<?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 Arnaldo De Moraes Pereira on Medium]]></title>
        <description><![CDATA[Stories by Arnaldo De Moraes Pereira on Medium]]></description>
        <link>https://medium.com/@arnaldopereira?source=rss-238ff14b7569------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*ulCvfwtuzjz0g2ehbtmNKg.png</url>
            <title>Stories by Arnaldo De Moraes Pereira on Medium</title>
            <link>https://medium.com/@arnaldopereira?source=rss-238ff14b7569------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 16:31:59 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@arnaldopereira/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[Leadership Anti-patterns: Don’t properly delegate]]></title>
            <link>https://medium.com/@arnaldopereira/leadership-anti-patterns-dont-properly-delegate-f8dedd2b83f0?source=rss-238ff14b7569------2</link>
            <guid isPermaLink="false">https://medium.com/p/f8dedd2b83f0</guid>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[dysfunction]]></category>
            <dc:creator><![CDATA[Arnaldo De Moraes Pereira]]></dc:creator>
            <pubDate>Fri, 12 May 2023 13:32:14 GMT</pubDate>
            <atom:updated>2023-05-12T13:32:14.078Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why</h3><p>These are some quick thoughts around delegating work.</p><p>Why even talk about this at all? We basically take it for granted, or even talk about this in a very simplistic, binary way: a person delegate their work or doesn’t, but what, how and when? That makes a difference in one’s work.</p><p>As in almost any leadership subject, we could overthink it, do a lot of bikeshedding and end up nowhere. We could also understand some dysfunctions and improve our work.</p><h3>What do delegate</h3><p>Usually this is easy to get wrong for new leaders. What kind of tasks should you delegate?</p><p>If you delegate only easy tasks, you’ll end up with the hard ones and won’t have time to manage your teams. At the same time, the exact opposite might not work, as well.</p><p>Focus on more strategic and work that scales over time seems to make sense here. At the same time, don’t be afraid to delegate other tasks when necessary — have in mind that every work is valuable.</p><h3>When to delegate</h3><p>Delegating something as fast as you can, only because you’re doing many things at once is doing the right thing at the wrong time. Who’s never heard the reasoning “you have too much in your plate”, instead of things like “you’ll be more effective”, “this process will flow more smoothly”, etc.?</p><p>The mindset change is important.</p><p>There’s what I see as almost a <strong>fire and forget</strong> pattern in leaders that have an almost desperate behavior: they delegate tasks as soon as they receive and almost forget about them, until close to the deadline.</p><p>On the other hand, there are the <strong>lazy loaders</strong>: they receive the task, but don’t delegate until close to the deadline.</p><h3>How to delegate</h3><p>It’s not only about that, but you also have to make sure you’re passing along something with a clear outcome. The project/task should be ready (ready is different than done), with a clear outcome and context, and your peer should also be ready to receive and work on it — if she’s busy or doesn’t have the necessary resources to do it, it’s a failure before the start and you, as the leader, have to able to predict that.</p><p>Let’s not forget that it’s all about communication, it’s our most essential way of exchanging information. So it’s crucial to make sure your peer understands what you’re asking of her. Some leaders usually talk in their own dialect and don’t generally care if they’re being understood, they just assume they are. Or ask simplistic questions such as “Got it? Cool.”</p><p>Mechanisms are crucial to this part. Amazon and AWS have been using the “<a href="https://www.allthingsdistributed.com/2006/11/working_backwards.html">Working backwards</a>” method to illustrate ideas about projects and features since ~2006 and it seems to be working nicely. I’ve used it in some companies and love it.</p><p>For me, the rule of thumb is: the more strategic your position is, the more you must invest in work that scales over time. This is a thought that everybody has to have to do good work, but some people just can’t have the luxury not have it, this mindset should be nonnegotiable for leaders.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f8dedd2b83f0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Startups anti-patterns]]></title>
            <link>https://medium.com/@arnaldopereira/startups-anti-patterns-30b5d52ac9fb?source=rss-238ff14b7569------2</link>
            <guid isPermaLink="false">https://medium.com/p/30b5d52ac9fb</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[startup-lessons]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Arnaldo De Moraes Pereira]]></dc:creator>
            <pubDate>Tue, 28 May 2019 18:16:54 GMT</pubDate>
            <atom:updated>2019-05-29T11:34:06.320Z</atom:updated>
