<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Wale Ayandiran (walecloud) on Medium]]></title>
        <description><![CDATA[Stories by Wale Ayandiran (walecloud) on Medium]]></description>
        <link>https://medium.com/@walecloud?source=rss-d500c312e4bb------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*Kbiip40WobMK6n_5.jpeg</url>
            <title>Stories by Wale Ayandiran (walecloud) on Medium</title>
            <link>https://medium.com/@walecloud?source=rss-d500c312e4bb------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 17:16:18 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@walecloud/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Reigniting it, totally out of your control btw.]]></title>
            <link>https://medium.com/@walecloud/reigniting-it-totally-out-of-your-control-btw-8a677022d84d?source=rss-d500c312e4bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/8a677022d84d</guid>
            <category><![CDATA[work]]></category>
            <category><![CDATA[journey]]></category>
            <category><![CDATA[life]]></category>
            <category><![CDATA[growth]]></category>
            <category><![CDATA[life-lessons]]></category>
            <dc:creator><![CDATA[Wale Ayandiran (walecloud)]]></dc:creator>
            <pubDate>Mon, 16 Nov 2026 01:16:48 GMT</pubDate>
            <atom:updated>2026-11-17T10:31:27.952Z</atom:updated>
            <content:encoded><![CDATA[<h3>Finding Your Spark in Hardship — Reflections on Growth</h3><p>Sparks and sparkles… what even is that?</p><p>You find your spark when you’re usually not looking. And it often comes after you’ve gone through some form of turbulence that lingered longer than you imagined it would.</p><p>Sometimes these hardships are visible to the world. Often, they’re solely internal, and the closest person to you might have zero clue.</p><p>Think about someone who carries a considerable part of their family’s responsibilities. The person making sure things work, playing the parental role, looking after everyone, remembering what needs to be done and somehow finding a way to do it.</p><p>They come home and play with the kids. They’re bubbly, making jokes, asking everyone about their day. There’s food to sort out, something in the house needs fixing, someone needs help with something that apparently cannot wait until tomorrow.</p><p>And they’re there for all of it.</p><p>Meanwhile, work has become a source of pressure they don’t quite know how to explain. The workload keeps growing. Even away from it, part of their mind is occupied by what remains unfinished and what tomorrow will demand of them.</p><p>But the children still get a playful parent. Their partner still gets someone who participates. At work, things still get delivered.</p><p>From the outside, what exactly has changed?</p><p>Perhaps very little. Which makes it quite difficult for anyone to notice that something has.</p><p>And I don’t think the laughter necessarily has to be fake. You can enjoy playing with your children and still carry something painful that the enjoyment doesn’t resolve. Being lively around people you love doesn’t mean everything within you is equally alive.</p><p>You’re still doing the things. Some of them quite well, actually. Yet you can feel a difference that you cannot point to in any of the things you’ve done.</p><p>Then, at some point, you say, “I have my spark back.”</p><p>What do we mean when we say that?</p><p>For me personally, I think we don’t lose our spark and find it again. I mean, we do, but perhaps not in the sense that it left.</p><p>Think about it. How often do we lose “this spark” and find it again and again and again?</p><p>Did we ever lose it in the first place? Where exactly does it go while we continue being a parent, a partner, a friend, an employee and all the other things we’ve agreed to be?</p><p>Could it be that what we sometimes describe as losing our spark is a lack of the almost obsessive discipline required to follow through on a promise we’ve made to ourselves?</p><p>We wanted to accomplish something. We were excited about it. Then the excitement wore off, and continuing became considerably less interesting than starting.</p><p>Motivation is a great feeling, but one that evaporates as time goes by. Relying on it to keep going is like expecting the wind to blow in your direction for the entire journey.</p><p>There are things you simply have to continue doing after the feeling that started them has disappeared. I believe that.</p><p>But hold on… the person I described earlier hasn’t stopped doing anything.</p><p>They’re following through every day. People depend on them, and they keep showing up. If discipline were the entire explanation, surely they would have enough of it by now to feel like themselves.</p><p>So why did I reach for discipline?</p><p>It started with a question about something we feel within ourselves and somehow arrived at whether we’re doing enough. Interesting.</p><p>Perhaps I’m mixing together experiences that happen to share the same expression.</p><p>Losing interest in a particular ambition. Being worn down by responsibilities. Finding very little enjoyable. Being unable to begin something you still care about. We might describe all of these as losing our spark, but I’m not sure they’re asking the same thing of us.</p><p>And if we don’t know which one we mean, how do we know what we’re trying to reignite?</p><p>For the person carrying their family and their workload, another promise to become more disciplined might simply become another thing to carry.</p><p>There’s also something curious about how we notice its return.</p><p>Sometimes you realise you’ve spent an hour doing something without having to negotiate with yourself first. You become curious about a thing that has no immediate relevance to any of your responsibilities. You look forward to a day without first checking what you’ll accomplish in it.</p><p>Nothing particularly remarkable to report. You might even have got less done.</p><p>Would you still say your spark was back?</p><p>Would the people around you recognize it if its return meant you were less available to do everything they had become accustomed to you doing?</p><p>That question makes the whole thing slightly less comfortable.</p><p>Because being dependable can become such a familiar part of how people experience you that there’s very little reason for them to question what it takes. Things are working. You’re laughing. You said you were fine.</p><p>And perhaps you don’t know what else to say. Explaining it would require understanding it first, and you’ve barely had time to sit with it yourself.</p><p>I’m beginning to think there are limits to what we can demand back from ourselves through effort alone. We can make decisions, change what is possible to change, keep certain promises. But I don’t know that we can instruct ourselves to feel interested, light or hopeful on a particular date.</p><p>The spark has refused the calendar invite. Very unprofessional.</p><p>Still, I’m not entirely convinced it disappeared.</p><p>Perhaps what changed was how much of our attention and energy was already spoken for before we had a chance to decide what to do with it. Perhaps that explains some of it. It certainly doesn’t explain every season in which a person feels far away from themselves.</p><p>I wonder how much of that we can deliberately bring about.</p><p>Because if I knew exactly what brought it back, surely I would know what to do the next time it disappeared. Yet somehow, each time, I’m back here trying to understand something I thought I’d already figured out.</p><p>And there’s something about the word “back” too.</p><p>It assumes I know the version of myself I’m trying to return to. That the person before the turbulence is still the person I want to be after it.</p><p><em>Here’s a picture of furry friends back in the days… the year was 2020</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*N_rUnORVFMOoIHul" /></figure><p><em>Originally published at </em><a href="https://wales-note.beehiiv.com/p/sparks-sparkles"><em>https://wales-note.beehiiv.com</em></a><em> on November 16, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8a677022d84d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I Know What I Meant — Structure your agent work]]></title>
            <link>https://medium.com/@walecloud/i-know-what-i-meant-structure-your-agent-work-381e9546c502?source=rss-d500c312e4bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/381e9546c502</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[agents]]></category>
            <category><![CDATA[onbuddy]]></category>
            <category><![CDATA[openai]]></category>
            <dc:creator><![CDATA[Wale Ayandiran (walecloud)]]></dc:creator>
            <pubDate>Mon, 26 Oct 2026 23:37:27 GMT</pubDate>
            <atom:updated>2026-10-27T09:07:59.802Z</atom:updated>
