nHuman beings were trying to behave as human beings and not as cogs in the capitalist machine\nAccumulation / centralisation is not good: wealth, information, access, knowledge,\nCommoditisation of Carbon \u0026amp; Biodiversity \u0026amp; nature Reductionism, NFT, web3, privatisation, Political role of companies Inverse tragedy of the commons Private companies donate / tax / fund institutions with electable boards? Executive institutions, policy making institutions, checking institutions Financialisation cost based accounting\nDebt Debt implies Capital, negative capital, borrowed. The interest is the problem, but only if it impedes growth (\u0026lt;\u0026ndash; growth is a problematic term?)\nAdd to debt blog Financial Technical Organisational debt Consultancy is aflossingsvrije hypotheek we lack trust not urgency Structure replaces trust The more trust, the less structure is needed Cultural capital Social capital the organisation as part of society, not disconnected from society Gradual Appreciate the insignificant life changing moments Those moments are not announced by the trumpets of the apocalypse You have to recognise them to appreciate them Quote Pieter Bieri (??) nachttrein naar Lissabon Heb vertrouwen dat de toekomst zal brengen wat je ervan verwacht.\nErrors are inevitable. Incrementalism makes it easier to correct them.\nIncrementalism calls for ceaseless change: kaizen\nThe consultant effect Description:New product in new market Higher is better Accuracy instead of consistency\nConsultants writing RFPs\nParadox: refer to SVPG post about how McKinsey will block transformational change\nWhy misunderstandings are ok The idea that you will understand a person is wrong Energy spent on making your point is fruitless Confidence \u0026amp; Trust ","permalink":"https://thishollow.earth/ideas/","summary":"\u003cp\u003eRaw, unordered and unstructured. Let me know which one you want me to write about.\u003c/p\u003e\n\u003ch2 id=\"1-step-forward-2-steps-back\"\u003e1 step forward 2 steps back\u003c/h2\u003e\n\u003cp\u003eThe illusion of progress\u003c/p\u003e\n\u003ch2 id=\"saving-the-forest-and-black-friday\"\u003eSaving the forest and Black Friday\u003c/h2\u003e\n\u003cp\u003eThe cloud doesn’t exist, there’s only other peoples computers\u003c/p\u003e\n\u003ch2 id=\"nl-and-eu-leading-digitisation\"\u003eNL and EU leading digitisation\u003c/h2\u003e\n\u003cp\u003eDigitalisation, NL, EU, analogous processes\u003c/p\u003e\n\u003cp\u003eIn NL we are already used to the Digital only paradigm while probably a lot of our customers are have many analogue operational processes or parts of processes.\u003c/p\u003e","title":"ideas"},{"content":"We\u0026rsquo;re back.\n\u0026ldquo;Arjen, what happened to your newsletter?\u0026rdquo;\nWell, this happened. And this.\nAnd now we\u0026rsquo;re back.\nLet\u0026rsquo;s pick it back up slow and easy kids.\nScience, research, conferences\nThe scientific community and their conferences can be a crazy bunch. My French friend Sonia went to this conference organised by a person who identifies as a pirate. Walks around the conference dressed up as a pirate. Very cool research though.\nAt the same conference, the, apparently, (in)famous Johan Rockström hang out (or maybe I mix up events). Never heard of the kind man (impossible personality) but he is (one of the) originators of the Planetary Boundaries Framework.\nThe planetary boundaries framework highlights the rising risks from human pressure on nine critical global processes that regulate the stability and resilience of the Earth\nVery relevant, very cool, very serendipic.\n\u0026ldquo;You have too many books!\u0026rdquo;\nI found there\u0026rsquo;s a word for my condition: But luckily Umberto Eco comes to the rescue with his antilibrary.