<?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 Rifki Razick on Medium]]></title>
        <description><![CDATA[Stories by Rifki Razick on Medium]]></description>
        <link>https://medium.com/@rifki_razick?source=rss-c9724c48626------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*8nToEHQZ-h-XV4PRzxjneQ.jpeg</url>
            <title>Stories by Rifki Razick on Medium</title>
            <link>https://medium.com/@rifki_razick?source=rss-c9724c48626------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 21:38:12 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@rifki_razick/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[Mythos: Should we brace for a CVE deluge?]]></title>
            <link>https://medium.com/@rifki_razick/mythos-should-we-brace-for-a-cve-deluge-e426d73675f9?source=rss-c9724c48626------2</link>
            <guid isPermaLink="false">https://medium.com/p/e426d73675f9</guid>
            <dc:creator><![CDATA[Rifki Razick]]></dc:creator>
            <pubDate>Sat, 11 Apr 2026 03:57:43 GMT</pubDate>
            <atom:updated>2026-04-11T04:02:30.198Z</atom:updated>
            <content:encoded><![CDATA[<p>This week, Anthropic unveiled <strong>Mythos</strong>, their most powerful model to date. It’s so powerful that they decided not to release it to the wider public. Through project <a href="https://www.anthropic.com/glasswing">Glasswing</a>, Anthropic is making the preview version available only to a select group of companies including NVIDIA, Amazon, Google, Microsoft et al. and in my mind quite importantly <strong>The Linux Foundation</strong></p><p>Mythos represents a specific shift in the landscape of software security. It isn’t just their best coding agent; early indicators suggest it is the most formidable model yet for identifying security defects.</p><p>From the Glasswing blog post…</p><blockquote>Mythos Preview has already found <strong>thousands </strong>of high-severity vulnerabilities, including some in <em>every major operating system and web browser</em>.</blockquote><p>Some critics call this Anthropic/Dario <a href="https://www.youtube.com/watch?v=INGOC6-LLv0">theatre</a> which sometimes feels like it, however I’ve always been a proponent of their more responsible approach to AI from the very first Claude models which had <a href="https://www.anthropic.com/news/claudes-constitution">constitutions </a>baked in during post-training. And of course they’ve alway had the best coding models in my mind. Agentic coding in the early Claude models was amazing but Opus 4.6 just blew my mind.</p><p>For software vendors, this is a “double-edged sword” moment. While Mythos can help us build more secure software, it is also about to be unleashed on the world’s existing codebases. We should likely prepare for a “CVE deluge” — a massive spike in reported vulnerabilities across Software Bills of Materials (SBOMs) everywhere.</p><p>So, how should enterprise software teams prepare for this impending wave of security findings?</p><h3>1. Ensure Full SBOM Transparency</h3><p>Before you can fix vulnerabilities, you must know where they are. Maintaining a comprehensive and transparent Software Bill of Materials (SBOM) is no longer optional. It is the foundational map you will need to navigate the coming surge of reports.</p><h3>2. Fine-Tune Your Scanning Pipelines</h3><p>This should already be standard practice for enterprise software, but now is the time to ensure your SAST (Static Application Security Testing) and DAST (Dynamic Application Security Testing) scanning pipelines are well-oiled and fully integrated. If your pipelines are slow or brittle, they will become a bottleneck when the volume of findings increases.</p><h3>3. Clear the Technical Debt</h3><p>Don’t wait for the deluge to start. Use this window to fix all existing CVEs in your current codebase. This typically involves upgrading open-source libraries and uplifting your code to remain compatible. Testing your code against these new versions now ensures you aren’t fighting “old” fires when the “new” ones arrive.</p><h3>4. Scale Fixes with Agentic Orchestration</h3><p>To keep up with a model-driven deluge, you need model-driven solutions. Start working on orchestration flows to fix security defects autonomously.</p><ul><li><strong>The Workflow:</strong> Utilize <strong>Claude Opus 4.6</strong> (via tools like Cursor or Claude Code) to analyze and patch defects at scale.</li><li><strong>Parallelization:</strong> Think in terms of sub-agents working in parallel to address multiple vulnerabilities simultaneously.</li><li><strong>The Safety Loop:</strong> Ensure every automated PR is reviewed by a second model for logic and security, with a final human-in-the-loop sign-off before merging.</li></ul><h3>5. Prioritize Test Coverage and Automation</h3><p>A deluge of CVEs means a deluge of Pull Requests. You cannot manually verify every fix without slowing your release cycle to a crawl. Your unit test coverage and test automation pipelines must be in prime shape to provide the necessary confidence that automated security fixes haven’t introduced regressions.