<?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[Jose Duarte - Medium]]></title>
        <description><![CDATA[Digital Marketing | Brand Strategy | Product Management - Medium]]></description>
        <link>https://medium.com/jose-duarte?source=rss----4d3434a1bc45---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Jose Duarte - Medium</title>
            <link>https://medium.com/jose-duarte?source=rss----4d3434a1bc45---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 08:10:38 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/jose-duarte" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[I Just Spoke an App Into Existence With Claude Code]]></title>
            <link>https://medium.com/jose-duarte/i-think-i-accidentally-built-the-first-of-a-new-kind-of-app-the-kind-you-speak-into-existence-cc8398539756?source=rss----4d3434a1bc45---4</link>
            <guid isPermaLink="false">https://medium.com/p/cc8398539756</guid>
            <category><![CDATA[claude-code]]></category>
            <category><![CDATA[ios-app-development]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai-code-generation]]></category>
            <category><![CDATA[vibe-coding]]></category>
            <dc:creator><![CDATA[Jose Duarte]]></dc:creator>
            <pubDate>Wed, 28 Jan 2026 11:13:34 GMT</pubDate>
            <atom:updated>2026-01-28T11:13:33.119Z</atom:updated>
            <content:encoded><![CDATA[<p>Last Friday I spent 35 minutes describing a travel app to an AI. Then I went on a Zoom call. When I looked up, the app was running on my phone.</p><p>Then it warned me about a scheduling conflict.</p><p>I never wrote conflict detection code.</p><p>I hadn’t even told it that “move” meant changing a time.</p><p>The AI figured that out from the data structure alone.</p><h3>What Happened</h3><p><strong>Friday, January 23, ~8:15 AM</strong>: I’ve been playing around with Ralph Wiggum loops — an agentic coding technique where you let Claude run in a loop until a task is done. There’s a companion plugin called Lisa that handles the planning phase. Lisa interviews you until she thinks she has enough to write your spec. I thought the Simpsons pairing was cute, so I’d been using them together.</p><p>I start a Lisa session. I want a native SwiftUI travel app for my Japan trip. She asks questions: What’s the navigation structure? What data do you need? How should the AI modify the itinerary? I answer for about 35 minutes.</p><p><strong>~9:00 AM</strong>: My user research call was supposed to start. No-show. I’m chatting with coworkers on Zoom while the Ralph loops run in the background. I’m half-watching, hitting Enter to approve build steps.</p><p><strong>~9:05 AM</strong>: The call participant never joins. I look up.</p><p>The app is done. Not feature one. Not the schema. All three features I’d planned for the week. Done. While I was distracted.</p><p>I deploy to TestFlight. It works.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FRgqEq_fX_ok%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fshorts%2FRgqEq_fX_ok&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FRgqEq_fX_ok%2Fhq2.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="640" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/368b656f2c1a970e99dae40c61bf90d1/href">https://medium.com/media/368b656f2c1a970e99dae40c61bf90d1/href</a></iframe><p>I immediately put in my Claude API key and type “Move ghibli museum back 2pm” into the chat.</p><p>The app shows a confirmation dialog warning me about a conflicting event.</p><p>I had not programmed time-conflict detection.</p><p>I went back to the system prompt. Here’s the section that tells the AI how to respond to modification requests:</p><pre>{<br>  &quot;type&quot;: &quot;modify&quot; | &quot;add&quot; | &quot;delete&quot;,<br>  &quot;changes&quot;: {<br>    &quot;time&quot;: &quot;&lt;new time or null&gt;&quot;,<br>    &quot;place&quot;: &quot;&lt;new place or null&gt;&quot;<br>  },<br>  &quot;warnings&quot;: [&quot;&lt;any concerns about the request&gt;&quot;]<br>}</pre><p>That’s it. No explanation of what “move” means. No conflict detection function. Just: you can modify things, changes include time, you can include warnings about “any concerns.”</p><p>The AI understood:</p><ol><li>“Move” = type: &quot;modify&quot;</li><li>“back 2pm” = changes: { time: &quot;2:00 PM&quot; }</li><li>This would conflict with another event</li><li>That conflict was worth warning about</li></ol><p>From nothing but property names and a warnings array.</p><h3>The Apple Accident</h3><p>I’m not a developer. I’m a product marketer. But I’ve used Apple products my whole life, and I have opinions about what makes software feel right — even if I couldn’t explain why technically.</p><p>When I finally had a good reason to try using Claude to make an iPhone app, I had this instruction I kept giving to the Lisa loop: <strong>“Just make the Apple choice.”</strong></p><p>I didn’t know what I was asking for technically. I just knew Apple makes decisions that feel obvious in retrospect. Clean property names. Clear object boundaries. Things named what they are.</p><pre>// Not this<br>class ReservationDataTransferObject {<br>    var rsrvId: String<br>    var rsrvType: Int<br>}<br><br>// This<br>struct Reservation {<br>    var id: UUID<br>    var type: ReservationType<br>}</pre><p>“Make the Apple choice” meant: name things clearly, don’t abbreviate, use enums instead of magic numbers.</p><p>I thought I was asking for aesthetic decisions.</p><p>Turns out I was accidentally encoding domain vocabulary in a way AI can read.</p><p>Every time I said “make the Apple choice,” I was unknowingly saying “make the objects self-describing.” And self-describing objects are AI-readable objects.</p><p>The AI didn’t need a detectConflict() function because it could read what a ScheduleItem with a time property means. It understood the domain from the vocabulary.</p><p>I was aiming for Apple aesthetics. I hit AI legibility by accident.</p><h3>What I Think Is Actually Going On</h3><p><em>Note: This is what I’ve been able to reason out of Claude while trying to figure out what happened. I’m still just a marketer.</em></p><p>Traditional software architecture says: extract logic OUT of objects into service layers and utilities. Separation of concerns.</p><p>But AI doesn’t read service layers. AI reads schemas.</p><p>If your schema is just dumb data containers with abbreviated property names and magic numbers, AI has nothing to reason about. It needs to be told everything explicitly.</p><p>But if your objects describe themselves — clear names, domain vocabulary, properties that explain what they are — AI can reason about what operations make sense.</p><pre>struct Ad {<br>    var spend: Decimal<br>    var conversions: Int<br><br>    var costPerConversion: Decimal {<br>        spend / Decimal(conversions)<br>    }<br>}</pre><p>The logic isn’t hidden in a MetricsCalculator utility. It&#39;s on the object, in domain vocabulary. The AI can read it and understand the relationships.</p><p>When data describes itself clearly enough, AI can reason about it without being told how.</p><p>The conflict detection is the proof. I never told the AI what a scheduling conflict was. It inferred the concept from ScheduleItem objects with time properties.</p><h3>The Modal That Made Me Think This Generalizes</h3><p>During the Lisa interview, I said I wanted a modal that would confirm any changes before they happened. AI suggests changes, you approve or reject.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/404/1*_1HT1Ur0PDi9mDF_FO8NkQ.png" /></figure><p>I expected I’d still need to write logic telling the AI when to use it. shouldAskForConfirmation(), that kind of thing.</p><p>I didn’t write any of that. The AI used the modal correctly from the first build.</p><p>It understood from the object structure that changing a checkout time is the kind of operation that should be presented for review. No explicit instruction.</p><p>That’s when I started thinking this wasn’t specific to travel apps.</p><h3>The Weekend Tests</h3><p>I spent Saturday and Sunday figuring out if this was a fluke. I built three demos testing the same hypothesis: can AI reason about operations from schema alone?