            <content:encoded><![CDATA[<p>I’ve been working on the OnBuddy Workspace product, the website and OnBuddy Journey this weekend. Alongside that work, I’ve been using Workspace for some of the work of building OnBuddy itself.</p><p>That has given me something quite specific to think about.</p><p>My own use is useful to me, but it is only one way into the product. A customer’s use case has made that distinction more concrete.</p><p>In my case, I give , the coding agent, a task, and it logs that work in my . It could be a bug, an improvement, a quick fix or a feature. The documentation about the work also has somewhere to live. When I prompt Codex, it can pick tasks up from OnBuddy and work through them. <a href="https://onbuddy.ai/workspace">OnBuddy workspace</a></p><p>Some of those tasks begin as quite vague thoughts. There is something I want changed, or an improvement I want to return to, without necessarily working through all the details at that moment.</p><p>The workspace gives those thoughts somewhere to become work I can come back to.</p><p>And because I’m the one working on the product, I already have a lot of the surrounding context. I know why I raised something and what I was trying to achieve when I raised it.</p><p>A customer has a different use case.</p><p>They use meeting notes imported from Google Meet to create tasks and activities, with OnBuddy inferring which members should own the tasks.</p><p>It looks similar to my own use if you describe both as creating and managing tasks. But the work around those tasks is different.</p><p>In my workflow, I’m bringing an intention to a coding agent and asking it to record or carry out the work. In the meeting workflow, the starting point is a conversation involving people. What someone agreed to do, who should take responsibility and what needs to happen afterwards all become relevant.</p><p>Even ownership means something different in those circumstances.</p><p>The person who mentions an issue may not be the person who should resolve it. An idea discussed during a meeting may not yet be a decision to do anything. Those distinctions matter when a conversation becomes assigned work.</p><p>I find this part of building interesting because it makes a familiar feature less simple the longer you look at it.</p><p>You can say that a product turns a document into tasks, and that can be an accurate description of what it does. It still leaves quite a lot unexplained about what the person needs from it.</p><p>For me, a task might preserve a thought I want Codex to work through later. For a team, it might represent a responsibility emerging from a conversation.</p><p>The task can look much the same on a screen while carrying different expectations for the people involved.</p><p>And those expectations affect what it means for the product to work well.</p><p>This is why I don’t want to assume that everyone should use the workspace the way I do.</p><p>My own use lets me experience the product directly, but it also gives me a very particular view of it. I know the work, I know the language I use to describe it, and I know what I meant when I wrote something down.</p><p>Other people bring knowledge of their own work. They also bring habits and responsibilities that I may have no reason to encounter in mine.</p><p>That seems obvious when written down. It becomes more consequential when I’m making decisions about the product.</p><p>What feels like an unnecessary step in my workflow might have a purpose in somebody else’s. Something I would happily leave informal might need to be explicit when several people are involved.</p><p>I’m trying to pay attention to those differences without treating every difference as a request for another feature.</p><p>Sometimes there is a product change to make. Sometimes I need to understand more about the work before deciding what, if anything, should change.</p><p>One customer’s use case cannot tell me everything about how teams work, any more than my own use can.</p><p>But it gives me something real to examine alongside my own experience.</p><p>I can look at the same task in both workflows and ask where it came from, what someone expects it to mean and what they will do with it next.</p><p>The more I do that, the more I find myself returning to the conversation around the work.</p><p>What happened before this became a task? What did the people involved understand that the task alone might not tell me?</p><p>Those are the questions I’m taking back into the product as I continue building.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*SE8KRPq0hhvfhBms" /></figure><p><em>Originally published at </em><a href="https://wales-note.beehiiv.com/p/i-know-what-i-meant"><em>https://wales-note.beehiiv.com</em></a><em> on October 26, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=381e9546c502" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[On Taking Breaks — And the strange privilege of being able to stop.]]></title>
            <link>https://medium.com/@walecloud/on-taking-breaks-and-the-strange-privilege-of-being-able-to-stop-adf04f1917c2?source=rss-d500c312e4bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/adf04f1917c2</guid>
            <category><![CDATA[philosophy]]></category>
            <category><![CDATA[life]]></category>
            <category><![CDATA[creativity]]></category>
            <category><![CDATA[work]]></category>
            <category><![CDATA[work-life-balance]]></category>
            <dc:creator><![CDATA[Wale Ayandiran (walecloud)]]></dc:creator>
            <pubDate>Mon, 19 Oct 2026 06:01:01 GMT</pubDate>
            <atom:updated>2026-10-19T06:01:01.610Z</atom:updated>
            <content:encoded><![CDATA[<p>On Taking Breaks — And the strange privilege of being able to stop.</p><p>There is something mildly absurd about me writing about taking breaks. I am not particularly good at them.</p><p>In more than three decades of existence, I could probably count the genuine breaks I’ve taken. The rest have been pseudo-breaks at best. Time away from earning a living, perhaps, but not necessarily away from the things that make earning a living necessary.</p><p>And yet, I’ve increasingly come to appreciate what happens when we step away from our usual endeavours, even temporarily.</p><p>There is a certain clarity that comes with it. Things that previously seemed urgent become less so. Problems that occupied our minds for weeks begin to look different. Sometimes, the decision we had been struggling to make no longer seems particularly difficult.</p><p>But there is also the guilt that accompanies the decision to stop.</p><p>Am I resting, losing momentum, avoiding something, or quietly giving up? If I leave something alone for long enough, will I still want to return to it? And if circumstances force the break upon me, can I even call it a break?</p><p>I’ve been thinking about these questions, particularly the relationship between taking extended holidays and the important decisions we make in life.</p><p>Interestingly, the word <em>holiday</em> comes from the Old English <em>hāligdæg</em>, meaning a holy day. Originally associated with religious observances, its meaning eventually expanded to include time away from ordinary labour.</p><p>I find that origin rather fascinating. The idea that certain days could be set apart from the usual demands of existence.</p><p>Now, I’m particularly interested in breaks that last longer than a weekend or the occasional two-week holiday. The kind where you step away from earning a living for months, perhaps even a year, without necessarily knowing what you’re going to do with the time.</p><p>Something I’ve observed about people whose accomplishments we admire is that extended periods away from their usual endeavours sometimes appear in their stories before significant changes in their lives.</p><p>When I watched the <em>Jobs</em> movie, I became curious about Steve Jobs’ journey to India.</p><p>In 1974, before Apple existed, Jobs travelled to India with his friend Daniel Kottke. He spent several months exploring spirituality, visiting ashrams and trying to make sense of questions that had little to do with building computers.</p><p>It wasn’t a conventional holiday, but neither was it a professional undertaking.</p><p>Then there is Phil Knight.</p><p>In <em>Shoe Dog</em>, Knight recounts his journey around the world in 1962, visiting Japan, India and several other countries before Nike existed.</p><p>What makes his account especially interesting is that he already had the idea of importing Japanese running shoes before leaving America. His travels were partly an opportunity to explore that idea and partly an attempt to experience a world he knew relatively little about.</p><p>His father also helped finance the journey.</p><p>That last detail is important, and I’ll come back to it.</p><p>Both men eventually went on to build extraordinary things. It is tempting to look at their travels and conclude that stepping away from conventional work somehow prepared them for what followed.</p><p>I’ve certainly been tempted to make that connection.</p><p>But perhaps we’re giving these journeys importance because we already know how their stories ended.</p><p>If Jobs hadn’t founded Apple, would his months in India have meant anything to us? If Knight hadn’t built Nike, would his journey around the world be regarded as anything more than the adventure of a young man trying to find his place?</p><p>How many people have taken similar journeys, returned home, and simply continued with their lives?</p><p>And would that make their experiences any less valuable?</p><p>I don’t know.