</p><h3>6. Adjust Your Roadmap</h3><p>Finally, be realistic about your capacity. The overhead of managing a sudden spike in security maintenance will impact feature velocity. Add necessary “slack” to your planning for 2026 to account for the time your team will spend triaging and merging these fixes.</p><p>The arrival of Mythos means the speed of vulnerability discovery is about to accelerate. By hardening your pipelines and embracing agentic workflows now, you can turn a potential deluge into a manageable, automated process of continuous improvement.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e426d73675f9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Telling LLMs to self evaluate]]></title>
            <link>https://medium.com/@rifki_razick/telling-llms-to-self-evaluate-ca06a4588ec5?source=rss-c9724c48626------2</link>
            <guid isPermaLink="false">https://medium.com/p/ca06a4588ec5</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[genai]]></category>
            <dc:creator><![CDATA[Rifki Razick]]></dc:creator>
            <pubDate>Thu, 10 Jul 2025 05:30:57 GMT</pubDate>
            <atom:updated>2025-08-03T07:23:11.063Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/559/1*c1_T2WM1Ejq4QCr9t48ZOw.jpeg" /></figure><p>Experimenting with LLMs for quite a while now, lately I’ve been trying out an interesting technique that just popped in my head while doing a slightly complex prompt. The basic idea is to tell the LLM to evaluate its own confidence level of it’s response. This is more like a ‘self eval’ of sorts. Now this is not a replacement for proper <a href="https://python.langchain.com/docs/concepts/evaluation/">evaluations and testing</a> when building production grade Gen AI apps, however it can and probably should be used in combination when building apps.</p><p>Its pretty simple in practice, when prompting I tell the model to give it’s confidence level of it’s own prediction as part of its output.</p><p>This works quite well for many use cases especially when you prompt for structured output, but I guess it can be used in general as well.</p><p>Here’s a simple example prompt to make this clearer:</p><blockquote><em>Analyze the attached CSV data containing customer feedback. For each entry, determine the sentiment as either ‘Positive’, ‘Negative’, or ‘Neutral’. </em><strong><em>Additionally, provide a confidence rating for the sentiment analysis on a scale of 1 to 5, where 1 is “Very Low Confidence” and 5 is “Very High Confidence.</em></strong><em>” Respond with a new CSV containing the original feedback, the ‘Sentiment’ in a new column, and the ‘Confidence’ rating in another new column.</em></blockquote><p>This can be used for almost any prompt from sentiment analysis tasks like mentioned above to analyzing annual reports and other relevant sources in order to understand the financial health of public quoted companies.</p><p><em>Update (03/08/25): The paper </em><a href="https://arxiv.org/abs/2503.02863"><em>SteerConf: Steering LLMs for Confidence Elicitation</em></a><em> seems to suggest a similar approach in more detail, worth a read.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ca06a4588ec5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What’s the Chorus of your Product?]]></title>
            <link>https://medium.com/@rifki_razick/whats-the-chorus-of-your-product-3ddd74d358db?source=rss-c9724c48626------2</link>
            <guid isPermaLink="false">https://medium.com/p/3ddd74d358db</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[Rifki Razick]]></dc:creator>
            <pubDate>Sun, 06 Jul 2025 02:25:33 GMT</pubDate>
            <atom:updated>2025-07-06T02:25:33.472Z</atom:updated>
            <content:encoded><![CDATA[<p>After listening to a song you like what part of the song plays over and over in your head?</p><p>What part of the song do you keep on singing in the car to work or in the shower?</p><p>What part of the song do you hum to your friends when telling them how great it is?</p><p>It’s the <strong>Chorus</strong>!</p><p>The chorus or hook of a song gives it a unique identity. It is the most important ‘feature’ that makes it different from other songs. It’s <a href="http://whatis.techtarget.com/definition/unique-selling-point-USP">USP.</a></p><p>So what can the software industry learn from the music industry?</p><p>Let’s first have a look at this fascinating video (starts at 0:15) of how Sia comes up with a new song. (in fact it is the inspiration behind this blog post)</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F59nwMeoYCkQ%3Fstart%3D16%26feature%3Doembed%26start%3D16&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D59nwMeoYCkQ&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F59nwMeoYCkQ%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/8cf9889a1eef2d544236587e1d42af49/href">https://medium.com/media/8cf9889a1eef2d544236587e1d42af49/href</a></iframe><p>The cool thing to note here is how she mainly focuses on identifying the chorus and getting it to sound awesome, even without any lyrics.