</p><p><strong>Test 1: Role-Based Access</strong></p><p>An AI canvas with roles: Artist, Curator, Viewer. Simple structure:</p><pre>enum Access {<br>    case heldTool    // Can use without asking<br>    case lockedTool  // Requires permission<br>}</pre><p>Artist can draw (held) but not delete (locked). I told the Artist to delete.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pTgWtOwDbhGnwzfpjB8uwQ.png" /></figure><p>The AI refused:</p><blockquote><em>“I cannot delete content from the canvas. As an artist, I can only add to the canvas by drawing with colors.”</em></blockquote><p>No conditional logic. The fence exists as data. The AI read it and reasoned about it.</p><p><strong>Test 2: Emergent Conflict Detection</strong></p><p>A calendar with a dead simple schema:</p><pre>struct ScheduleItem {<br>    var title: String<br>    var time: Date<br>}</pre><p>I asked it to schedule two things at the same time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JIilRzxp1fATvU1Rgo_3gQ.png" /><figcaption><em>“I notice there’s already a ‘Code Review’ scheduled at 2:00 PM, so adding another activity at that time would create a conflict.”</em></figcaption></figure><p>No detectConflict() function. The AI computed &quot;conflict = two things at same time&quot; from the schema.</p><p><strong>Test 3: Business Logic as Plain Text</strong></p><p>I defined business logic as a formula: CAC = spend ÷ conversions</p><p>Not as code. As a string.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-PvJtpW-hmrAIOcKrcthfw.png" /></figure><p>The AI reads it and calculates. If I change the formula, it calculates differently. No recompilation. The formula IS the logic.</p><h3>What This Might Mean</h3><p>I don’t know if this scales. But if it does, I think I accidentally stumbled into a new kind of app — the kind you speak into existence, because AI agents can understand your intent well enough to assemble the software pieces.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cc8398539756" width="1" height="1" alt=""><hr><p><a href="https://medium.com/jose-duarte/i-think-i-accidentally-built-the-first-of-a-new-kind-of-app-the-kind-you-speak-into-existence-cc8398539756">I Just Spoke an App Into Existence With Claude Code</a> was originally published in <a href="https://medium.com/jose-duarte">Jose Duarte</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How do you solve a problem like El Salvador?]]></title>
            <link>https://medium.com/jose-duarte/how-do-you-solve-a-problem-like-el-salvador-4dc3e1dc407f?source=rss----4d3434a1bc45---4</link>
            <guid isPermaLink="false">https://medium.com/p/4dc3e1dc407f</guid>
            <dc:creator><![CDATA[Jose Duarte]]></dc:creator>
            <pubDate>Sun, 11 Nov 2018 04:07:38 GMT</pubDate>
            <atom:updated>2018-11-16T17:57:15.852Z</atom:updated>
            <content:encoded><![CDATA[<p>Perspectives from a Salvadoran abroad.</p><p>I was visiting Switzerland with family about 10 years ago, when we drove past a large plot full of shacks. They looked a lot like the “zonas marginales” (slums) that have peppered every nook and cranny of San Salvador since our civil war in the 1980s. So much so, that one of my family members remarked, “You can understand why these people don’t like immigrants coming over. They come and ruin their cities like this.”</p><p>I’d later find out that this was actually a city farm, and it looked dirty because of, you know, the dirt.</p><p>At the time I didn’t know what to say. I chalked it up to her being from an older generation. Born in 1986, I was technically born during El Salvador’s bloody civil war, but young (and fortunate) enough not to remember many details. As a child, I mostly remember growing up in one of El Salvador’s rare peaceful boom periods, right after a hard fought peace, but before years of corruption and neglect returned El Salvador to its status as one of the most violent countries in the world.</p><p>Ask the average American about El Salvador today and they most likely will talk about caravans and “<a href="http://time.com/5100058/donald-trump-shithole-countries/"><em>shitholes</em></a><em>”</em>, or MS-13. Hopefully a few will tell you about our beautiful beaches and friendly people. While gang violence and migration are huge problems that need to be addressed in North and Central America, they are merely symptoms of problems that were never fully addressed after the civil war ended.</p><p>And if you’re an American, those root problems would scare you too. This is a bit oversimplified, but stop me if this sounds familiar:</p><p>For years, the 1% corrupted and manipulated the government to do whatever they wanted: prevented competitors from entering their markets, squashed labor movements, collected zero taxes, and suppressed votes to remain in power. They bought and enabled politicians to meet their needs, and they weakened or outright destroyed institutions to ensure good people trying to play by the rules were silenced or disillusioned.</p><p>The biggest difference is that the Salvadoran 1% was much more brutal and effective at achieving their goals.</p><h3><strong>Solving the Root Issues</strong></h3><p>ICE has a yearly budget of $5.7 billion. That’s about $200 million more than the ~entire~ Salvadoran general budget. The US spends more money trying to keep Salvadorans (and others) out of their country as El Salvador does on its people.</p><p>That also means that if the US diverted even a fraction of its ICE budget into a better and more efficient Foreign Aid program to countries like El Salvador, it could have a huge impact on living conditions for millions of people. People who wouldn’t feel the need to leave their families and everything they know to escape death by violence — economic, physical, sexual.</p><p>But you can’t just throw money at the problem either. There’s a pernicious culture of corruption that has persisted through governments left and right, moderate and extreme. Foreign donations routinely find themselves lining the pockets of the same people who perpetrate the cycle of violence.</p><h3><strong>The Necessary Elements</strong></h3><p>The US needs to rethink its approach to foreign aid if it truly wants to solve the immigration issue. Here’s a few broad points on how a better program could be structured:</p><p><strong>Direct Aid.</strong> For any program to be successful, it needs to have direct foreign oversight and execution. This will be a bitter pill to swallow in both countries: Americans will think it unnecessarily increases overhead costs, and Salvadorans will see it as an intrusion in sovereignty.</p><p>But it is absolutely necessary for any aid program to succeed. The Salvadoran government has proven itself completely untrustworthy to manage any kind of financial aid. El Salvador has had five former presidents since the 1992 peace accords: 2 two were jailed for corruption schemes in the 100s of millions of dollars, and a third fled the country under similar suspicions.</p><p><strong>Top down culture change. </strong>That direct aid should have an expiration date, though. As a condition to enter this program, the Salvadoran government should commit to transparency and anti-corruption. An International Commission Against Impunity like the one that has worked well for Guatemala should be instituted and protected.</p><p><strong>A bottom up approach. </strong>The focus of any program should be at the ground level: education, small and medium business growth, security and affordable housing.</p><p><strong>Strategic. </strong>Many Salvadorans in-country yearn for the days of old when El Salvador got most of its income from Coffee exports. The reality is that in a new world economy, El Salvador can’t rely on Agriculture, much less on one crop, to sustain its economy.</p><p>As the Salvadoran diaspora grows stateside, opportunities for tourism, e-commerce, customer service, investment and more open up if given the right room to grow. There are about 1.5 million Salvadorans in the US, and about 7 million in the El Salvador. Remittances amount to about 15% of GDP, but is often spent on bare necessities or luxury goods. Salvadorans outside El Salvador need better channels to maximize their investment in their country and their families.</p><p>The US can’t solve all of El Salvador’s problems, but it would be a benefit to help the country get out of its rut. With few exceptions, most Salvadorans don’t want to leave El Salvador, and they hate the MS-13 too.</p><p>The hysteria around immigration and caravans can impede fruitful discussion. And the lack of basic investment in the American people makes it hard to advocate for that same aid to be given elsewhere.</p><p>I think that proves the point more than work against it. Americans and Salvadorans are more alike than they know. And it’s time for politicians and the upper class of both countries to realize that investing in its people is the best way for all parties to prosper in peace.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4dc3e1dc407f" width="1" height="1" alt=""><hr><p><a href="https://medium.com/jose-duarte/how-do-you-solve-a-problem-like-el-salvador-4dc3e1dc407f">How do you solve a problem like El Salvador?</a> was originally published in <a href="https://medium.com/jose-duarte">Jose Duarte</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Making Equality Go Viral]]></title>