</p><p>There is some evidence that time away from work can improve aspects of cognitive flexibility, but that doesn’t necessarily explain why certain experiences appear to precede life-changing decisions.</p><p>It reminds me of something considerably more ordinary.</p><p>When we wake up after a good night’s sleep, we often find ourselves better equipped to think through things that seemed difficult the previous evening.</p><p>What exactly changes?</p><p>And if a night’s sleep can offer that kind of mental refreshment, what happens when we extend the distance between ourselves and our usual responsibilities?</p><p>Is there a threshold?</p><p>Would taking a nap every three hours improve our ability to think throughout the day? Or would our minds eventually adapt, making the supposed benefit less noticeable?</p><p>Perhaps an extended holiday offers something sleep alone cannot. The opportunity to experience life without constantly returning to the same obligations.</p><p>Or perhaps I’m confusing two fundamentally different things.</p><p>What I do find curious is that, whenever I manage to take a break from earning my keep in the committee of the human species, I tend to become more engaged with the things I enjoy doing for their own sake.</p><p>Running, writing, building software.</p><p>Things that could reasonably be described as work, except that I am not necessarily doing them because someone needs them done or because they must produce an economic outcome.</p><p>I’m doing them because I enjoy doing them.</p><p>It makes me wonder whether the clarity I associate with taking breaks comes from doing less or simply having greater freedom to decide what deserves my attention.</p><p>But this brings me back to the uncomfortable part of the whole observation.</p><p><strong>The ability to step away from earning a living without worrying about the consequences is a luxury that isn’t evenly distributed.</strong></p><p>Someone might argue that everyone can afford a holiday suited to their financial circumstances.</p><p>And I wouldn’t necessarily disagree.</p><p>But that isn’t quite what I’m talking about.</p><p>I’m talking about the possibility of taking an entire year away from earning a living without spending that year worrying about what happens afterwards.</p><p>What happens to your employment? Will the job still exist? Will the skills you’ve spent years acquiring remain relevant in a world where technological change, particularly AI, is already making people question the future of their work?</p><p>What happens to the responsibilities that do not disappear simply because you’ve decided you need some time away?</p><p>One could argue that a year is excessive.</p><p>Why not three months? Six months? Or even two?</p><p>But the question remains.</p><p>How much of that time would genuinely belong to you if you spent it worrying about what awaits you when you return?</p><p>And what about someone whose break wasn’t a choice at all?</p><p>Someone who lost their job, whose business failed, or whose circumstances simply made continuing impossible.</p><p>They might have an entire year away from work, but would they experience the clarity I’ve been describing?</p><p>There is a difference between having nothing to do and having the freedom to do nothing.</p><p>Perhaps that difference matters more than the duration of the break itself.</p><p>I’m also beginning to question my original assumption that vacations should lead to clarity or important decisions.</p><p><strong>Why must a break produce something valuable before we consider it worthwhile?</strong></p><p>The moment I describe extended holidays as a superpower for making better decisions, I’ve already found a way to turn them into another instrument of productivity.</p><p>Take time off, gain clarity, return with a better idea, make something extraordinary.</p><p>Even our attempts to escape the demands of work somehow end up being justified by what they might enable us to accomplish afterwards.</p><p>What if nothing extraordinary happens?</p><p>What if someone takes a year away and returns without a business idea, a new career, or some profound understanding of their existence?</p><p>Has that year been wasted?</p><p>I’m not entirely sure why we are so quick to measure time this way.</p><p>Perhaps it’s because many of us have spent most of our lives learning that our time must be accounted for, that there must be something to show for it.</p><p>And perhaps that’s also why taking a break comes with so much guilt.</p><p>As with most things I write about, I don’t have a solution to offer.</p><p>I can only reflect on my own experiences, the lives I’ve observed, and the questions they leave me with.</p><p>Personally, I hope to someday have the luxury of taking at least a year away from earning a living.</p><p>Not necessarily to discover the next big idea, reinvent myself, or return as a more productive person.</p><p>Just to experience what it means to have that much time without having to worry about the state of my existence afterwards.</p><p>I look forward to that possibility, hopefully in the not-so-distant future.</p><p>Although, there is one thing I haven’t quite figured out.</p><p>After spending so much of my life learning how to keep going, <strong>would I even know what to do with the freedom to stop?</strong></p><p>Here’s a picture I took while i was away at Leeds over the weekend after work.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*nD_9Ugd5iCpxZW97" /><figcaption>A picture taken away at Leeds</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=adf04f1917c2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The cost of being Human]]></title>
            <link>https://medium.com/@walecloud/the-cost-of-being-human-50e5b32411e1?source=rss-d500c312e4bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/50e5b32411e1</guid>
            <category><![CDATA[thoughts]]></category>
            <category><![CDATA[purpose]]></category>
            <category><![CDATA[creativity]]></category>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[life]]></category>
            <dc:creator><![CDATA[Wale Ayandiran (walecloud)]]></dc:creator>
            <pubDate>Tue, 13 Oct 2026 05:30:45 GMT</pubDate>
            <atom:updated>2026-10-13T08:25:49.161Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*K7N9DuyYoEOEIK4q" /></figure><p>Lets talk about work…Not necessarily the work we’re paid to do, but the work we would do if nobody paid us. The work that occupies our minds long after we’ve closed our laptops, left our offices, or finished whatever it is we do to earn a living.</p><p>The visible and invisible work. The passion projects. The things we once wanted to make, discover, build, or simply spend our lives doing.</p><p>And perhaps more importantly, the things that have disappeared entirely because life got in the way.</p><p>Thinking about the relationship between creativity and the means to pursue it.</p><p>For most people, indulging in a passion project requires funding. And funding, despite what the media and the more optimistic proponents of capitalism would have us believe, isn’t something one simply comes by.</p><p>Someone has to pay for it.</p><p>Sometimes that someone is you, through savings from a job you may or may not particularly enjoy. Sometimes it’s an institution, a wealthy individual, a government, or an organization that sees enough potential value in what you’re doing to justify supporting it.</p><p>I usually use examples like; for you to indulge in passion projects, it has to be funded. the fact that someone or some institution has to fund something you care about already defeats the purpose. If the funding gets pulled what happens to your project?</p><p>Does the work cease to be important? Does your curiosity suddenly become less valid? Or does it simply mean that whatever you were trying to bring into the world no longer satisfies someone else’s economic interests?</p><p>This is a complex topic with varying degrees of tangents in human evolution but for the sake of this short essay, lets talk about… <strong><em>How much human creativity have we lost by making the ability to create dependent on the ability to pay for it?</em></strong></p><p>Consider the Wright brothers.</p><p>Two men who became obsessed with the possibility of human flight. They weren’t employees of some aviation conglomerate commissioned to invent an aeroplane. They ran a bicycle business, and the income from that business helped finance their experiments.</p><p>Their curiosity preceded the commercial opportunity.</p><p>One could say something similar about countless artists, inventors, musicians, and thinkers throughout history. Although their circumstances differed considerably, many pursued ideas long before those ideas had any obvious commercial value.</p><p>Picasso didn’t need someone to establish a market opportunity for Cubism before experimenting with the form. That doesn’t mean he existed outside the economics of his time, or that patronage and money played no role in his career. They certainly did.</p><p>But there is an important difference between creating something because you believe it should exist and creating something because somebody has agreed to pay for it.