<?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 Andrei Simion on Medium]]></title>
        <description><![CDATA[Stories by Andrei Simion on Medium]]></description>
        <link>https://medium.com/@andreiashu?source=rss-4e06f8b7393d------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*EqBZp0IAmIbYFOE0E9yMjw.jpeg</url>
            <title>Stories by Andrei Simion on Medium</title>
            <link>https://medium.com/@andreiashu?source=rss-4e06f8b7393d------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 22:47:41 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@andreiashu/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[What is the future of Intelligent Virtual Assistants?]]></title>
            <link>https://medium.com/@andreiashu/what-is-the-future-of-intelligent-virtual-assistants-9a90e8594298?source=rss-4e06f8b7393d------2</link>
            <guid isPermaLink="false">https://medium.com/p/9a90e8594298</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[data]]></category>
            <category><![CDATA[privacy]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Andrei Simion]]></dc:creator>
            <pubDate>Fri, 19 Jul 2019 10:22:34 GMT</pubDate>
            <atom:updated>2019-07-19T10:23:52.755Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*vL6mbpU5WIUCtkc-" /><figcaption>Photo by <a href="https://unsplash.com/@dariusbashar?utm_source=medium&amp;utm_medium=referral">Darius Bashar</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><blockquote>Alexa, Siri, Cortana and Google Assistant are examples of Intelligent Virtual Assistants (IVAs). In this article I explore the idea that a focus on user privacy could enable such tools to progress from simple task managers (shopping, weather, play specific song, schedule a meeting) to software agents that can provide meaningful insight into our lives based on our individual behaviour.</blockquote><h3>Context</h3><p>We are at a tipping point where privacy and interoperability are crucial for the development of the future IVAs.</p><p>The current tools cannot make full use of the vast personal data that can be available to them because there is close to 0 interoperability between them. Additionally the potential to capture more data is limited by a lack of trust that users have in the way the owners of these tools are handling their data.</p><h3>Why</h3><p>Do you know how your meals in the past week affected your mood? Do you even know what, when, where and with whom you ate this past week? Why would that be of any use you ask?</p><p>Whether it’s food and drinks that we put in our body, or time spent with friends or colleagues at work or the environment we spend time in, everything affects our health, mood and how we ultimately relate to the outside world.</p><p>If you knew that taking the bus to work instead of the underground makes you an overall more calm person would you do it? Would you still do this even though you knew it’ll increase your commute time by 20 mins daily? What if using the bus you are more likely to engage in talking with people during your commute or read a book? And that on days when you commute by bus, your productivity at work and general mood improves?</p><p>Just the awareness of certain patterns in your life means you will more likely change undesired behaviours and reinforce desired ones. When you can correlate perceived state of mind and results to habits and environmental influence, improvements in your life can happen organically.</p><p>If you find the idea of tracking everything is your life creepy then you are correct. With most of our current technology tracking your life IS CREEPY. Because your data is currently accessible by the owners of the service. The biggest offenders are big tech like Google, Facebook, Amazon, Apple although this is an industry wide practice / business model.</p><p>The good news is that this is slowly starting to change. Businesses (especially the biggest offenders in the space) are starting to change the way they record, store and use (cough… sell cough…) your data.</p><h3>Obstacles</h3><p>Some people are already tracking their life to a certain degree(*). The manual work required to do this is currently one of the barriers to mainstream adoption.</p><p>Another issue is the lack of a clear business model that would support development in this direction. Currently our data is being used as the gas that fuels the advertising engine of big tech companies. For example Facebook and Google are highly dependant on this relationship between your data, their advertising algorithm accuracy and their revenue. Without your data their business model wouldn’t make sense (no data → low ads and search results accuracy → low revenue → need to charge usage based fees → incentive for users to look for alternative, free, products).