            <link>https://medium.com/jose-duarte/making-equality-go-viral-e89a5a59ea5a?source=rss----4d3434a1bc45---4</link>
            <guid isPermaLink="false">https://medium.com/p/e89a5a59ea5a</guid>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[love]]></category>
            <dc:creator><![CDATA[Jose Duarte]]></dc:creator>
            <pubDate>Wed, 28 Mar 2018 16:41:43 GMT</pubDate>
            <atom:updated>2018-04-04T15:05:26.014Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*8UnvEz3Gwr7T9mxQnHLKYQ.jpeg" /></figure><p>There’s no question Latin American countries are well on their way to recognize and celebrate LGBT equality. <a href="https://www.fundacionigualitos.org/">Fundación Igualitxs</a> was founded in 2016 with the intent to educate and fight for equal rights in Central America. I have been fortunate enough to consult in social media and branding strategies for this important work.</p><p>The foundation has achieved important legal milestones, including a constitutional lawsuit filed in Supreme Courts across Central America. However, their education initiatives are slowly raising awareness and opening hearts and minds.</p><p>Costa Rica held its presidential election on February 4, 2018. However, since neither candidate achieved 50% of the vote, a second round was called for April 1st. LGBTI issues quickly became <a href="https://www.reuters.com/article/us-costarica-election/costa-rica-center-left-easily-wins-presidency-in-vote-fought-on-gay-rights-idUSKBN1H80XC">the main issue</a> on a campaign between a pro-LGBTI center left candidate and a conservative Evangelical preacher.</p><p>On March 2018, Igualitos launched a video campaign titled “¿Quién es el Oscar en tu Familia?” (“Who’s the Oscar in your family”). The video showcases a family’s path to loving, accepting and embracing Oscar, the brother, son, uncle who happens to be gay. The goal of the campaign was to humanize LGBTI issues for Costa Ricans through the relatable video.</p><p>Using some of my video optimization techniques, Fundación Igualitos managed to create a video that hooked viewers, generated high levels of positive sentiment and achieved significant viral reach. Later, I helped Igualitxs pour fire on that gasoline through smart budgeting and retargeting. The results? Increased watch time and video completion, and insane reach: almost 500,000 Costa Ricans (10% of the population!) have seen the video on their Facebook Feeds.</p><p>On April 1st, Costa Ricans went to the polls and elected the center-left candidate, ensuring that marriage equality would be implemented in the country. The Oscar campaign managed to not just move hearts and minds, but popular opinion as well.</p><p>Two key takeaways from this campaign:</p><ul><li>Don’t create just one version of your video. Igualitxs created videos with different hooks up front, eventually striking gold with a variant that started with an emotional bang as opposed to a slow but steady setup.</li><li>Virality is organic, but smart marketers can help foster it. We maximized the promotional budget by creating smart audiences that targeted the video to people likely to watch or share it.</li></ul><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FS0JGZxvDmDE%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DS0JGZxvDmDE&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FS0JGZxvDmDE%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/9dfaf109042f30fa2e2fed43d5d6efce/href">https://medium.com/media/9dfaf109042f30fa2e2fed43d5d6efce/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e89a5a59ea5a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/jose-duarte/making-equality-go-viral-e89a5a59ea5a">Making Equality Go Viral</a> was originally published in <a href="https://medium.com/jose-duarte">Jose Duarte</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The 23 Questions Facebook Needs to Answer About The $100,000 Russian Ads]]></title>
            <link>https://medium.com/jose-duarte/the-23-questions-facebook-needs-to-answer-about-those-100k-russian-ads-e6f404c7dbe6?source=rss----4d3434a1bc45---4</link>
            <guid isPermaLink="false">https://medium.com/p/e6f404c7dbe6</guid>
            <category><![CDATA[cybersecurity]]></category>
            <category><![CDATA[russian-hacking]]></category>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[politics]]></category>
            <category><![CDATA[facebook]]></category>
            <dc:creator><![CDATA[Jose Duarte]]></dc:creator>
            <pubDate>Fri, 08 Sep 2017 20:27:49 GMT</pubDate>
            <atom:updated>2017-09-08T20:28:11.964Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JAqQvZbSO1ph-NQQuK6Nlw.jpeg" /></figure><p>Facebook is getting a lot of flack for their recent disclosure that a Russian company bought <a href="http://money.cnn.com/2017/09/07/media/what-russian-troll-army-could-buy-facebook-ads/index.html">$100,000 worth of Facebook ads</a> related to the 2016 presidential election.</p><p>The biggest criticism is that their disclosure seemed incomplete, insencere and opened more questions than it answered. Other than the dollar amount and the number of ads (~3,300), Facebook hasn’t shared much else.</p><p>As many savvy Social Media buyers know, $100,000 can get you really far on Facebook if you know what you’re doing. Dennis Yu, cofounder of Blitzmetrics (a Facebook ad agency) told <a href="http://www.thedailybeast.com/russias-facebook-fake-news-could-have-reached-70-million-americans">The Daily Beast</a> he estimated those ads helped Russians reach over <strong>70 million Americans. </strong>My own experience buying thousands of dollars in ads from Facebook agrees with Yu’s assessment.</p><p>The truth is, knowing Russians bought 3,300 ads and paid $100,000 to place them on Facebook doesn’t really tell us much about the <em>impact </em>these ads had. The true impact has to consider how the Facebook Algorithm helped amplify and further distribute these fake news. If you need a primer on how the Facebook algorithm works, <a href="https://medium.com/jose-duarte/why-facebooks-fake-news-button-won-t-help-against-russia-it-s-all-about-the-algorithm-10cb8799c4c5">read this</a>.</p><h4>With that in mind, here are the questions Facebook still needs to answer:</h4><ol><li>What exactly were the 3,300 ads?</li><li>How many Americans did each of these Ads reach?</li><li>How many of them went ‘viral’ and were distributed for free outside of the ad spend?</li><li>How many Americans were re-targeted with subsequent fake ads after the first time?</li><li>How many ads did the average person see?</li><li>What about in battleground states?</li><li>What was the geographic distribution of impressions? How many people did these ads reach in each state / battleground state?</li><li>What role did fake profiles play in artificially increasing the reach of these ads?</li><li>What were the targeting settings for the ads?</li><li>Did these ads use any kind of personal data in a customer file to micro-target voters?</li><li>If personal data / customer files were used, did they match efforts made by other ad accounts?</li><li>Were stolen voter rolls uploaded to Facebook in order to target specific voters in specific states?</li><li>Were there any patterns to the ad spend?