</p><p>And increasingly, the former appears to require the latter.</p><p>The strange business of earning a living</p><p>Humans without “means” find it increasingly difficult to exist as humans.</p><p>That sentence is quite disturbing!</p><p>If someone isn’t paying you “wage/salary” or living off “inheritance”, you merely can’t live, its practically impossible.</p><p>If all you truly wanted to do was paint, or play musical instruments, or build sculptures or anything that you fell in love with as a kid and there’s something about putting colours on a canvas that makes being alive worthwhile to you.</p><p>Not become a famous painter. Not build a personal brand around your paintings. Not sell courses teaching people how to paint, or create content documenting your artistic journey.</p><p>But now you’re a grown human specie and can’t exist without “ <strong><em>earning a living</em></strong> “ which means unless your service is valued by someone, organizations, you’re doomed to not earn your keep as a member at the top of the food chain.</p><p>You need somewhere to live. You need food, electricity, healthcare, clothing, and perhaps a few other things that make existence bearable.</p><p>None of those things are particularly interested in your paintings.</p><p>Your landlord certainly isn’t.</p><p>How ironic don’t you think?</p><p>So you must find something people are willing to pay you for. And if painting isn’t one of those things, you must spend a considerable part of your life doing something else.</p><p>Your passion becomes something you indulge in after you’ve finished doing what society considers economically useful.</p><p>If you’re lucky, you have some energy left for it.</p><p>If you’re very lucky, you make enough money to buy back some of your time.</p><p>And if you’re extraordinarily lucky, somebody eventually decides your paintings are worth paying for.</p><p>We’ve somehow arrived at a point where being human requires earning the financial means to remain human.</p><p>We’ve evolved so much that we have to earn to be able to stay on earth, the alternative you would ask? well, be incarcerated or in an asylum as “we’ve” agreed we should have no true freedom of existence except you’ve amassed a great amount of human created material wealth.</p><p>We’re supposedly the most intelligent species on the planet. We’ve built machines that can traverse continents, instruments that can observe galaxies, and computers capable of performing extraordinary intellectual tasks.</p><p>We have become exceptionally good at improving the tools of human existence without necessarily improving the experience of being human.</p><p>Now, there’s another interesting dimension to this.</p><p>Or more precisely, the promise of wealth.</p><p>We’ve built an entire culture around the idea that financial freedom is available to anyone willing to work hard enough, take enough risks, or make enough sacrifices.</p><p>Work now so you can live later.</p><p>Build wealth so you can pursue your passions.</p><p>Sacrifice your twenties so you can enjoy your forties.</p><p>Sacrifice your forties so you can retire comfortably in your sixties.</p><p>There’s always another milestone between you and the life you actually want to live.</p><p>And while some people do manage to accumulate enough wealth to escape the necessity of working, <strong>that possibility is far from evenly distributed.</strong></p><p>The circumstances of one’s birth, access to education, geography, inherited wealth, social connections, health, and plain luck all influence how realistic that promise is.</p><p>Yet we continue to sell the possibility as though it were simply a matter of individual determination.</p><p>It’s a rather clever arrangement, isn’t it?</p><p><strong>We sell people the dream of eventually escaping the very conditions under which they must spend most of their lives.</strong></p><p>And the price of that dream is often more of their lives.</p><p>I don’t think there’s anything inherently wrong with wanting wealth. Nor do I think being paid for work somehow makes that work less meaningful.</p><p>Its deeper than that… the cost of being human is so high now more than it has ever been.</p><p>The painter must first become financially stable enough to paint.</p><p>The musician must first earn enough to afford the time to practice.</p><p>The inventor must somehow convince someone that an idea nobody has ever seen is worth financing.</p><p>And the thinker? Well, they’d better find a way to monetize those thoughts… Loool.</p><p>I wonder how many extraordinary things have never existed because the people capable of creating them were too busy earning a living.</p><p>How many remarkable musicians never picked up an instrument again after childhood?</p><p>How many potential breakthroughs died quietly in the minds of people who couldn’t afford to experiment?</p><p>How many artists abandoned their craft, not because they stopped loving it, but because the demands of existence left them with neither the time nor the energy?</p><p>We tend to celebrate the extraordinary individuals who managed to overcome these circumstances.</p><p>The stories are inspiring, certainly.</p><p>But they’re also convenient.</p><p>Because for every person who makes it through, there are countless others whose work we’ll never encounter.</p><p>Not because they lacked talent, ambition, or curiosity.</p><p>But because curiosity doesn’t pay the rent.</p><p>And perhaps that’s the part of human progress we don’t discuss enough.</p><p>What exactly have we been progressing towards?</p><p>Human beings have always had to work to survive. Our ancestors weren’t exactly spending their afternoons painting landscapes while someone else gathered food.</p><p>And modern society has made possible extraordinary freedoms that previous generations couldn’t have imagined.</p><p>But surely the purpose of progress cannot simply be to create more sophisticated ways of earning the right to exist.</p><p>At some point, shouldn’t advancement translate into more freedom to be human?</p><p>Not necessarily freedom from all responsibility or contribution, but freedom to contribute in ways that aren’t immediately reducible to economic value.</p><p>And somewhere along the way, the things we once wanted to do simply because we loved doing them became things we hoped to afford someday.</p><p><strong>The cost of being human has become extraordinarily high.</strong></p><p>Not just in money.</p><p>But in all the things we might have been, had we been free to find out and express.</p><p><em>Originally published at </em><a href="https://wales-note.beehiiv.com/p/the-cost-of-being-human"><em>https://wales-note.beehiiv.com</em></a><em> on October 13, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=50e5b32411e1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Experience Should Refine You, Not Rewrite You — Values & Growth]]></title>
            <link>https://medium.com/@walecloud/experience-should-refine-you-not-rewrite-you-values-growth-dd86b0c92d50?source=rss-d500c312e4bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/dd86b0c92d50</guid>
            <category><![CDATA[values]]></category>
            <category><![CDATA[experience]]></category>
            <category><![CDATA[growth]]></category>
            <category><![CDATA[life-lessons]]></category>
            <category><![CDATA[sacrifice]]></category>
            <dc:creator><![CDATA[Wale Ayandiran (walecloud)]]></dc:creator>
            <pubDate>Sun, 04 Oct 2026 11:29:18 GMT</pubDate>
            <atom:updated>2026-10-04T11:34:30.859Z</atom:updated>
            <content:encoded><![CDATA[<p>Certain experiences have more impact on how you want to interact with humans going forward. People usually term it as “lesson learnt”.</p><p><strong>But should this be a way to live?</strong></p><p>The short answer is no.</p><p>The slightly more elaborate answer is that you can’t live your life on the back foot of what happens to you as a result of other people’s actions.</p><p>This is close to something I find interesting about Adlerian psychology: the idea that what happened to you doesn’t have to become the reason you don’t become or do what you want. <em>We aren’t simply pushed forward by causes from our past. There is also the direction we choose to move in.</em></p><p>Otherwise, that just sounds like a terrible way to live.</p><p>Everyone should develop and own their core values. Values that serve as a harness for how you want the world to experience you and, equally, how you want to experience the world.</p><p>Your response to certain situations will differ from event to event, of course. But those responses should still run within the harness of your values so you don’t lose yourself while dealing with all the variants of human behaviour.</p><p>Someone takes advantage of your kindness, so you decide never to be that kind again.</p><p>Someone betrays your trust, so the next person now has to pay for what they did.</p><p>Someone doesn’t appreciate what you gave, so next time you decide to give less.</p><p>I don’t think that’s a great way to live.</p><p>You should learn from experience, definitely.</p><p>But experience should improve your judgment, not completely rewrite who you are.</p><p>Perspective is a funny thing</p><p>Imagine you have $100.