</p><h3>A brave new world?</h3><p>What if you can be guaranteed that you’re the only one that has access to your private data? That it cannot be accessed without your consent? Would recording most things in your life still be creepy if there’s a promise of valuable insights in such exercise? An Intelligent Virtual Assistant (that cannot share any of your data with the outside world) can provide valuable insights into your life.</p><p>I wouldn’t take an IVA long before it can surface the fact that more than a certain amount of time spent on social media daily will affect everything from productivity at work to how we interact with our friends and family. There’s a continuous feedback loop between how we interact with the world and how later the world “responds” back to us.</p><p>Are the new business models of the world going to encourage true privacy of data and interoperability of apps in order to allow their users to trust their products enough?</p><p>I believe that once a company cracks the business model challenge then the rest will adopt it soon after.</p><p>When enough incentives are in place for businesses to focus on this model of doing business then products that take better advantage of the much more personal and in-depth data will become available. Users will follow since the benefits of this type of IVAs will increase significantly.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ZyysMjt3OKf4pJr1" /><figcaption>Photo by <a href="https://unsplash.com/@franki?utm_source=medium&amp;utm_medium=referral">Franki Chamaki</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>More to come</h3><p>The dynamics between privacy, interoperability and business model innovation in this area is complex but deeply interesting to me. This article only focused on a few ideas in the space, mainly related to how privacy might enable IVAs to have access to a much larger personal dataset. Armed with this, future IVAs can be more standalone: no need for them to share much if any data with their creators or any outside party.</p><p>Things that I’d like to cover in future writing are more in-depth thoughts about how might apps share data with the IVAs, interoperability (between apps themselves or through the IVA), incentive systems, and potential business models.</p><p>I’m curious to hear your thoughts in this space. What do you think the future of personal data and IVAs is? Are business models that create privacy focused products a pipe dream or the inevitable next step we’re heading towards?</p><p><em>(*) “I Tracked Every Single Thing I Ate for a Full Year. Here are 10 Quick A-Ha’s.” by Mike Bechtel: </em><a href="https://medium.com/@mikebechtel/i-tracked-every-single-thing-i-ate-for-a-full-year-here-are-10-quick-a-has-2c259dd2f57b"><em>https://medium.com/@mikebechtel/i-tracked-every-single-thing-i-ate-for-a-full-year-here-are-10-quick-a-has-2c259dd2f57b</em></a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9a90e8594298" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Fun First, Friction Later (if at all): why you should delay the blockchain experience in your app]]></title>
            <link>https://medium.com/meme-wars/fun-first-friction-later-if-at-all-why-you-should-delay-the-blockchain-experience-in-your-app-b46866cdecc9?source=rss-4e06f8b7393d------2</link>
            <guid isPermaLink="false">https://medium.com/p/b46866cdecc9</guid>
            <category><![CDATA[user-experience-design]]></category>
            <category><![CDATA[memes]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[games]]></category>
            <category><![CDATA[user-research]]></category>
            <dc:creator><![CDATA[Andrei Simion]]></dc:creator>
            <pubDate>Tue, 08 Jan 2019 19:24:34 GMT</pubDate>
            <atom:updated>2019-01-08T19:24:34.317Z</atom:updated>
            <content:encoded><![CDATA[<h3>Fun First, Friction Later (if at all): Why you should delay the blockchain experience in your app</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/540/1*BK3hp7b5dne0hogmCP8mxA.jpeg" /></figure><p>by <a href="https://twitter.com/pricegumb">Tyler Gumb</a> and <a href="https://twitter.com/andreiashu">Andrei Simion</a></p><h3>Show me value first</h3><p>We are at the beginning of the Web 3.0 revolution. As it is the early days of this revolution there is a lot of friction around onboarding. This friction has likely been one of the main factors holding dapps back from achieving a scale typical of web 2.0 apps.