</li><li>Were there surges in spending around times when Donald Trump was receiving bad press?</li><li>Were there other ad accounts that showed similar spending patterns?</li><li>Who were the administrators of the ad accounts and pages?</li><li>Were any Americans involved in placing or managing these ads?</li><li>How many fake accounts did the Russians use to place these buys?</li><li>How much did the algorithm favor these ads as a result of fake interactions by fake accounts?</li><li>How many people reported the ads as fake?</li><li>Were those reports investigated?</li><li>Are there any other advertisers still promoting links to these problematic websites?</li><li>What was the <strong><em>organic </em></strong>reach of the pages associated with these accounts?</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e6f404c7dbe6" width="1" height="1" alt=""><hr><p><a href="https://medium.com/jose-duarte/the-23-questions-facebook-needs-to-answer-about-those-100k-russian-ads-e6f404c7dbe6">The 23 Questions Facebook Needs to Answer About The $100,000 Russian Ads</a> was originally published in <a href="https://medium.com/jose-duarte">Jose Duarte</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Open the Safe! ARKit Mixed Reality Puzzle]]></title>
            <link>https://medium.com/jose-duarte/qrsafe-arkit-mixed-reality-puzzle-d07098475adb?source=rss----4d3434a1bc45---4</link>
            <guid isPermaLink="false">https://medium.com/p/d07098475adb</guid>
            <category><![CDATA[ios]]></category>
            <category><![CDATA[arkit]]></category>
            <category><![CDATA[apple]]></category>
            <category><![CDATA[augmented-reality]]></category>
            <category><![CDATA[qr-code]]></category>
            <dc:creator><![CDATA[Jose Duarte]]></dc:creator>
            <pubDate>Wed, 30 Aug 2017 20:43:49 GMT</pubDate>
            <atom:updated>2017-08-30T22:24:54.353Z</atom:updated>
            <content:encoded><![CDATA[<p>We’ve been following AR Kit closely at <a href="http://dose.com/">Dose</a> Labs; and we’ve been experimenting with a few applications for our websites &amp; clients. One of the AR applications that excites us the most is mixed reality content. That is, AR content that is affected by things that happen in the real world, not just on your screen.</p><p>Enter QRSafe:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F3IK2dJdEUPU%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D3IK2dJdEUPU&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F3IK2dJdEUPU%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/aed96daa26b18fff29ec8d67dfd09d63/href">https://medium.com/media/aed96daa26b18fff29ec8d67dfd09d63/href</a></iframe><p>The premise is pretty simple: complete the puzzle, scan it, and try to open the safe. For simplicity, we used a QR code cut in half. However, more elaborate puzzles are possible. A more advanced version of this app would use machine learning to identify objects, doing away with QR codes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pb6wWe--_ui0iKxR7faFjw.png" /><figcaption>Mixed Reality Fun</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d07098475adb" width="1" height="1" alt=""><hr><p><a href="https://medium.com/jose-duarte/qrsafe-arkit-mixed-reality-puzzle-d07098475adb">Open the Safe! ARKit Mixed Reality Puzzle</a> was originally published in <a href="https://medium.com/jose-duarte">Jose Duarte</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Desire Path Design — Twitter Threads]]></title>
            <link>https://medium.com/jose-duarte/desire-path-design-twitter-threads-aa9606ca714c?source=rss----4d3434a1bc45---4</link>
            <guid isPermaLink="false">https://medium.com/p/aa9606ca714c</guid>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[twitter]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Jose Duarte]]></dc:creator>
            <pubDate>Fri, 14 Jul 2017 16:10:24 GMT</pubDate>
            <atom:updated>2017-07-14T16:10:24.326Z</atom:updated>
            <content:encoded><![CDATA[<p><em>This article is part of a series on </em><a href="https://medium.com/@joseduarte/desire-path-product-design-2e1aacaac5b4"><em>Desire Path Product Design</em></a><em>. If you’ve ever created a Twitter thread, check out </em><a href="http://tweethreader.com/"><em>TweetThreader.com</em></a><em> for the easiest way to create twitter threads.</em></p><p>Tweet Threads — or tweetstorms — have defined Twitter since the practice <a href="http://www.npr.org/sections/alltechconsidered/2017/04/05/521835755/what-should-we-make-of-the-tweetstorm-or-thread-or-whatever-you-call-it">became popular</a> in 2014–15, to the point where Twitter trademarked the term. This article advocates that Twitter look at this feature as a “desire path” feature that its power users have popularized; as well as offer some feature solutions for how Twitter should cater to the users who want to create this vital part of the platform.</p><p>The Twitter Thread was popularized by <a href="http://www.npr.org/sections/alltechconsidered/2017/04/05/521835755/what-should-we-make-of-the-tweetstorm-or-thread-or-whatever-you-call-it">Marc Andreseen and Chris Dixon</a>, and they originally started as a simple way to get around the 140 character limit and share longer thoughts. If you run out of space, all you need to do is reply to your own tweet and it will automatically create a thread, making it easier to read and allowing users to share longer thoughts:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wt4g45OIosc0pcjPWSZ3vg.png" /><figcaption>If you didn’t know Twitter Threads by another name, you might recognize the numbered tweets, a staple of the format. <a href="https://twitter.com/joshtpm/status/874070385595219968">Josh Marshall</a></figcaption></figure><p>Twitter users quickly turned Threads from a simple hack to tweet longer into a real medium. Threads can be suspensful and enthralling, like in the viral thread from 2015, <a href="http://www.complex.com/pop-culture/2015/10/zola-twitter-insane-epic-story">Zola’s story</a>; reporters are able to report news with more nuance without making people click out of Twitter; and threads proved great comedic timing to elevate jokes.</p><p>Twitter Threads are a typical Desire Path feature. Users are driving the product forward by finding new and innovative ways to use the medium; and Twitter has created features <em>in response to the </em>the way people are using threads as opposed to proactively creating innovative thread features.</p><p>Much like hashtags, threads first emerged as a hack by power users, until Twitter finally built features into the system</p><p>The way Threads developed is reminiscent of another Desire Path feature: the hashtag. It was proposed by Google emplooyee <a href="https://en.wikipedia.org/wiki/Hashtag">Chris Messina</a> and while Twitter originally rejected the hasthag, the feature gained popularity among users until it became one of Twitter’s defining features.</p><p>Twitter has added some features to make <em>reading </em>threads a little bit easier, by connecting them with a bright blue line and making sure when you click into a threaded tweet, you see the oldest tweet on top, the complete inverse of the usual reverse chronological order.</p><p>Creating threads, however, is one of the glaring missing spots in this feature. The ‘correct’ way to create a thread is to reply to your own tweet; but not everyone is exactly clear on <em>how </em>they’re supposed to do it. Do you reply to the original tweet? Each subsequent one? How are you supposed to number them? Some people just send out 10 loose tweets and it’s impossible to read.</p><p>Enter Tweet Threader:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7k5J3pqZ8jY2FMmdHpeqOQ.png" /></figure><p>Tweet Threader was created with the intent of being the easiest way to create Twitter Threads. Just like in twitter, threads are created tweet by tweet; but Tweet Threader takes away all the guesswork and just lets you write.</p><p>Check it out and if you find it useful, share it with other people. We have ideas on better ways to <em>read </em>Twitter Threads, and we want to build them if there’s interest.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=aa9606ca714c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/jose-duarte/desire-path-design-twitter-threads-aa9606ca714c">Desire Path Design — Twitter Threads</a> was originally published in <a href="https://medium.com/jose-duarte">Jose Duarte</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[#TBT to the time I predicted a major iTunes feature]]></title>