</p><p>Someone asks you for financial support and you give them $50.</p><p>Another person has $1,000 and gives them $100.</p><p>From the receiver’s perspective, even if they knew you only had $100 yourself, they could still see the second person as having given more.</p><p>After all, $100 is more than $50.</p><p>But you gave them a whopping 50% of everything you had.</p><p>The other person gave 10%.</p><p>There is nothing wrong with the second giver.</p><p>There is also nothing necessarily wrong with the receiver.</p><p>Life is about perspective.</p><p>People experience what you give them from where they are standing, not necessarily from where you are standing.</p><p>They see the $50.</p><p>They may not fully understand what the $50 represented to you.</p><p>And even when they understand, they may still not appreciate it in the same way you expect them to.</p><p>Only people with a certain level of empathy or deeper conscience will sometimes look beyond the immediate gratification and ask what that thing actually cost the person who gave it.</p><p>That is why I don’t think the lesson should always be:</p><p>“I’ll never do that again.”</p><p>Maybe the better lesson is:</p><p>“I’ll be more intentional about who I do that for.”</p><p>There is a difference.</p><p>One changes your judgment.</p><p>The other changes your character.</p><p>Being good-hearted is not enough</p><p>Being good-hearted doesn’t cost much.</p><p>Share what you know.</p><p>Do things for people without turning every interaction into a transaction.</p><p>But being good-hearted doesn’t necessarily propel you at corporate work.</p><p>Unless everyone else is equally good-hearted, which sadly has never really been the case in human society, especially in group settings.</p><p>You can do a lot of invisible work.</p><p>You can help people.</p><p>You can unblock people.</p><p>You can prevent problems nobody will ever know were about to happen.</p><p>You can make other people better at what they do.</p><p>And still, somebody can turn around and ask:</p><p>“What exactly did you do?”</p><p>Not necessarily because they are bad people.</p><p>Sometimes, they simply didn’t see it.</p><p>People see what is visible to them.</p><p>So be good-hearted, but also make sure you have evidence of your work, your output and your outcomes.</p><p>Know what you did.</p><p>Know what changed because of it.</p><p>Know what you contributed.</p><p>Don’t leave all of that to somebody else’s memory or interpretation.</p><p>That isn’t being transactional.</p><p>That is just being sensible.</p><p>Don’t let people decide who you become</p><p>There will always be people who misunderstand what you gave.</p><p>People who don’t appreciate what it cost you.</p><p>People who take more than they return.</p><p>People who interpret kindness as weakness.</p><p>And there will also be people who completely understand it.</p><p>You can’t control which one you meet.</p><p>What you can control is who you become after meeting them.</p><p>Pay attention to patterns.</p><p>Be less naive where you need to be.</p><p>But don’t let every bad experience slowly turn you into somebody you don’t recognise.</p><p>There is a difference between learning a lesson and becoming a product of other people’s behaviour.</p><p>Experience should refine you.</p><p>It shouldn’t own you.</p><p>In another write-up, I’ll talk about human perspective, blind spots, their effects and how we can harness them better in our day-to-day lives.</p><p>For now, this is where I land:</p><p><strong>Learn from how people treat you, but don’t let how people treat you decide the kind of person you become.</strong></p><p><em>Living you with a Picture from 2022 on an evening walk in Zaandam Netherlands.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*TfNJEvz0BUBMpTw5" /></figure><p><em>Originally published at </em><a href="https://wales-note.beehiiv.com/p/experience-should-refine-you-not-rewrite-you"><em>https://wales-note.beehiiv.com</em></a><em> on October 4, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dd86b0c92d50" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I’ve Changed My Mind About Speed — A Technical Founder’s Advice]]></title>
            <link>https://medium.com/@walecloud/ive-changed-my-mind-about-speed-a-technical-founder-s-advice-757749695fdc?source=rss-d500c312e4bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/757749695fdc</guid>
            <category><![CDATA[product]]></category>
            <category><![CDATA[founders]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[Wale Ayandiran (walecloud)]]></dc:creator>
            <pubDate>Sun, 27 Sep 2026 07:56:51 GMT</pubDate>
            <atom:updated>2026-09-27T11:54:25.140Z</atom:updated>
            <content:encoded><![CDATA[<p>For a long time, I thought one of the biggest advantages of being a technical founder was simple:</p><p>If I had an idea, I could build it. I didn’t necessarily need to find an engineer, raise money to hire a team, write a detailed specification and wait several months to discover whether the thing in my head could actually work.</p><p>I could just start.</p><p>That still feels like an enormous advantage but I no longer think the real advantage is being able to build faster.</p><p>I think it is being able to <strong>learn faster</strong> , that sounds like a small distinction. I’m increasingly convinced it is a very important one.</p><p>AI changed the economics of building</p><p>The distance between an idea and a convincing piece of software has totally collapsed. An idea that might previously have required weeks or months of engineering can sometimes be explored in <strong>days</strong> or . <strong>hours</strong></p><p>AI can help write the code.</p><p>It can suggest the architecture.</p><p>It can build interfaces.</p><p>It can generate tests.</p><p>It can help debug problems, analyze data and write documentation.</p><p>Small teams can now attempt things that would previously have required significantly more people and money.</p><p>But it has created an interesting problem.</p><p>When building becomes cheaper, <strong>being able to build something becomes weaker evidence that it should be built.</strong></p><p>Almost every interesting thought can now be followed far enough to resemble a product of some sort and I think that makes judgment more important, not less.</p><p>The cost of being wrong is getting cheaper</p><p>The cost of being wrong is getting cheaper, it may sound entirely positive and mostly it is but I think it could also make us wrong more often than not.</p><p>When building something required six months, a team and a significant amount of capital, there was friction before you committed. That certainly didn’t prevent bad products anyway… we’ve spent decades proving that.</p><p>But today that friction is disappearing.</p><p>I can have a thought in the morning and have something fairly convincing running shortly afterwards. The temptation is to interpret that velocity as progress.</p><p>Integration.</p><p>But moving quickly and learning quickly aren’t necessarily the same thing.</p><p>You can build very quickly in completely the wrong direction.</p><p>I used to optimize for execution speed</p><p>This is probably where my thinking has changed the most… sike! actually no! it hasn’t but its refined in a different way.</p><p>I still strongly hold that adequate consistent motion is better than a well thought-out execution that comes sporadically at time-friction-cost of solution.</p><p>I used to see the ability to move quickly from idea to implementation as an advantage but now I think the better use of that ability is to shorten the distance between:</p><p><strong>assumption → experiment → evidence.</strong></p><p>Build enough to learn.</p><p>Not enough to prove how sophisticated the architecture could become.</p><p>Build the uncomfortable version that answers the question you’re avoiding.</p><p>Put it in front of somebody.</p><p>Watch what they actually do.</p><p>Then decide what deserves to happen next.</p><p>That requires a different relationship with the things you build.</p><p>You have to be willing to throw them away</p><p>And throwing software away can be particularly difficult when you’re the person who knows exactly how much thinking went into creating it.</p><p>Restraint might be the new leverage</p><p>As execution becomes cheaper, I’m beginning to think restraint becomes more valuable.</p><ul><li>Knowing what not to build. (No one ever does in the first try *or a veryyyy few have).</li><li>Knowing when a two-day integration isn’t worth two hours.</li><li>Knowing when someone asking for a feature is actually describing a different product.</li><li>Knowing when ten people saying “interesting” is weaker evidence than one person desperately trying to use something unfinished.</li><li>Knowing when you need another iteration.</li><li>And knowing when the evidence is telling you to stop.</li></ul><p>That last one is difficult for engineers.</p><p>Partly building companies.</p><p>Mostly the accumulation of watching the difference between things that are technically possible and things people genuinely care about.</p><p>I’m probably capable of building more today than at any other point in my career.</p><p>The tools are better.