            <content:encoded><![CDATA[<p>Startups always catches me, usually because I find them interesting, fun and challenging.</p><p>On one I&#39;ve seen, the definition of <strong>culture</strong> was a very long text glued at all urinals of the company (probably at the doors on the women restroom as well). On another, high management frequently spoken of the <strong>awesome culture</strong> the company had, but nobody actually knew anything about it until more than a third of the company was fired and everything abruptly changed. Startups.</p><p>Below are two anti-patterns I&#39;ve seen in more than one. There are some others I might post another time, such as <strong>the rich partner</strong>, <strong>mind the gap </strong>and<strong> corporate mindset</strong>.</p><p>Mind you that I consider these more dysfunctions than anti-patterns, but thankfully I googled &quot;startups dysfunctions&quot; and found out that there are already articles that are just and plainly copies of the &quot;five dysfunctions of a team&quot; renamed to &quot;five dysfunctions of startups&quot;. Which are still applicable anyway.</p><p>The <a href="https://www.amazon.com/Five-Dysfunctions-Team-Leadership-Fable/dp/0787960756">five dysfunctions of a team</a> is a very good book and describes simple dysfunctions that are present in most teams, you should read if you lead a team or plan to. Or even if aren’t a leader, so you can help identify and fix the team dynamic, where possible.</p><p>On the anti-patterns.</p><h3>Product for the founders</h3><p>Listen to users is one of the most important things to do and most companies fail to do it, even the ones that say to do it. Just saying you do something over and over won&#39;t magically turn that into a true fact, it&#39;s crazy some people don&#39;t seem to know that.</p><p>It&#39;s common to prioritize the development of features over other things, so implementing mechanisms to listen to users get down on the backlog. The UX, Product Managers, Product Owners, if they even exist on the company, don&#39;t get to actually own the products — they don&#39;t make decisions.</p><p>So the product ends up being created for the founders to use, not for the users. This is a very known mistake, but people are still spending time and money on it.</p><p>If you don&#39;t have mechanisms to see which feature is being used, quantitatively and qualitatively, how will you deprecate a feature? If you don&#39;t deprecate a feature and only write new ones, how will you maintain a whole bunch of features down the line?</p><h3>The watchdog</h3><p>He have the owners trust above everyone and everything else, even his skills and deliveries. He probably thinks he dominates some specific field, and that idea might even be fed by the owners themselves, but they&#39;re alone on that — the rest of the company thinks otherwise and don&#39;t actually respect him, they only respect the authority he represents.</p><p>The role of the watchdog is to be the ears and eyes of the owers, try and override teams decisions and report what&#39;s happening to owners. All that in a ghost position, so nobody actually understands what he does.</p><p>Some VCs apply this pattern professionally. Once a company I had got funded by a VC that had watchdogs, inside the VC they really used the name <strong>bulldog</strong> and applied the pattern throughout almost all the companies they funded.</p><p>Good points on that pattern within that specific VC: between the founders and the VC we kind of knew the bulldog role, his limits and so on — so the rules were clear. I left my own company less than a year after it got funded. Best decision ever.</p><p>Some startups are amateur on this: the founders hire their best friend forever for some random, but big, role. Chances are he was never on that role on any other company before, but he not only gets the job but also get into the company full of power to switch roles whenever whatever.</p><p>He never delivers what his roles demands, but it doesn&#39;t really matter. The actual employees don&#39;t get it, but what nobody gets is that his role is actually just to be a watchdog. Only it&#39;s implicit.</p><h3>So what?</h3><p>The first dysfunction, <strong>Product for the founders</strong>, doesn&#39;t make the startup a bad company to work for, per se. If the company have this dysfunction, it means the products it creates depends only on the vision of their founders — if they&#39;re remarkable geniuses, with visions that represents their users, or if they&#39;re lucky enough to not have competitors, etc.</p><p>It doesn&#39;t mean it does any good, in my opinion. But the second dysfunction is way worse.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=30b5d52ac9fb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Kubernetes and AWS Application Load Balancer]]></title>
            <link>https://medium.com/@arnaldopereira/kubernetes-and-aws-application-load-balancer-1b70bd2b2306?source=rss-238ff14b7569------2</link>
            <guid isPermaLink="false">https://medium.com/p/1b70bd2b2306</guid>
            <category><![CDATA[docker]]></category>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[kubernetes]]></category>
            <dc:creator><![CDATA[Arnaldo De Moraes Pereira]]></dc:creator>
            <pubDate>Wed, 08 Mar 2017 14:21:29 GMT</pubDate>
            <atom:updated>2018-08-02T16:53:18.091Z</atom:updated>