            <link>https://medium.com/jose-duarte/tbt-to-the-time-i-predicted-a-major-itunes-feature-379f6b841fb9?source=rss----4d3434a1bc45---4</link>
            <guid isPermaLink="false">https://medium.com/p/379f6b841fb9</guid>
            <category><![CDATA[iphone]]></category>
            <category><![CDATA[itunes]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[ios]]></category>
            <category><![CDATA[apple]]></category>
            <dc:creator><![CDATA[Jose Duarte]]></dc:creator>
            <pubDate>Thu, 13 Jul 2017 15:49:12 GMT</pubDate>
            <atom:updated>2017-07-13T15:50:36.995Z</atom:updated>
            <content:encoded><![CDATA[<iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F-wfv0OJ1oMQ%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D-wfv0OJ1oMQ&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F-wfv0OJ1oMQ%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/2431208089050493ede40f57570e81af/href">https://medium.com/media/2431208089050493ede40f57570e81af/href</a></iframe><p>In the lead-up to the 2009 release of iTunes and iOS, there were rumors that Apple was working on a way to manage apps in iTunes. This video is a simple Keynote presentation I made around that time to show how Apple could implement these features (and a couple more that I knew would never make it)</p><p>After making it, I started posting it to Mac forums and tech sites. It got a lot more pickup than I thought, making it into: <a href="https://forums.appleinsider.com/discussion/95546/solution-for-poor-iphone-app-icon-management-proposed">AppleInsider</a>, <a href="http://gizmodo.com/5158506/itunes-concept-shows-how-iphone-app-management-should-have-worked-from-the-start?tag=Iphone-app-management">Gizmodo</a>, <a href="https://www.google.com/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=4&amp;cad=rja&amp;uact=8&amp;ved=0ahUKEwib5bH474PVAhVBNSYKHUgTDjQQFgg0MAM&amp;url=http%3A%2F%2F9-bits.com%2Fpost%2F80905473%2Fapp-management-in-itunes-concept-via-svdomer09&amp;usg=AFQjCNH03P8kgRoeadTrlj7IgsCWzPizew">9Bits</a>, and <a href="https://www.google.com/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=12&amp;cad=rja&amp;uact=8&amp;ved=0ahUKEwjb8Jza8IPVAhXGOSYKHe7vByYQFghVMAs&amp;url=http%3A%2F%2Fwww.iphonehacks.com%2F2009%2F02%2Fconcept-itunes-based-solution-to-manage-iphone-apps.html&amp;usg=AFQjCNFwN5ZG4xBCoNW95PBNzwF7sQHJ2Q">a</a> <a href="https://www.google.com/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=16&amp;cad=rja&amp;uact=8&amp;ved=0ahUKEwjh9rC28IPVAhXMPCYKHSYwC9Y4ChAWCD8wBQ&amp;url=https%3A%2F%2Fwww.onemorething.nl%2F2009%2F02%2Fbeheerconcept-voor-apps-in-itunes%2F&amp;usg=AFQjCNHMLWRDNgwgxSuT-igJ0ehduJv4Qw">bunch</a> <a href="https://www.google.com/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=15&amp;cad=rja&amp;uact=8&amp;ved=0ahUKEwjh9rC28IPVAhXMPCYKHSYwC9Y4ChAWCDcwBA&amp;url=http%3A%2F%2Fwww.maxiapple.com%2F2009%2F02%2Finterface-itunes-genial-concept-video.html&amp;usg=AFQjCNFnQTTrjXI4CTgpGK1rXDfFTZ-QGA">more</a> <a href="https://www.google.com/url?sa=t&amp;rct=j&amp;q=&amp;esrc=s&amp;source=web&amp;cd=19&amp;cad=rja&amp;uact=8&amp;ved=0ahUKEwjh9rC28IPVAhXMPCYKHSYwC9Y4ChAWCFUwCA&amp;url=https%3A%2F%2Fhipertextual.com%2F2009%2F02%2Fcomo-nos-gustaria-administrar-aplicaciones-en-el-iphone&amp;usg=AFQjCNG4ohvFgfl6UgMc4GaIRF7EIuhVHw">sites</a>.</p><p>All in all the video got almost 500,000 views, which is WAY more than I ever expected. This is what that feature ended up looking like when Apple finally released it:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/586/1*FAjquZCbsGQnrVdrmtohUg.jpeg" /><figcaption>Apple</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=379f6b841fb9" width="1" height="1" alt=""><hr><p><a href="https://medium.com/jose-duarte/tbt-to-the-time-i-predicted-a-major-itunes-feature-379f6b841fb9">#TBT to the time I predicted a major iTunes feature</a> was originally published in <a href="https://medium.com/jose-duarte">Jose Duarte</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Taking iOS11’s Notification Pane to its logical conclusion]]></title>
            <link>https://medium.com/jose-duarte/rethinking-the-ios-11-notification-center-and-spring-board-e74e606b03bf?source=rss----4d3434a1bc45---4</link>
            <guid isPermaLink="false">https://medium.com/p/e74e606b03bf</guid>
            <category><![CDATA[design]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[apple]]></category>
            <category><![CDATA[ios]]></category>
            <dc:creator><![CDATA[Jose Duarte]]></dc:creator>
            <pubDate>Thu, 13 Jul 2017 00:49:17 GMT</pubDate>
            <atom:updated>2017-07-13T15:48:38.095Z</atom:updated>
            <content:encoded><![CDATA[<p>With iOS11, Apple has decided to combine the Lock Screen and the Notification center into one element called the “Notification Pane.” Apple had been hinting at this change with prior iOS releases. For example, the widgets section moved <em>out </em>of notification center in iOS10 and into the Lock Screen. It’s a bit of an unpopular opinion, but I actually like this change. If Apple’s gonna have both the lock screen and notification center duplicate information, then it makes sense to combine them.</p><p>The problem is Apple hasn’t quite fully embraced this change and the OS has become more confusing for it. Below you’ll find my thoughts on how Apple can take this joint Notification Pane concept to its logical conclusion. But first, we have to go all the way back to the last major iOS redesign (iOS7) to understand <em>why </em>the way Apple has implemented Notification Pane is confusing.</p><p>When Apple introduced iOS7, they also introduced the concept of ‘functional layers.’ The idea was that each element of the OS, like Control Center, the Home screen and Notification Center lived in a separate layer. This was a very clean and elegant design concept that instantly made sense, particularly with the heavy use of translucency throughout the interface.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*L1eYbWNgOmvTZp6Nhx_DDQ.png" /><figcaption>Functional Layers in iOS7. Apple</figcaption></figure><p>Every subsequent release has scaled back on the more extreme elements of the iOS 7 design paradigm (Everything is a little less flat now). As Apple has added new features like Widgets and interactive notifications, things have gotten a little more complicated as well.</p><p>The behavior of this new hybrid element is different depending on how you invoke it. Another weird duplicative feature is the Widgets screen; which has 2 different layouts, depending on how you got there as well.</p><p>If you map out iOS UI navigation, you get something like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/768/1*x0eQw2KK_KWR6S1V7UjU9g.jpeg" /><figcaption>There’s also Control Center and Search, but those work pretty well.</figcaption></figure><p>The <strong>Notification Center</strong> layer is by far the most inconsistent of the 3. By default, you see the <strong>Notification Pane</strong>. If you swipe to the left, you get the Big Clock version of the Widgets screen. If you swipe to the right, you get the Camera App.</p><p>Here’s where it can get even more confusing. If you just woke up your phone, then you get a locked version of the <strong>Notification Pane</strong> home screen (signified by a hard-to-see locked icon). You can swipe left and right for widgets and camera, but you can’t swipe up to go anywhere else.</p><p>On the other hand, if you get to <strong>Notification Pane </strong>by pulling down from the top, it allows you to get back to what you were doing by pulling up from the bottom of the screen.