</p><p>We are currently working on a mobile gaming dapp called <a href="http://www.memewars.me/">Meme Wars</a>. In testing an initial prototype with an audience of “non-crypto-enthusiasts,” it became clear that requiring users have a wallet already set up and have some crypto to start engaging with the dapp provides way too much friction. We believe this is in line with expectations set by Web 2.0 apps, where users are accustomed to fun first, friction later. Whether that friction be time or money, users expect the value of your app will be established before they invest any of their valuable time or money. Currently setting up Metamask and learning your way around is a non-trivial investment of time and brainpower for people that are new to the space and have come to use your dapp not learn about blockchain.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6PbixNwhY0puMjRJnKa4ww.png" /><figcaption>Sign up for our newsletter at <a href="http://www.memewars.me/">memewars.me</a></figcaption></figure><p>Based on some of our user testing, we’d like to argue that a better way is to provide a frictionless Web 2.0 version of your product that establishes value and fun for new users unfamiliar with blockchain. Delay the blockchain onboarding as long as possible and only when necessary introduce it. At this point if your dapp has been well-designed and truly serves a need you will have established that value to users and they will be more likely to go through the friction of onboarding and unfamiliar blockchain based concepts like “wallet”, “seed”, “there’s no password reset option”, “you’ll lose your money if you lose this password.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/458/1*lO3us_8a64Vz0_kmY_dKKw.jpeg" /></figure><h3>How it would work in practice</h3><p>Imagine your app (ingeniously named Dhappee after weeks of user research by your UX team) enables users to play board games online with their friends. A unique selling proposition for your product that differentiates it among others is that people can stake tokens on each games which are transferred to the winning party. And because said tokens are publicly tradable and can be converted to fiat money it might make the game more interesting for some people.</p><p>In an ideal Web 3.0 world, such users would already actively manage their identity and different product personas on a blockchain based solution and would just use that to access your product and manage those juicy tokens. But we’re not there yet. So one alternative that we think would be useful is to push the blockchain part of your product as late as possible in the experience of the game.</p><p>Imagine the following user onboarding flow (which we aptly call ahem… WEB 2.5):</p><ul><li>user registers and gets an account on your app</li><li>user gets by default a bunch of 100 game tokens to play with their friends</li><li>when the free tokens run out then the user enters the “blockchain” onboarding flow where they have to download a wallet (or maybe use their existing one via Metamask) and top-up their account with new tokens (which might involve an exchange or a friend sending them some tokens)</li><li>at any point during the game a user can transfer their tokens by jumping on the blockchain onboarding flow</li></ul><p>What we do here is provide value up-front to justify the friction of setting up the blockchain aspects of the experience. Our approach onboards users to the blockchain in a gentler fashion. Forcing users to grapple with alien blockchain concepts in onboarding is a recipe for a high bounce rate.</p><h3>Transitioning to Decentralized</h3><p>Oftentimes there are core mechanics that can be accurately simulated using Web 2.0 technology. Refining these core mechanics are critical to the success of the product. Our approach is allowing us to focus our energy on the core features of Meme Wars. We are extremely excited about how Web 3.0 technology will augment Meme Wars and add value for users, but we also think the initial core mechanics of the game can be introduced, tested and refined with Web 2.0 technology.</p><p>“But this is not a decentralized app!” I hear you say. Well, maybe it shouldn’t be at this stage. We believe the best route for a new game like ours is to leverage the strong points of familiar centralized tech available now. Once that is established build in the transactional and self-sovereignty value that the blockchain provides today. We expect in some near future we will be able to port such an app to a fully Web 3.0 world.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b46866cdecc9" width="1" height="1" alt=""><hr><p><a href="https://medium.com/meme-wars/fun-first-friction-later-if-at-all-why-you-should-delay-the-blockchain-experience-in-your-app-b46866cdecc9">Fun First, Friction Later (if at all): why you should delay the blockchain experience in your app</a> was originally published in <a href="https://medium.com/meme-wars">Meme Wars</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[TIL: you can actually get a useful stacktrace on unhandled Promise rejections!]]></title>