</p><p>My experience is deeper.</p><p>AI has dramatically increased what one person can attempt.</p><p>But I increasingly want to measure progress by something other than how much software I can produce.</p><p>The technical founder’s advantage might not be:</p><p><strong>I can build this faster than you.</strong></p><p><strong>I can find out whether this should exist faster than you.</strong></p><p>That means technical ability still matters enormously but the scarce skill moves somewhere else, it moves towards judgment.</p><p>Toward understanding people.</p><p>Toward asking better questions.</p><p>Toward recognizing weak signals before turning them into six months of work.</p><p>And toward having enough detachment from your own ability to build something to decide that you shouldn’t.</p><p>That’s what I’ve changed my mind about.</p><p>I’m just much less interested in the speed of building than I used to be.</p><p>I’m becoming far more interested in the speed of learning.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Fk0D4g95t-vOGaIr" /><figcaption>A picture I took during my run on Thursday looking across from the bay.</figcaption></figure><p><em>Originally published at </em><a href="https://wales-note.beehiiv.com/p/i-ve-changed-my-mind-about-speed"><em>https://wales-note.beehiiv.com</em></a><em> on September 27, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=757749695fdc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Be Less Impressed By What You Build — Technical Founder Advice]]></title>
            <link>https://medium.com/@walecloud/be-less-impressed-by-what-you-build-technical-founder-advice-cebad3bf2099?source=rss-d500c312e4bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/cebad3bf2099</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[technical-founder]]></category>
            <category><![CDATA[founders]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <dc:creator><![CDATA[Wale Ayandiran (walecloud)]]></dc:creator>
            <pubDate>Sun, 20 Sep 2026 15:45:51 GMT</pubDate>
            <atom:updated>2026-09-20T16:36:38.504Z</atom:updated>
            <content:encoded><![CDATA[<p>For most of my career and generally for any software engineer Pre-llm-coding-agents era, being able to build things felt like a superpower.</p><p>Give me a problem, enough time, and access to the right tools, and I could usually figure out a way to make something work.</p><p>That ability has served me well as an engineer. It has opened doors, helped me lead teams, build products, start companies and turn ideas that existed only in conversations into software people could actually use.</p><p>For a long time, I also thought it gives a significant advantage as a founder.</p><p>I’m beginning to think there is a trap in that.</p><p>There is satisfaction in taking something messy and turning it into something elegant, refined &amp; structured.</p><p>But the business of market solutions doesn’t particularly care about elegance.</p><p>A technically interesting problem may not necessarily be an important problem.</p><p>A difficult problem is not necessarily a valuable problem.</p><p>And a product working exactly as intended is not proof that enough people care about it or want it.</p><p>I knew all of this intellectually but building has forced me to understand it differently.</p><p>When someone tells an engineer that something is inconvenient, <strong>we can hear a product requirement.</strong></p><p>When a prospect tells a founder that something is inconvenient, I’ve learned that the more useful response might be:</p><p>How painful is it?</p><p>How often does it happen?</p><p>What are you doing about it today?</p><p>What does doing nothing cost you?</p><p>Who inside the company cares enough to fight for this to change?</p><p>Will they pay to make it go away?</p><p>Those questions are increasingly more interesting to me now than whether I know how to build the solution.</p><p>“That’s impressive” is not the same as “I need this”</p><p>This is another thing I think technical founders have to unlearn.</p><p>You spend days or weeks building something complicated. You show it to someone and hear:</p><p>“I can see how companies would use this.”</p><p>Especially when you know how much work sits behind what looks like a simple demo.</p><p>But I’ve become much more interested in what happens after the compliment.</p><p>Does the person immediately start describing where this would fit into their company?</p><p>Do they introduce someone else who needs to see it?</p><p>Do they ask how quickly they can start?</p><p>Do they ask about security or integration because they’re already imagining deployment?</p><p>Do they try to find budget?</p><p>Do they actually come back and use it?</p><p>Or do they tell you you’ve built something impressive and carry on with their day?</p><p>Those are very different signals.</p><p>There is always another feature.</p><p>Another architecture improvement.</p><p>Another model to experiment with.</p><p>Another clever way of solving the problem.</p><p>And because you’re building, it feels like you’re progressing.</p><p>Sometimes you’re just extremely busy avoiding a much more uncomfortable question:</p><p>That is the question I’m learning to sit with for longer.</p><p>The engineer in me still loves building.</p><p>I don’t think that will change.</p><p>But I’m trying to become less impressed by the fact that I can build something.</p><p>The more interesting questions now are:</p><p>Who urgently needs it?</p><p>What changes for them if it works?</p><p>And what evidence do I have beyond the fact that I managed to make it work?</p><p>Maybe technical ability is most useful when it gives you the freedom to explore a problem quickly.</p><p>But it becomes dangerous when it gives you endless reasons to avoid admitting that the problem might not matter enough.</p><p>So these days, when I build something and my first reaction is:</p><p>I’m trying to follow it with a harder question:</p><p>Till Next time, <em>here’s a picture of a Dragon built using Lego in Cardiff</em></p><figure><img alt="Welsh Dragon Lego built." src="https://cdn-images-1.medium.com/max/1024/0*g1sMIkvyOOVQV0Dg" /></figure><p><em>Originally published at </em><a href="https://wales-note.beehiiv.com/p/be-less-impressed-by-what-you-build"><em>https://wales-note.beehiiv.com</em></a><em> on September 20, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cebad3bf2099" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Single-Flight Token Refresh]]></title>
            <link>https://medium.com/@walecloud/single-flight-token-refresh-2575da7b473e?source=rss-d500c312e4bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/2575da7b473e</guid>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Wale Ayandiran (walecloud)]]></dc:creator>
            <pubDate>Tue, 01 Sep 2026 17:00:50 GMT</pubDate>
            <atom:updated>2026-09-02T09:19:39.100Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*nJFQv5-ELFO4HDbD" /><figcaption>Photo by <a href="https://unsplash.com/@scottsdalemint?utm_source=medium&amp;utm_medium=referral">Scottsdale Mint</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Four invariants define correct behaviour:</p><p>These invariants are more useful than comments about individual lines. They state what must remain true after the implementation changes.</p><p>Each tenant owns one entry:</p><p>The entries live in a map keyed by a validated tenant ID:</p><p>is an absolute timestamp. Compute it once from the lifetime returned by the authorisation server:</p><p>Do not hardcode the token lifetime. Providers can change it. A hardcoded thirty-minute lifetime paired with a fifteen-minute token creates a silent outage window.</p><p>refreshInFlight stores a promise rather than a boolean. A boolean only tells callers that work exists. A promise gives them the exact work to await.</p><p>A token can pass a local check and expire during serialisation, network transit, upstream queueing, or a retry. Test whether it will remain alive after a safety window:</p><p>The skew should exceed the p99 upstream latency plus the retry budget. A two-minute skew on a thirty-minute token gives up about 6.7% of its life to remove the rollover race.</p><p>Make the skew a parameter. A long operation can request more headroom, and tests can move the boundary without mutating module state.</p><p>The refresh path is small:</p><p>The fast path is a map lookup and timestamp comparison. It has no network call, promise allocation, or mutex.</p><p>The critical section is the check followed by the assignment:</p><p>In a single-threaded event loop, nothing else runs between those statements because there is no . The check and publication act as one uninterrupted operation.</p><p>Do not insert awaited work between them:</p><p>That one change lets concurrent callers pass the check before any caller publishes its promise. The token storm returns.</p><p>The promise is the coordination handle. The first caller creates it. Later callers await the same promise. Only callers on the refresh path touch it.</p><p>Without it, a failed refresh leaves a rejected promise in the entry. Every later caller awaits the same old rejection. One brief network error then lasts until the process restarts.