</p><p>The weirdest thing about this layer is the Camera App. Writing this article I noticed that this instance of the Camera App is completely different from the one you access through the home screen. It makes for an amusing loop:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FXcbBzaTT-RQ%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DXcbBzaTT-RQ&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FXcbBzaTT-RQ%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="640" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/1091d3e3a2e884f49cbbea9fa5842580/href">https://medium.com/media/1091d3e3a2e884f49cbbea9fa5842580/href</a></iframe><p>On the <strong>Home Screen, </strong>you can swipe to the left to go to a <em>slightly different </em>widgets screen that has no large clock (there doesn’t seem to be a good reason for why there’s no clock in this view). And finally, you can tap any app icon to open an app in the <strong>app layer. </strong>From here, you can pull down to go to the <strong>Notification Pane </strong>again, or press the home button to go to the first screen of App icons.</p><h4>Simplifying the paradigm.</h4><p>I think Apple can easily remove one of the three layers. This would simplify and clarify the UI; and the functionality wouldn’t have to change much. This is how I would design this new two-layer paradigm:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/768/1*hyMLgpM_WIgZ5529nLIF4w.jpeg" /></figure><p>The combined <strong>Home </strong>/ <strong>Notification Layer </strong>makes navigation a lot simpler. Instead of the Camera App to the right in this layer, users would be able to navigate their traditional rows of icons. The Widgets section remains to the left (but only the Big Clock version).</p><p>The addition of a <strong>persistent dock </strong>is a natural extension of this new iPad feature. The Siri suggestion app would default to the Camera App when the phone is locked. This solves another issue with the lock screen camera app: the slide-to-the-right Camera we have right now is sticky and glitchy.</p><p>This is what this new interface would look like in action:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FZJZN1UGgfcM%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DZJZN1UGgfcM&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FZJZN1UGgfcM%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="640" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/e66ea3224e5a87f2b5f377dac3e28a41/href">https://medium.com/media/e66ea3224e5a87f2b5f377dac3e28a41/href</a></iframe><p>I think this makes navigating the OS a lot more fluid and quick. The only small usability concession that breaks this paradigm is the Notification Center fading in when sliding down from the top on the <strong>Home Layer.</strong></p><p>Another benefit of this paradigm is it makes the Home button completely unnecessary to navigate the interface; which would come in handy with the rumored home button-less iPhone 8.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e74e606b03bf" width="1" height="1" alt=""><hr><p><a href="https://medium.com/jose-duarte/rethinking-the-ios-11-notification-center-and-spring-board-e74e606b03bf">Taking iOS11’s Notification Pane to its logical conclusion</a> was originally published in <a href="https://medium.com/jose-duarte">Jose Duarte</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Desire Path Design]]></title>
            <link>https://medium.com/jose-duarte/desire-path-product-design-2e1aacaac5b4?source=rss----4d3434a1bc45---4</link>
            <guid isPermaLink="false">https://medium.com/p/2e1aacaac5b4</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[facebook]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[product-design]]></category>
            <category><![CDATA[social-media]]></category>
            <dc:creator><![CDATA[Jose Duarte]]></dc:creator>
            <pubDate>Wed, 12 Jul 2017 13:36:00 GMT</pubDate>
            <atom:updated>2018-03-28T16:43:37.646Z</atom:updated>
            <content:encoded><![CDATA[<p>Have you ever found yourself so frustrated with social media websites that your workflow includes 3 other apps just to get basic work done? Is your Photoshop workflow so custom, complicated and so full of shortcuts and third party plugins that few other mortals can decipher it? If you said yes; you probably discovered (or created) a Desire Path.</p><p><strong>In landscaping design,</strong> Desire Paths arise from unforeseen inefficiencies in the way the designers attempted to guide people through their paths. Something like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*D1R7QvewXmwUVytD." /><figcaption>A perfect example of a Desire Path. Who builds a walkway roundabout? photo credit: <a href="https://imgur.com/JIXACph">https://imgur.com/JIXACph</a></figcaption></figure><p>Desire paths can cause a rift between the people who design the landscapes and the people actually using them.</p><p>I saw a battle between landscapers play out while I was in my undergrad; the administration stubbornly tried to stop people from cutting across a lawn. Hedges and fences did nothing to stop us; in fact, it motivated us.</p><p>While I didn’t see the end of that fight while I was still in school, I checked in on that spot on Google Earth, and we won:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*A6Q8EQge_halCZyt." /><figcaption>You can tell they really didn’t want to make that path on the right because otherwise they would’ve aligned the mortar properly. Photo Credit: Google Earth.</figcaption></figure><h4>How do Desire Paths translate to product design?</h4><p>In product design, Desire Paths are all the tools, tips, and shortcuts people use to get work done faster with your product, or to use your product in ways you never intended.</p><p>A couple of examples, one big and one small of Desire Path design:</p><ol><li><strong>YouTube </strong>was originally intended as a <a href="https://www.washingtonpost.com/news/innovations/wp/2015/07/02/the-7-greatest-pivots-in-tech-history/">video dating site</a><strong>. </strong>The idea was, people would upload videos of themselves as their dating profile. Users <em>quickly </em>started uploading all sorts of videos to the site and YouTube pivoted to what their users wanted.</li></ol><p>2. The <strong>Facebook Branded Content Tool</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/391/0*z5X_glYd9KqLtBIX." /><figcaption>The “with Adobe Spark” copy is how Facebook complies with advertising guidelines and regulations about sponsored posts. Mari Smith.</figcaption></figure><p>It<strong> </strong>was created because people with big Facebook pages discovered an unintended way to use Facebook’s pages product: getting paid to promote stuff. Facebook didn’t want to thwart that desire path, but Influencers keep<a href="https://www.theverge.com/2017/4/19/15364828/instagram-celebrities-ftc-letters-ads"> getting in trouble with the FTC</a>. So they created a tool to disclose the fact that these were paid posts + some additional tracking features and shared insights.</p><p>Product managers should also look towards third party apps and plugins built for their products. These kinds of products are usually desire paths, particularly for users who use your platform daily, at high volume, or to make a living (aka power users). For example, It’s not a big leap to see Facebook’s post scheduling tool partly as a response to how popular tools like Hootsuite are with their power users.</p><h4>How to prioritize Desire Path features.</h4><p>Features inspired by your users’ desire paths should not automatically be treated as top priority; as any product manager knows, just because users <em>want </em>something doesn’t mean it is a good feature to build into your product.</p><p>Instead, Desire Path features should be prioritized in the same way you prioritize your other features, with a couple additional factors to consider:</p><ol><li><strong>How many people use the desire path?</strong></li></ol><p>Going back to my University example, the Desire Path that was ultimately paved over wasn’t the <em>only </em>desire path on campus. There were other, more manageable ones where people didn’t completely kill the grass or made a muddy mess when it rained. If your feature will pave over a muddy mess, you should prioritize it higher.</p><p><strong>2. Avoid the Power User pitfall.</strong></p><p>Desire Path features, almost by nature, will tend to skew more toward power users. They are the ones using your product the most, getting frustrated and finding those tips and shortcuts that make up product desire paths.