            <link>https://medium.com/@andreiashu/til-you-can-actually-get-a-useful-stacktrace-on-unhandled-promise-rejections-af7099008c33?source=rss-4e06f8b7393d------2</link>
            <guid isPermaLink="false">https://medium.com/p/af7099008c33</guid>
            <category><![CDATA[promises]]></category>
            <category><![CDATA[error-handling]]></category>
            <category><![CDATA[nodejs]]></category>
            <category><![CDATA[bluebird]]></category>
            <dc:creator><![CDATA[Andrei Simion]]></dc:creator>
            <pubDate>Wed, 24 May 2017 22:28:56 GMT</pubDate>
            <atom:updated>2017-05-24T22:28:56.418Z</atom:updated>
            <content:encoded><![CDATA[<p>TIL: you can actually get a useful stacktrace on unhandled Promise rejections!</p><p>For some reason, at the time of writing, nodejs (v7.10.0) does not bail out when an unhandled promise rejection happens. It outputs a warning message but that’s it. I totally dislike the idea of my script just happily continuing with</p><blockquote>(node:7758) UnhandledPromiseRejectionWarning: Unhandled promise rejection (rejection id: 1): Error: […]</blockquote><blockquote>(node:7758) DeprecationWarning: Unhandled promise rejections are deprecated. In the future, promise rejections that are not handled will terminate the Node.js process with a non-zero exit code.</blockquote><p>To be fair it does say that in the future it’ll work in a more sane fashion. So in order to fast forward into the future you can use <a href="http://bluebirdjs.com/docs/getting-started.html">Bluebird</a> to sort this mess out:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/138fad21a4d753c502a33040bf344868/href">https://medium.com/media/138fad21a4d753c502a33040bf344868/href</a></iframe><p>Now you should get a nice stacktrace when you forget to handle rejections:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VLN5Nfm8kKHUu8ZJ0eEg8w.png" /><figcaption>Lovely stacktrace isn’t it?</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=af7099008c33" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Docker, AWS ECS optimized instances and Disk space]]></title>
            <link>https://medium.com/simi-says/docker-aws-ecs-optimized-instances-and-disk-space-9cbc42cd433c?source=rss-4e06f8b7393d------2</link>
            <guid isPermaLink="false">https://medium.com/p/9cbc42cd433c</guid>
            <category><![CDATA[aws]]></category>
            <category><![CDATA[docker]]></category>
            <category><![CDATA[ces]]></category>
            <dc:creator><![CDATA[Andrei Simion]]></dc:creator>
            <pubDate>Fri, 14 Apr 2017 13:19:16 GMT</pubDate>
            <atom:updated>2017-04-24T07:43:22.211Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-A6_VSKiS27O9ptC2hWZMg.png" /><figcaption>that time when prod goes down</figcaption></figure><p>On a recent project we started working with Docker, AWS ECS (container orchestration from AWS) and a few other buzzwords. Needless to say that we ran into several interesting issues.</p><p>One such issue only showed up in production recently: Docker will not release disk space to the host instance when is de-allocated (read rm -f) from within the container.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/541/1*g33I-cXS_bb5L4jdhrpSsQ.gif" /><figcaption>You see you see!?!</figcaption></figure><p>In order to fix this you either stop and remove the container (which is not an option in most cases) or use fstrim in a cron job on the host:</p><blockquote>docker ps -qa | xargs docker inspect --format=’{{ .State.Pid }}’ | xargs -IZ fstrim /proc/Z/root/</blockquote><p>This is a docker issue rather than an AWS one. It might have been fixed in newer versions of Docker but until AWS upgrades their ECS Optimised AMIs (we are using the latest one which at the moment is <strong>amzn-ami-2016.09.g</strong>) the workaround above should suffice.</p><p>Another thing to note is that <em>df -h</em> on both the host and container will show plenty of disk space left even though in reality your docker containers might be out of space. This is because the ECS optimized AMI has a separate LVM partition for the docker volumes and associated metadata. The side effect of this is that your EC2 monitoring tools will not give you an alert for low disk space.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9cbc42cd433c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/simi-says/docker-aws-ecs-optimized-instances-and-disk-space-9cbc42cd433c">Docker, AWS ECS optimized instances and Disk space</a> was originally published in <a href="https://medium.com/simi-says">simi.says</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>