</p><p>It’s amazing how we can stimulate product strategy in our own industry by looking at other almost unrelated industries. The key is to not isolate your knowledge to your own industry alone but widen it to other industries and try to draw parallels that will help with your own strategy.</p><p>So if you are building a consumer (or even enterprise) software product make sure you identify it’s <em>Chorus,</em> the feature that makes it unique. Then focus most of your effort and resources on it to make your users fall in love with it and talk about it with other potential users.</p><p>Product leaders should always strive towards helping their teams identify the chorus of their product they are building and set that as the north star for what they do.</p><p>What is the ‘chorus’ of the product you are building? Can your entire team articulate it in a single sentence? If not, that is your first and most critical task.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3ddd74d358db" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Speculation]]></title>
            <link>https://medium.com/@rifki_razick/speculation-3ddefa94dc35?source=rss-c9724c48626------2</link>
            <guid isPermaLink="false">https://medium.com/p/3ddefa94dc35</guid>
            <dc:creator><![CDATA[Rifki Razick]]></dc:creator>
            <pubDate>Wed, 04 Jun 2025 00:22:22 GMT</pubDate>
            <atom:updated>2025-06-04T00:29:26.273Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Speculation</em> is something I highly recommend as a form of personal/career development exercise. It’s a form of curiosity. It has the ability to sharpen your thought process and drives you to come up with <strong>your own future view of the world</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/540/1*B49rfnVzC4nhgtBpqL4naQ.png" /></figure><p>At the beginning you would probably suck at it, but with practice you will be amazed at how good you can get at it. After all speculation is free! (at least initially)</p><p>So what should you be speculating about?</p><p>Well, like me you could:</p><p><strong><em>Speculate about the economy</em></strong><em> (e.g. what is the biggest </em><a href="https://en.wikipedia.org/wiki/Foreign_direct_investment"><em>FDI</em></a><em> area in the next 5 years?)</em></p><p><strong><em>Speculate about your industry</em></strong><em> (e.g. who will win/dominate the AI race ? OpenAI, Anthropic, Microsoft, Google? ..and Why? How?)</em></p><p><strong><em>Speculate about other industries</em></strong><em> (e.g. when will autonomous vehicles be the norm?)</em></p><p><strong><em>Speculate about the next disruption</em></strong><em> (e.g. when will all major banks trade Bitcoin?)</em></p><p><strong><em>Speculate about your company</em></strong><em> (e.g. where are your next customers likely to come from?)</em></p><p><strong><em>Speculate about your career</em></strong><em> (e.g. what is the impact of AI to your career?)</em></p><p>…and the list goes on.</p><p>Some personal guidelines I follow on speculation:</p><ul><li>Speculation should be un-emotional, just look at the current context and figure out where something is heading.</li><li>Try and figure out how your speculation of a certain area would affect you</li><li>If your speculation is strong, adjust and adapt yourself to take advantage of it</li><li>Be judicious when making big changes/bets based on speculation (remember <a href="https://en.wikipedia.org/wiki/Financial_crisis_of_2007%E2%80%932008">sub prime loans and the financial crisis</a>?)</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3ddefa94dc35" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Edge Case Bias]]></title>
            <link>https://medium.com/@rifki_razick/edge-case-bias-0aa3ef63c137?source=rss-c9724c48626------2</link>
            <guid isPermaLink="false">https://medium.com/p/0aa3ef63c137</guid>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Rifki Razick]]></dc:creator>
            <pubDate>Sun, 06 Apr 2025 17:10:55 GMT</pubDate>
            <atom:updated>2025-04-06T17:10:55.544Z</atom:updated>
            <content:encoded><![CDATA[<p>An interesting thing I have noted when pitching or demoing a software solution to a potential group of end users is something I like to call “Edge Case Bias”.</p><p>This is when your users provide feedback considering exceptional use cases that are rare and which usually have simple workarounds.</p><p>This sort of feedback usually sounds like <em>“Nice solution! This solution works under the contexts A, B and C but *sometimes* we have situations like Z and this will not work then”</em></p><p>This bias towards edge cases by potential users is quite common but should not be just ignored mainly due to two reasons.</p><p>1. The edge case may not be an edge case after all<br>2. Ignoring the feedback may leave your users with negative feelings</p><p>In my mind product management is not just managing the product per se but also “managing” the users of the product.