\nAI is not \u0026hellip;\nSemantic stretching can be tricky. You think you agree but you might be talking about completely different topics. Especially in product design and development. Platform, dashboard, portal, insight, user friendly. Everyone agrees, no-one has a clue.\nBut lately I find myself increasingly annoyed with semantic narrowing. For example, \u0026ldquo;sport\u0026rdquo; used to include all sorts of leisure activities like running, rowing, biking, swimming, football or tennis. No, not tennis, tennis is a game. But nowadays, when people say they go \u0026ldquo;sporting\u0026rdquo;, they mean going to the gym (at least, in The Neterlands). Same for AI. AI used to be the umbrella term for regression models, machine learning, deep learning, decision trees, neural networks, computer vision, natural language processing and what not.\nNow, when people (or organisations) ask: \u0026ldquo;Do you use AI?\u0026rdquo;, they mean: \u0026ldquo;Do you use generative AI based on Large Language Models (LLMs)?\u0026rdquo;. Which is not only a subset of AI, it\u0026rsquo;s also a dead end. LLMs are probably not the foundational building blocks towards the singularity (or, General AI, or AGI). The kind folks at Apple kindly did some research showing how and why in their paper (The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity](https://machinelearning.apple.com/research/illusion-of-thinking).\nRead it. And pick a real sport.\nAn evening stroll\nI was walking my dog the other day when we ran into the chicken\n\u0026ldquo;Good morning chicken!\u0026rdquo;, I said\n\u0026ldquo;Good morning \u0026hellip;\u0026rdquo;, the chicken mumbled\n\u0026ldquo;I have to leave. If they ask, you didn\u0026rsquo;t see me.\u0026rdquo;\nI looked at the chicken\n\u0026ldquo;I don\u0026rsquo;t want to go. But I have to go. I must go.\u0026rdquo;\n\u0026ldquo;I know chicken. I understand.\u0026rdquo;, I said, \u0026ldquo;I never saw you\u0026rdquo;\n\u0026ldquo;Goodbye chicken \u0026hellip;\u0026rdquo;\nTrust\nThe currency of productivity.\nStuff to look into\nLinkedIn is a Fucked up circus. Flee. Already happened but needs to be shared: My want of you partakes of me Data Science at the Command Line Note: Test 1 - Test 23 were published on an internal company forum. I will revisit some of that stuff going forward. Maybe.\nFinally, remind me to write about this some day, I was testing Claude against ChatGPT sycophancy by stating that Disney is worse than crack. Guess what they said?\n","permalink":"https://thishollow.earth/posts/test_24/","summary":"\u003cp\u003eWe\u0026rsquo;re back.\u003c/p\u003e\n\u003cp\u003e\u0026ldquo;Arjen, what happened to your newsletter?\u0026rdquo;\u003c/p\u003e\n\u003cp\u003eWell, \u003ca href=\"https://thishollow.earth/\"\u003ethis\u003c/a\u003e happened. And \u003ca href=\"https://aneveningstroll.com/\"\u003ethis\u003c/a\u003e.\u003c/p\u003e\n\u003cp\u003eAnd now we\u0026rsquo;re \u003ca href=\"https://thishollow.earth/posts/test_24\"\u003eback\u003c/a\u003e.\u003c/p\u003e\n\u003cp\u003eLet\u0026rsquo;s pick it back up slow and easy kids.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eScience, research, conferences\u003c/strong\u003e\u003cbr\u003e\nThe scientific community and their conferences can be a crazy bunch. My French friend Sonia went to this conference organised by a person who identifies as a \u003ca href=\"https://piratepeel.github.io/\"\u003epirate\u003c/a\u003e. Walks around the conference dressed up as a pirate. Very cool research though.\u003c/p\u003e\n\u003cp\u003eAt the same conference, the, apparently, (in)famous Johan Rockström hang out (or maybe I mix up events). Never heard of the kind man (impossible personality) but he is (one of the) originators of the \u003ca href=\"https://www.stockholmresilience.org/research/planetary-boundaries.html\"\u003ePlanetary Boundaries Framework\u003c/a\u003e.\u003c/p\u003e","title":"Test 24"}]