</p><p>clears the handle on success and failure. The next request after a failure gets a clean attempt.</p><p>Do not add negative caching by default. A fixed failure window can lock out a tenant after a short partner blip. If failures cause harmful retry amplification, add explicit bounded backoff with metrics.</p><p>The fetch function resolves credentials for one tenant and stores the token only after a successful response:</p><p>Authentication failure is an expected operating condition. Returning lets the caller produce a deliberate instead of leaking an opaque .</p><p>Never log credentials or tokens. Log the tenant, credential source label, cache-hit state, response status, and outcome.</p><p>An in-memory map grows for every new key. Tenant identity must be validated against an allowlist before it reaches the cache.</p><p>If the key comes from an unchecked header, user ID, or customer value, an attacker can fill memory by sending unique values. Validate against a known tenant set or use a bounded least-recently-used cache.</p><p>Per-tenant state also contains failure. Bad credentials for tenant A must not stall or invalidate the token for tenant B.</p><p>Fetching a token when an instance starts can remove latency from its first request. It must remain optional:</p><p>Never make service readiness depend on a third party. If partner authentication is down during deployment, instances should still start and serve work that does not need that partner.</p><p>Use a delayed mock to keep the race window open:</p><p>The call-count assertion proves the invariant. Equal return values alone do not. Three separate refreshes can return the same mocked token.</p><p>An immediate mock may also hide a broken race through lucky scheduling. The delay makes the test meaningful.</p><p>Pin the clock with fake timers:</p><p>The last assertion is the purpose of the skew. The token is technically alive at minute twenty-nine, but it does not have enough life left for safe use.</p><p>Also test recovery after failure:</p><p>Removing the cleanup should make this test fail.</p><p>This design keeps one token per active tenant per service instance. Use fleet-wide coordination instead when:</p><p>In those cases, a shared cache and distributed lock may earn their cost. They also add a network hop, another outage source, secret storage risk, and an operational runbook.</p><p>The coordination cost should not exceed the value of the work being coordinated.</p><p>The implementation is small because it solves the correct problem. Token refresh looks like caching, but the dangerous edge is concurrency at expiry. Name that edge correctly, and the design becomes one shared promise, an early-expiry check, and careful cleanup.</p><p><em>Originally published at </em><a href="https://wales-note.beehiiv.com/p/single-flight-token-refresh"><em>https://wales-note.beehiiv.com</em></a><em> on September 1, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2575da7b473e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building a Multi-Tenant Integration Platform for Enterprise APIs]]></title>
            <link>https://medium.com/@walecloud/building-a-multi-tenant-integration-platform-for-enterprise-apis-b6a599ce67e6?source=rss-d500c312e4bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/b6a599ce67e6</guid>
            <category><![CDATA[engineering-leadership]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[Wale Ayandiran (walecloud)]]></dc:creator>
            <pubDate>Tue, 01 Sep 2026 04:00:25 GMT</pubDate>
            <atom:updated>2026-09-02T08:48:10.244Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*kaq__OzO70wuRfwp" /><figcaption>Photo by <a href="https://unsplash.com/@jasonthedesigner?utm_source=medium&amp;utm_medium=referral">Jason Richard</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>The obvious approach is to let each team integrate with the partner directly. That often looks faster at the start. Over time, it creates a familiar set of problems:</p><p>At that point, the question is no longer, “How should this team connect to the partner?”</p><p>The better question is, “What operating model should the organisation use for partner access?”</p><p>That change in framing led us to build a shared, multi-tenant integration platform.</p><p>Instead of asking each team to build and operate its own integration, we created one service that could support many consumers.</p><p>The design had a small set of principles:</p><p>A new team could now onboard to an existing capability. It did not need to create another integration stack.</p><p>The platform currently serves four teams across three business domains. We also designed the tenancy model so it could support external partner organisations in the future.</p><p>That future requirement mattered. If all tenants were internal teams, tenant identity could be treated as a reporting label. Once an external organisation can become a tenant, identity becomes part of the security boundary.</p><p>Credentials, tokens, logs, failures, and usage data must remain isolated. A problem affecting one tenant must not affect another.</p><p>Designing that boundary at the start added little cost. Adding it after the platform had grown would have required a major rewrite.</p><p>The most difficult part of “one integration, many tenants” was authentication.</p><p>Partner APIs usually issue access tokens with a fixed lifetime. The service caches a token and requests a new one when the old token expires.</p><p>That sounds like a simple caching problem. Under load, it becomes a concurrency problem.</p><p>Imagine that a token expires while hundreds of requests are being processed. Without coordination, every request can notice the expiry at the same time. Each request then attempts to refresh the token.</p><p>The result is a token refresh storm:</p><p>This issue is hard to see in low-volume testing. It appears at the expiry boundary, under real traffic, and then returns each time the token expires.</p><p>Multi-tenancy increases the risk. Each tenant has its own credentials and token lifetime. The platform is not managing one expiry cycle. It is managing several independent cycles.</p><p>We considered five approaches before choosing a design.</p><p>This is easy to implement, but it doubles partner traffic and adds authentication latency to every call. It also puts the most pressure on the endpoint most likely to have strict rate limits.</p><p>The cost grows directly with traffic. Success makes the design more expensive.</p><p>This performs well during normal operation. However, concurrent requests can all observe the same expired token and start separate refreshes.</p><p>It works most of the time, but fails at the exact moment coordination matters.</p><p>A background task can refresh tokens before they expire. This removes refresh latency from the request path.</p><p>It also refreshes tokens for idle tenants. In a scaled service, every running instance may refresh every tenant whether there is demand or not. Timers also add lifecycle concerns such as drift, shutdown handling, and cold starts.</p><p>A shared cache such as Redis can hold one token per tenant for the entire service fleet.</p><p>This is a reasonable enterprise pattern, but it has a cost. It adds a network call to the authentication path and creates another availability dependency. If the cache is unavailable, authentication may also become unavailable.</p><p>It also places bearer tokens from several tenants in shared storage. Concurrent service instances can still race during refresh, so a distributed lock is usually required as well.</p><p>A distributed design is justified when token lifetimes are very short, token issuance is metered, or a new token invalidates the old one. None of those conditions applied here.</p><p>The cost of coordination was greater than the value of the work being coordinated.</p><p><strong>Refresh after an authentication failure</strong></p><p>The service could wait for a , refresh the token, and retry the request.</p><p>This makes failure part of normal operation. At least one request fails at every rollover, and concurrent retries can create another storm.</p><p>We did not want to spend error budget on an event we could predict.</p><p>We combined an in-memory token cache with two controls:</p><ol><li><strong>Single-flight refresh:</strong> only one refresh operation can run for a tenant at a time.</li><li><strong>Expiry skew:</strong> the service treats a token as expired shortly before its actual expiry time.</li></ol><p>When several requests find a token near expiry, the first request starts the refresh. It publishes the in-progress operation so later requests can wait for the same result.</p><p>Exactly one refresh call is sent to the partner. Every waiting request receives the new token when that call completes.</p><p>The in-progress reference is cleared whether the refresh succeeds or fails. A temporary partner error therefore does not leave the tenant stuck in a failed state. The next request can try again.</p><p>This approach protects the refresh path without adding a lock to the normal path. A valid cached token requires only a map lookup and a timestamp comparison. There is no network call and no new infrastructure dependency.