</p><p>When considering power user features, make sure you don’t build features targeted to such a niche group of Power Users that it does not further your overall product goals, nor justify the time and resource investment. That’s the <strong>Power User pitfall.</strong></p><p>When my company, dose., ran a big network of user-generated content sites (like<a href="http://smartphowned.tumblr.com/"> smartphowned.com</a> and<a href="https://www.facebook.com/givesmehope/"> givesmehope.com</a>), we ran into this issue all the time; even though we didn’t realize it at the time.</p><p>The websites were successful because we allowed power users to submit a lot of content, other power users would vote that content up or down, and we’d publish the best, based on our results.</p><p>In building features for the site, we constantly looked toward building features that encouraged power users to submit and vote more because that was the lifeline of our network of sites.</p><p>Along the way, there were multiple instances when we both avoided, and fell into the power user pitfall. I’ll give one example of each.</p><ol><li><strong>We fell into the pitfall </strong>when we tried to build a robust User Account system for our network of sites. We heard from a small number of power users that they wanted a system that allowed them to track their submissions and comments across our sites. We got excited with this idea (as we also were power users of the sites ourselves) and after a long development time we had a User Account system with a ton of bells and whistles that made no impact on our bottom line or our traffic. We had misjudged how many of our users <em>actually </em>wanted the feature; and to make matters worse, we built more than what they asked us to do.</li><li><strong>We navigated successfully </strong>with the “builders” we developed for each site. They allowed power users to quickly and easily create the content in our site format. Take this one from Smartphowned, as an example:</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/430/0*A1WlaoQoJm6K4hCE." /><figcaption>Yes, if you’ve seen a Smartphowned post, it’s probably fake and someone made it in this builder. Dose.</figcaption></figure><p>The builders took a <strong>lot </strong>of time and resources to build. However, they worked <em>great </em>at encouraging our power users to submit content; the Smartphowned builder in particular was a massive success for us and catapulted the site to 60m+ monthly pageviews for a period of time, on the back of a strong number of submissions.</p><p><strong>The best way</strong> to avoid this pitfall is to have a great ROI analysis (Return On Investment) and to never forget: You should be your own product’s #1 power user; but this can make certain features look more useful than they actually are.</p><p>A good way to improve a feature’s ROI is to scale back its scope and build something that a higher percentage of your power users will use.</p><p>A great example of this is Facebook’s scheduling tool:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*muer2wC1BBqJoq5g." /><figcaption>The Facebook scheduler only lets you schedule one post at a time. However, it’s still the best way to schedule a video post.</figcaption></figure><p>Facebook created this feature as a response to third party scheduling tools like <a href="http://hootsuite.com/">Hootsuite</a> or <a href="http://buffer.com/">Buffer</a>. These two third party services let social media managers (the “0.01%” of Facebook power users) schedule hundreds of posts to multiple accounts using just an Excel spreadsheet.</p><p>Facebook’s scheduler, on the other hand, only lets you schedule one post in one account at a time.</p><p>Naturally, Facebook’s feature didn’t replace Hootsuite for social media managers. It did, however, help the thousands of small business owners and influencers who want to schedule 3–4 posts a day and have no need for something more complex.</p><p>Once I started looking for Desire Path features, I started finding ideas everywhere. This is hopefully the first in a series of desire path product feature ideas; first up: Tweetstorms.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2e1aacaac5b4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/jose-duarte/desire-path-product-design-2e1aacaac5b4">Desire Path Design</a> was originally published in <a href="https://medium.com/jose-duarte">Jose Duarte</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Facebook’s “Fake News” button won’t help against Russia. It’s all about the algorithm.]]></title>
            <link>https://medium.com/jose-duarte/why-facebooks-fake-news-button-won-t-help-against-russia-it-s-all-about-the-algorithm-10cb8799c4c5?source=rss----4d3434a1bc45---4</link>
            <guid isPermaLink="false">https://medium.com/p/10cb8799c4c5</guid>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[fake-news]]></category>
            <category><![CDATA[facebook]]></category>
            <category><![CDATA[politics]]></category>
            <category><![CDATA[donald-trump]]></category>
            <dc:creator><![CDATA[Jose Duarte]]></dc:creator>
            <pubDate>Wed, 12 Jul 2017 13:32:03 GMT</pubDate>
            <atom:updated>2017-09-08T17:11:30.458Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/640/0*-TdDf1qk7tYx5TOZ." /><figcaption><a href="https://pixabay.com/en/fake-fake-news-media-laptop-1909821/">Pixabay</a></figcaption></figure><p>The extent of Russia’s fake news hack of the 2016 election is widely misunderstood. The news media doesn’t understand the underlying technology that hackers used and as a result has not properly explained it to the public. The operation wasn’t just about writing <a href="http://www.slate.com/blogs/future_tense/2017/01/05/russia_used_fake_news_to_influence_the_election_james_clapper_says.html">thousands of fake stories</a> that were harmful to Clinton and helpful to Trump: Russian hackers socially engineered a massive <a href="http://www.slate.com/blogs/future_tense/2017/01/05/russia_used_fake_news_to_influence_the_election_james_clapper_says.html">propaganda campaign</a> on social media networks like Facebook, Twitter and Reddit.</p><p>The response from social networks, particularly <a href="https://www.facebook.com/zuck/posts/10103269806149061">Facebook’s</a>, has been to declare war on fake news by creating stronger filters and integrating fact checking sites into their News Feed. While these measures will help, they ignore the other half of the issue.</p><p>Russian hackers exploited Facebook’s algorithm (which it uses to decide what content appears in your news feed) to maximize their impact. I recognize a lot of the techniques they used, as I’ve been studying the Facebook algorithm for years at Dose. I’ve run hundreds of experiments and regressions trying to predict virality and understand people’s behavior. I used that knowledge to grow our Facebook pages from 400,000 to 4,000,000 in just four months.</p><p>I hope that by sharing what I know, we can open a conversation on how to improve this new social “infrastructure” that has become an integral part of our lives.</p><h4><strong>The power of fake news</strong></h4><p>So how successful were the Russians in spreading fake news? Very. The top 115 fake news articles got <a href="https://web.stanford.edu/~gentzkow/research/fakenews.pdf">at least</a> 30 million shares, and <a href="http://www.motherjones.com/politics/2017/06/russian-active-measures-trump-propaganda-conspiracy-theories/">some</a> were even picked up by the Trump campaign.</p><p>But the goal of any propaganda campaign is influence. Did fake news change enough votes to sway the election? That’s where it gets tricky. A <a href="https://web.stanford.edu/~gentzkow/research/fakenews.pdf">recent</a> Stanford study concluded that fake news might not have had a huge impact on the election.</p><p>On the other hand, fewer than 78,000 votes in three states <a href="http://www.weeklystandard.com/the-election-came-down-to-77744-votes-in-pennsylvania-wisconsin-and-michigan-updated/article/2005323">decided the election</a>, so the effect didn’t have to be big to be impactful.