            <content:encoded><![CDATA[<p>Kubernetes and AWS Application Load Balancer</p><p><strong>Update</strong>: Two months after I wrote this post, Ticketmaster and CoreOS released the first implementation of ALB support for Kubernetes in the form of an ingress controller. This implementation is now maintainted by the SIG-AWS and is available in <a href="https://github.com/kubernetes-sigs/aws-alb-ingress-controller">https://github.com/kubernetes-sigs/aws-alb-ingress-controller</a></p><p>So right now, in August 2018, Kubernetes doesn&#39;t natively support AWS ALBs — the users have to install the ingress controller to have it running in their clusters, but it <strong>seems</strong> to be stable and ready to be used in production.</p><p>What follows is my original post about the reasoning on why ALBs are better than classic ELBs, specially when used with Kubernetes:</p><p>AWS has three types of load balancers:</p><ul><li>ELB: the classic Elastic Load Balancer</li><li>ALB: Application Load Balancer</li><li>NLB: Network Load Balancer (out of the scope of this post)</li></ul><p>ELB works at layer 4 of the OSI model, while ALB, as it&#39;s name implies, works at layer 7. We can look at ELB as a dumb service to access our applications: basically you map one internet-facing protocol and port with one instance protocol and port.</p><p>If you run containers inside that instance, you&#39;ll have to run a proxy inside it to properly forward the requests to them.</p><p>ALB has the ability to keep a map of one internet-facing port to many listening ports on the instance. That way your containers, or Kubernetes pods, can directly receive traffic from the ALB without any proxy doing port forward between them.</p><p>Kubernetes still doesn&#39;t natively support ALB: <a href="https://github.com/kubernetes/kubernetes/issues/30518">https://github.com/kubernetes/kubernetes/issues/30518</a> — if you&#39;re able to help getting it done or vote up for it, go ahead, that&#39;s gonna be useful in many ways.</p><p>Besides eliminating a hop, ALB has many other advantages over the classic ELB, as well pointed by this article: <a href="https://www.sumologic.com/aws/elb/aws-elastic-load-balancers-classic-vs-application/">https://www.sumologic.com/aws/elb/aws-elastic-load-balancers-classic-vs-application/</a></p><p>One of the really useful ones was released early december: WAF support. As we move more and more applications to Kubernetes that&#39;s gonna be very important, as for a few we use WAF and don&#39;t want to lose our set of rules and the feature itself.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1b70bd2b2306" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Full stack, full slack]]></title>
            <link>https://medium.com/@arnaldopereira/full-stack-full-slack-6166f4d582a5?source=rss-238ff14b7569------2</link>
            <guid isPermaLink="false">https://medium.com/p/6166f4d582a5</guid>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[tech]]></category>
            <dc:creator><![CDATA[Arnaldo De Moraes Pereira]]></dc:creator>
            <pubDate>Mon, 11 Jul 2016 21:28:10 GMT</pubDate>
            <atom:updated>2016-07-21T15:28:05.402Z</atom:updated>
            <content:encoded><![CDATA[<p>It seems there&#39;s a mutually exclusive law governing developers:</p><ul><li>If outstanding, can&#39;t be full stack.</li><li>If full stack, can&#39;t be outstanding.</li></ul><p>You can replace &quot;outstanding&quot; by &quot;great&quot; or even &quot;very good&quot;. The fact is that all the developers I&#39;ve met that call themselves &quot;full stack&quot; are average at best — and usually limited, which is actually the opposite of what you&#39;d expect from them.</p><p>The notion of a full stack developer is simply based on the use of a set of popular libraries and services that have big communities around and have a whole lot of documentation and are very easy to use.</p><p>Hence anyone would be a full stack anything if they choose to use them. So, that just doesn&#39;t make any sense.</p><h4>A truly full stack developer these days would know how to choose good cloud or service providers, would have very good relations with ops teams, would know how to automate ops works and, very importantly, would think and care about it constantly in his workflow as he cares about his code quality — as all of this has direct impact in his quality of life and ultimately his employer’s products success.</h4><p>And of course, a developer who knows one language knows none. But that&#39;s something nobody has to say, right?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6166f4d582a5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The startup dream]]></title>
            <link>https://medium.com/@arnaldopereira/the-startup-dream-1b31dbd19fc5?source=rss-238ff14b7569------2</link>
            <guid isPermaLink="false">https://medium.com/p/1b31dbd19fc5</guid>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <dc:creator><![CDATA[Arnaldo De Moraes Pereira]]></dc:creator>
            <pubDate>Wed, 06 Jul 2016 18:48:04 GMT</pubDate>
            <atom:updated>2023-05-11T14:39:51.966Z</atom:updated>