</p><p>So when considering the “A,B,C and Z” example above a reasonable approach will be to ask your users to come up with rough numbers indicating the frequency which A, B, C (covered by solution) and Z (not covered by solution) occurs.</p><p>Once you get the numbers and find out that “Z” is very rare and has a reasonable workaround you can easily convince the users that changing the solution dramatically to accommodate “Z” is probably not the best way forward.</p><p>Of course if “Z” is a high frequency use case then you may have to do some changes to the product. But from what I have experienced in agile development this is seldom the case.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0aa3ef63c137" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Applying Locality Principle to UX Design]]></title>
            <link>https://medium.com/@rifki_razick/locality-principle-to-ux-design-2d6acd4e36dc?source=rss-c9724c48626------2</link>
            <guid isPermaLink="false">https://medium.com/p/2d6acd4e36dc</guid>
            <category><![CDATA[ux]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[computer-science]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[enterprise-software]]></category>
            <dc:creator><![CDATA[Rifki Razick]]></dc:creator>
            <pubDate>Sun, 30 Mar 2025 00:13:45 GMT</pubDate>
            <atom:updated>2025-03-30T00:22:38.462Z</atom:updated>
            <content:encoded><![CDATA[<p>Designing a great user experience (UX) for complex business software is non-trivial to say the least.</p><p>Screens with complex functionality along with large amounts of data in such systems do not make things easier.</p><p>One of the key elements of good UX is making it easy for the end-user to find “things” in the system that are relevant to them.</p><p>This is where I think the Locality Principle comes into play. It’s a well-known principle in computer science and has implications in memory access optimization.</p><p>So what does this principle tell us? In essence, the Locality Principle states that once a particular piece of data is accessed, there’s a high probability that nearby data will be accessed soon. To keep things super simple, have a look at this video which explains the gist of the principle.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2Fz9LHetPW0Vs%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3Dz9LHetPW0Vs&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2Fz9LHetPW0Vs%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/cf34e2c99e46185ee6ff4fb5cafdf5d5/href">https://medium.com/media/cf34e2c99e46185ee6ff4fb5cafdf5d5/href</a></iframe><p>Now how do we apply this concept when designing user experience (UX) in business software?</p><p>End-users of a business software system typically work with (or “reuse”) a relatively small set of data (and screens) during a given time period (e.g., during a business week). Therefore, the information they search for and work with during this period should be fairly predictable.</p><p>If we can leverage AI to identify information access patterns of users in the system, we could use that “intelligence” to suggest the next possible interaction on the current UI for a given user. This would alleviate the need for the user to “find” the next screen to work with in the system and make it easy to “jump” between screens.</p><p>It must be noted that this is richer than “Recent Screens” functionality, which merely puts the last used screens of one user in a stack. Think something more like Amazon Recommendations instead.</p><p>So the idea is simple: we should leverage the principle of locality in UX design to figure out the next information requirement (e.g., next screen) based on users’ previous UI interaction patterns.</p><p>Finally, the ultimate goal will be to take away the need of human interaction altogether based on these identified patterns and replace them with automation using Agentic AI.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2d6acd4e36dc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Innovator’s Dilemma Amplified in an AI-irst Future]]></title>
            <link>https://medium.com/@rifki_razick/the-innovators-dilemma-amplified-in-an-ai-irst-future-8ef3792ee946?source=rss-c9724c48626------2</link>
            <guid isPermaLink="false">https://medium.com/p/8ef3792ee946</guid>
            <category><![CDATA[generative-ai]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[innovators-dilemma]]></category>
            <category><![CDATA[innovation]]></category>
            <category><![CDATA[software]]></category>
            <dc:creator><![CDATA[Rifki Razick]]></dc:creator>
            <pubDate>Sun, 23 Mar 2025 01:12:11 GMT</pubDate>
            <atom:updated>2025-07-10T04:11:59.224Z</atom:updated>