</p><p>A token can be valid when the service checks it and expire before it reaches the partner.</p><p>Between those two moments, the request may pass through serialisation, network transit, proxies, queues, and retries.</p><p>We therefore apply a two-minute safety window to a thirty-minute token. A token inside that window is treated as expired even though the partner may still accept it for a short time.</p><p>This uses about 93% of the token lifetime while removing the risk of sending a token that is about to expire.</p><p>The two controls solve different parts of the problem:</p><p>Skew without coordination would still allow many refresh calls. Coordination without skew could still send tokens that expire in transit.</p><p>The token cache is keyed by tenant. Each tenant has its own credentials, cached token, expiry time, and refresh operation.</p><p>This gives us several important properties.</p><p>The tenant key is validated before credential resolution or token access. It is never inferred from an untrusted request payload.</p><p>Not every team could complete partner credential setup on the same timeline. The platform therefore supported a temporary fallback to shared credentials.</p><p>The fallback allowed delivery to continue without waiting for the slowest external dependency. However, an invisible fallback would quickly become permanent debt.</p><p>Every fallback request emits a structured audit event. The event records the tenant, request ID, environment, operation, outcome, status, and an explicit fallback marker. It never records the secret or token.</p><p>This turns migration progress into measurable data. We can answer which tenants still use shared credentials and how much traffic they generate.</p><p>It also detects configuration drift. If fallback use rises for a tenant that had already migrated, we know its dedicated credentials may have disappeared during a deployment.</p><p>The platform also replaced an older SOAP integration that used hardcoded shared credentials and provided no tenant attribution.</p><p>We retired it through a gradual and reversible process.</p><p>First, legacy responses included standard deprecation information: a deprecation flag, a sunset date, and a link to the migration guide. Consumers could detect the change through the protocol instead of relying on an email.</p><p>Later, a configuration switch could return a clear “gone” response with the migration guide attached.</p><p>Both stages were environment-specific and reversible without a deployment. If a consumer experienced an unexpected problem, the change could be undone in seconds.</p><p>That reversibility made the migration safer. A risky change should be recoverable faster than it can be escalated.</p><p>The value went beyond token management.</p><p>The fourth team did not need a fourth authentication implementation, monitoring stack, or independent discovery of the token-storm problem. It onboarded to an existing platform capability.</p><p>The design is intentionally process-local. Each service instance can hold one token per active tenant.</p><p>We accepted that because token lifetimes are moderate, token issuance is not metered, and the partner allows more than one active token.</p><p>We would revisit the decision if:</p><p>Under those conditions, a shared cache and distributed lock could justify their added latency and operational cost.</p><p>Naming the conditions that would change a decision prevents architecture from becoming folklore. It gives future teams evidence for when to keep the design and when to replace it.</p><p>The token cache looked like a caching problem. The failure mode showed that it was a concurrency problem.</p><p>Once the problem was named correctly, the solution became small: one in-flight refresh per tenant, an early expiry window, and clean recovery after failure.</p><p>The wider platform followed the same principle. We did not solve four team-level integration problems. We created one operating model with shared governance, strong isolation, and lower marginal cost for each new consumer.</p><p>That is the difference between shipping an integration and building organisational leverage.</p><p>Good engineering judgement is not measured by how much infrastructure a design introduces. It is measured by whether the design solves the real problem, contains risk, and remains easy to change when its assumptions no longer hold.</p><p><em>Originally published at </em><a href="https://wales-note.beehiiv.com/p/building-a-multi-tenant-integration-platform-for-enterprise-apis"><em>https://wales-note.beehiiv.com</em></a><em> on September 1, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b6a599ce67e6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Origin Story — The Path]]></title>
            <link>https://medium.com/@walecloud/origin-story-the-path-f22fbcfeb437?source=rss-d500c312e4bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/f22fbcfeb437</guid>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[founders]]></category>
            <dc:creator><![CDATA[Wale Ayandiran (walecloud)]]></dc:creator>
            <pubDate>Mon, 04 May 2026 01:29:07 GMT</pubDate>
            <atom:updated>2026-05-04T01:29:07.613Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nd0vw7_alwnQ2HG2UPimuw.png" /><figcaption>OnBuddy origin</figcaption></figure><p>The startup journey of a startup (<em>pun intended</em>) is never straightforward or planned to the T. Things happen, and you happen to things, as you continue to build and learn. Our story at <a href="https://medium.com/u/8e2111580847">Proteus AI</a> is no different.</p><p><a href="https://medium.com/u/9f443f4eab96">Franklin Chieze</a> and I have been building together since 2018, back when we were both learning facilitators at <a href="https://medium.com/u/a4d81c24b23">Andela</a>, collaborating on engineering tools and training the next generation of African engineers. What bound us beyond the job description was a shared instinct: find the data, understand the relationships within it, and figure out how to make it useful in the real world.</p><p>One early proof of that instinct was <a href="https://github.com/andela/live-pulse/tree/client">Live Pulse</a>, an internal tool we built to monitor engineer and team performance, alert the right stakeholders when key numbers drifted past a pre-determined threshold, and surface negative trends before they hardened into negative incidents. It was a performance tracking system built for Andela’s specific context, but the thinking behind it, giving organizations a real-time brain for their operations, would quietly follow us for years without us realizing it.</p><p>Fast forward to 2023. Before the AI boom became impossible to ignore, Franklin and I started tinkering with LLM APIs, particularly OpenAI’s. I had just completed a self-sponsored MSc in Artificial Intelligence with a focus on deep learning, and we had enough curiosity and enough stubbornness to start something. We built <a href="https://proteusai.app/">proteusai.app</a> (<em>deprecated</em>) as a crude but genuine proof of concept: a platform that let people and businesses upload their own documents and put AI to work on real problems. It was early, it was imperfect, and it grew to just over a thousand free and paying users.</p><p>As time went by and AI advancement grew, we realized business were finding it difficult to truly harness its power in her day-to-day operations and providing a means to crunch documents alone wasn’t sufficient. Uploading documents and getting answers was a starting point, not a solution. The gap between what AI could do and what organizations actually needed was wider than a chatbot could bridge.</p><p>That led us to conducting Over 30 organizational interview sessions while we continued building, gave us a window into what businesses actually needed as AI evolved and moved. What we found consistently, and what surprised us in its simplicity, was that the loudest pain wasn’t in strategy or tooling choices. It was in operations, particular how teams work and onboarding. Teams drowning in repetitive, manual tasks. New hires left to figure things out alone, navigating a maze of forms, tools, and unwritten context with very little structured support. The chaotic work was not an edge case. It was the norm.</p><p>Having lived through versions of that chaos ourselves across multiple organizations, those interviews led us to progressively build OnBuddy from what started as a very first onboarding manager into what it is now; a full <strong>business brain operator</strong> for organizations, living inside the tools teams already use, Slack, Microsoft Teams, WhatsApp, and quietly handling the operational work that no one has time to do manually, Its been termed as the “Linear for work”.</p><p>But we could only connect that dots backward. When people ask what’s our origin story, I summarize it by saying we’ve always wanted to build a business brain for every organizational processes before we know what that was from Live pulse to Yutars, to Proteusai App and now OnBuddy. The vision as always been clear but the path to it is bound to be, and probably always will be, forever dynamic.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f22fbcfeb437" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>