</p><h4><strong>Conditions were ripe for attack</strong></h4><p>Part of the reason the attack worked so well was due to some long term trends in the US:</p><ul><li>Increasing distrust of traditional media: <a href="http://www.journalism.org/2016/05/26/news-use-across-social-media-platforms-2016/">A recent Pew Research survey</a> found almost two-thirds of Americans get their news primarily from social media. (Even worse, more than half get all news from one site.)</li><li>Online, people don’t get exposed to opposing points of view due to <a href="https://arstechnica.com/science/2017/03/the-social-media-echo-chamber-is-real/">self-segregation</a> — in other words, we tend to tune out people who have political views that are different from ours.</li><li>The <a href="https://www.strategyand.pwc.com/reports/content-democratization">democratization of content</a> on social networks means that articles from real and fake sources appear the same in your feeds.</li><li>Social networks show you what they <em>think you want</em> to see — not a balanced offering of verified news and viewpoints.</li></ul><h4><strong>Understanding how Facebook decides what content to show you</strong></h4><p>Facebook <a href="https://newsfeed.fb.com/welcome-to-news-feed?lang=en">says</a> the goal of the news feed is to “show you the stories that matter most to you every time you visit Facebook.” To do that, it measures literally everything you do on the site and knows more about you than you may realize. For example, Facebook revealed last year that <a href="https://www.theverge.com/circuitbreaker/2016/8/24/12621784/facebook-political-preferences-ads">it labels your political preferences</a> based on the information they have on you.</p><p>Based on my own research, It’s safe to assume Facebook has hundreds of similar categories it sorts us all into. We’ve also found these are some of the metrics that Facebook actively measures:</p><ul><li>What kind of content you engage with the most. Do you like pictures more often than videos?</li><li><em>Whose</em> content you engage with the most. Is it your siblings’ photos, Tasty videos or a fake news page?</li><li>How long you spend reading a link you clicked. (If you click on a link and immediately return to your news feed, Facebook interprets that as you NOT liking that content)</li><li>How long you spend watching a video (down to the second, they know when you stopped watching)</li></ul><h4><strong>How content spreads on Facebook</strong></h4><p>When a Facebook page posts something, the algorithm won’t automatically show it to everybody who likes the page. Rather, Facebook will show it to a sample of people who like the page. Based on how they respond, Facebook will either continue to show it to more people or kill it entirely. They do this millions of times a day, for all content on their site.</p><p>To illustrate this, let’s use the hypothetical example below. Imagine a Facebook page with four followers. The page’s administrator posts something about green diamonds. Facebook will sample only one person, before deciding whether to show the content to anyone else.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/864/0*sLAEF2jnE8teB17a." /><figcaption>Jose Duarte</figcaption></figure><ul><li>Mary doesn’t<strong> </strong>like green diamonds and doesn’t engage with the post. So Facebook will stop showing it. Same for Rob.</li><li>John, however,<em> likes</em> green diamonds and engages. Facebook sees this as a good sign and shows the post to some of John’s friends. None of John’s friends like green diamonds. The post dies.</li><li>Nia, on the other hand, is a gold mine. She and all her friends love diamonds. The post will probably go viral.</li></ul><h4>What else did the Russians do?</h4><p>There are three active measures (that we know of!) that Russian hackers used to maximize the spread of fake news and artificially engineer virality.</p><ul><li>They successfully infiltrated <a href="https://www.washingtonpost.com/world/national-security/homeland-security-official-russian-government-actors-potentially-tried-to-hack-election-systems-in-21-states/2017/06/21/33bf31d4-5686-11e7-ba90-f5875b7d1876_story.html?utm_term=.abb0eef9ea8f">voter rolls</a> in at least 21 states.</li><li>They used that data to create thousands of <a href="http://www.npr.org/sections/alltechconsidered/2017/04/03/522503844/how-russian-twitter-bots-pumped-out-fake-news-during-the-2016-election">fake profiles</a>. They used these fake profiles to impersonate different groups of people in the US. Their goal was to fool Facebook users — and Facebook itself — into thinking these were real people. As an example, fake profiles that appeared to be <a href="http://www.npr.org/sections/alltechconsidered/2017/04/03/522503844/how-russian-twitter-bots-pumped-out-fake-news-during-the-2016-election">Republicans in the Midwest</a> seem to have been particularly effective as a way to make people implicitly trust that the fake news articles were real.</li><li>They may have <a href="https://www.dailykos.com/story/2017/5/23/1665234/-FEC-looks-into-whether-Trump-and-Russia-spread-fake-news-on-Facebook">illegally</a> paid Facebook to place ads promoting fake news that microtargeted key swing voters.</li></ul><p><strong>Putin it all together</strong></p><p>Imagine that page posting green diamonds was actually a front for a Russian fake news operation. Now, let’s also add a few fake users to see how the algorithm was fooled (the Russians are in red):</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/856/0*s_o3a5jeK4vY2es4." /><figcaption>Jose Duarte</figcaption></figure><p>Using all the data they exploited, the Russians were able to craft a fake news article that Nia — a fake user — shared to her network, and a bunch of the people who liked the green diamonds post were served fake news meticulously crafted to ensure Facebook would show it to them.</p><p>It’s true that a single fake news article couldn’t have changed the result of the election. However, the fake news operation was insidious, long and way more effective than has been reported so far. Russians know very well how much Americans have isolated themselves in bubbles and echo chambers, and how our social networks are designed to exploit that isolation by mostly showing content from our bubbles.</p><p>This vacuum allowed them to bombard swing voters with fake news over a period of months<em>, </em>slowly but surely flipping people from Clinton to Trump (or getting them to stay home).<em> </em>What’s worse, most other people <em>had no idea </em>this was happening because social networks have become so effective at hiding things outside our bubbles. For example, liberals in New York had no idea that the Pope “<a href="http://www.factcheck.org/2016/10/did-the-pope-endorse-trump/">endorsed Trump</a>.”</p><p>Imagine your TV was only tuned to CNN, and CNN changed the kind of news it broadcast to fit what it <em>thought </em>you wanted to see — and it did that for everyone in America. That is what social networks have become.</p><p>So how do we fix it?</p><p>The people who want to abuse Facebook’s algorithm already know what levers to pull to exploit it for their own nefarious ends. But most Americans probably don’t even know people are doing that. And if you don’t know that, you just think Facebook is showing you everything in a natural, organic way. So the first step is teaching Americans the basics of how the Facebook algorithm works.</p><p>The second step is deciding what to do about it. Should the algorithm be regulated similar to broadcast TV? Should there be an independent panel in charge of shaping the algorithm so that it can’t be so easily exploited? Should there be no algorithm at all?</p><p>These are questions we have to answer. Only by having a more open conversation with people — regular people, outside of Facebook and Twitter — can we start to solve this problem before the next election.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=10cb8799c4c5" width="1" height="1" alt=""><hr><p><a href="https://medium.com/jose-duarte/why-facebooks-fake-news-button-won-t-help-against-russia-it-s-all-about-the-algorithm-10cb8799c4c5">Why Facebook’s “Fake News” button won’t help against Russia. It’s all about the algorithm.</a> was originally published in <a href="https://medium.com/jose-duarte">Jose Duarte</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>