            <content:encoded><![CDATA[<h3>The Innovator’s Dilemma Amplified in an AI-first Future</h3><p>Large companies always had a problem. Even when they were aware of early disruption in their target markets, they couldn’t help but continue improving existing products and carry on extracting higher profits. As the late Clay Christensen observed in “The Innovator’s Dilemma,” organizational structures and processes of these big players were geared purely for <strong>optimizing the status quo</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*pAJiz69HsgLT5Sjeh01JxA.jpeg" /></figure><p>In the meantime, nimbler startups kept experimenting and innovating in tiny segments of the market that weren’t relevant to the big players. They were able to do so due to the flexibility (or even lack) of processes and rigid organizational structures. They also didn’t have the burden of satisfying all customers all the time.</p><p>For bigger firms, allocating management focus and resources to try out disruptive innovations did not make economic sense on the surface. But does that mean large players should totally forget about disruption? Definitely not. One possible approach would be to become early investors in potential disruptors with perhaps plans for more involved future partnerships or even a merger/acquisition.</p><p>In such a scenario the larger firm should ensure that the smaller firm (startup) can operate independently with its innovation game and failure is not frowned upon. Consider how Microsoft’s investment in OpenAI allowed the startup to develop generative AI technologies while maintaining its innovative edge.</p><p>Another approach for large companies is to establish focused “innovation labs” — essentially autonomous “startups” within the organization that become breeding grounds for disruptive ideas. The key is not treating these as throwaway experiments but funding and managing them properly with clear intention of market disruption. Google X, established by Google as a separate innovation unit within the company is a classic example. X operates with significant autonomy and has its own leadership structure, including a dedicated CEO</p><p>The recent rise of generative AI represents precisely the kind of disruptive force Christensen described — initially serving niche markets but rapidly evolving toward mainstream applications. For incumbents across industries, the choice is clear: actively participate in this disruption through strategic investments, acquisitions, or internal innovation labs, or risk obsolescence. The most successful organizations will be those that balance optimization of their core business while simultaneously cultivating the disruptive innovations that will define tomorrow’s markets. In this accelerating landscape of technological change, standing still is no longer an option.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8ef3792ee946" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Product strip-down meetings]]></title>
            <link>https://medium.com/@rifki_razick/product-strip-down-meetings-ef19d29dcba5?source=rss-c9724c48626------2</link>
            <guid isPermaLink="false">https://medium.com/p/ef19d29dcba5</guid>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[product-development]]></category>
            <dc:creator><![CDATA[Rifki Razick]]></dc:creator>
            <pubDate>Sun, 26 Jul 2015 05:21:02 GMT</pubDate>
            <atom:updated>2025-07-13T06:24:35.461Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/561/1*fY6-0a5wuWjlLrEOXDFkKg.jpeg" /></figure><p>Just adding new features to software products is not enough! Here’s why…</p><p>When working with software products over a long time I’ve seen that development teams are pretty much focused on only adding shiny new features. Software products tend to “bloat” overtime with features that are seldom used by the end user.</p><p>Although such features provide almost no value externally it comes at a huge cost internally in the form of maintenance costs since the code-base may get unwieldy and complicated to work with.</p><p>There are times I have seen where such “no value/low value” features cause poor performance of the product resulting in bad UX externally as well.</p><p>I think development teams should spend a reasonable amount of time to discuss product features that are obsolete or even hardly used by end users. After analyzing the trade-offs of removing such features the team can narrow down candidates for removal.</p><p>So in addition to all the “Agile” meetings that we have today, I’d like to suggest <strong>product strip-down meetings</strong> that identify “low value” or “no value” features which can be removed from the product. Teams can set aside a reasonable amount of time to remove such features during sprints/iterations. Having a solid test suite will be of immense help for this effort.</p><p>This will not only result in an easy to maintain code-base for the development team internally, a simpler product may also provide a better UX to the end-user externally as well!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ef19d29dcba5" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>