            <content:encoded><![CDATA[<p>The first result for a google search for &quot;the startup dream&quot; is a post titled &quot;<a href="https://medium.com/swlh/how-quitting-my-corporate-job-for-my-startup-dream-f-cked-my-life-up-3b6b3e29b318#.9n6nwwget">How quitting my corporate job for my startup dream f*cked my life up</a>&quot; — and that&#39;s not a coincidence.</p><p>Today we already have many startup stories to show us that we cannot simply just leave our jobs and have a startup and be happy and rich.</p><p>Yesterday I gave mentorship in technology for entepreneurs seeking funding from FIESP (Federation of Industries of the State of São Paulo) in an event called CAF, the biggest angel investment event in latin america, for the second time.</p><p>I&#39;ve spent (invested?) 10 years of my career in startups and in all that time I&#39;ve known only one type of startup: the silicon valley one, the one we dream about.</p><p>Lets pause and go back in time a little.</p><p>From 14 to 18 years old I&#39;ve worked with electronics in the car industry, where I had to work from 8am to 5:45pm with nonsense strict rules and poor work conditions. We couldn&#39;t talk much to each other, work seated, or read anything to seek knowledge — because if you get smarter you could get your boss&#39; job, right? That was the thought back then, back there.</p><p>I left the car industry and electronics for good when I was 18 and switched to computing programming, networking administration, then VoIP, Wifi, startups, cloud and so on.</p><p>I&#39;ll get to the point. I&#39;ve never seriously associated companies that don&#39;t have the internet as the main focus with the term <strong>startup</strong>. Specially if the company&#39;s focus is on building industrial machines and stuff — if that&#39;s the case I see it as far away as it can be from a startup. That&#39;s a big mistake, as I&#39;ve learned.</p><p>In my 10 years in startups I&#39;ve seen way more entrepreneurs that built the companies just because they wanted to have them than to actually solve problems. Most of them just don&#39;t admit that and end up with zombie startups, but that&#39;s just the truth.</p><p>Yesterday I&#39;ve had a great surprise to see people outsite the startup dream<em> </em>very well prepared to solve real problems. Don&#39;t get me wrong: the startup dream exist and it does solve real problems, but it also bias a whole lot of naive people.</p><p>My perception is that the enterpreneurs I&#39;ve saw yesterday care less about the sillicon valley and more about doing it right. And that&#39;ll make all the difference — and may eventually ironically get them to sillicon valley, for the right reasons.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1b31dbd19fc5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My git sed]]></title>
            <link>https://medium.com/@arnaldopereira/my-git-sed-a7ed7c69ddce?source=rss-238ff14b7569------2</link>
            <guid isPermaLink="false">https://medium.com/p/a7ed7c69ddce</guid>
            <category><![CDATA[git]]></category>
            <category><![CDATA[vim]]></category>
            <category><![CDATA[development]]></category>
            <dc:creator><![CDATA[Arnaldo De Moraes Pereira]]></dc:creator>
            <pubDate>Sun, 28 Feb 2016 22:10:47 GMT</pubDate>
            <atom:updated>2016-02-29T14:32:19.577Z</atom:updated>
            <cc:license>http://creativecommons.org/publicdomain/zero/1.0/</cc:license>
            <content:encoded><![CDATA[<p>My git sed</p><p>My equivalent of a `git sed` is a combination of `git grep`, bash and VIM. To replace `sudo` by `become` in a git repository, for example:</p><pre>vim -c ‘argdo %s/sudo/become/g | update’ -c q $(git grep sudo | cut -d: -f1 | sort -u)</pre><p>Roughly explaining the commands, backwards:</p><pre>git grep sudo | cut -d: -f1</pre><p>Returns the files that matches the grep for `sudo`. The file name will be printed as many times as it matches the search, such as:</p><pre>roles/app/tasks/main.yml<br>roles/app/tasks/main.yml<br>roles/app/tasks/main.yml<br>roles/redis/tasks/main.yml<br>roles/redis/tasks/main.yml</pre><p>To generate a unique file list that matches the grep for `sudo`, we add the `sort` command, having:</p><pre>git grep sudo | cut -d: -f1 | sort -u</pre><p>That cuts the initial file list to something as:</p><pre>roles/app/tasks/main.yml<br>roles/redis/tasks/main.yml</pre><p>This is the file list we want to process.</p><p>On VIM: it also works non-interactively, by processing commands received in the `-c` option, so we just have to write the search/replace command and pass the file list to it:</p><pre>vim -c ‘argdo %s/sudo/become/g | update’ -c q $(git grep sudo | cut -d: -f1 | sort -u)</pre><p>As VIM opens multiple files with this command, there are some things to notice here:</p><ul><li>We must use `argdo` to have the command runned on all files in the argument list.</li><li>`update` saves the buffer (file) to disk after it&#39;s changed. It is necessary to do that before switching to the other file as VIM will stick to the current buffer if it has unwritten changes.</li></ul><p>`q` of course only makes VIM exit.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a7ed7c69ddce" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>