<?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 Lorinc &quot;Lorenzo&quot; Nyitrai on Medium]]></title>
        <description><![CDATA[Stories by Lorinc &quot;Lorenzo&quot; Nyitrai on Medium]]></description>
        <link>https://medium.com/@lorinc?source=rss-32b4c99e2825------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*X-vuH-JIQtkNofJHY-Sp9A.jpeg</url>
            <title>Stories by Lorinc &amp;quot;Lorenzo&amp;quot; Nyitrai on Medium</title>
            <link>https://medium.com/@lorinc?source=rss-32b4c99e2825------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 02:20:12 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@lorinc/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[Architecture of a real chatbot]]></title>
            <link>https://lorinc.medium.com/architecture-of-a-real-chatbot-37d8849ffd5a?source=rss-32b4c99e2825------2</link>
            <guid isPermaLink="false">https://medium.com/p/37d8849ffd5a</guid>
            <dc:creator><![CDATA[Lorinc "Lorenzo" Nyitrai]]></dc:creator>
            <pubDate>Wed, 13 May 2026 08:19:49 GMT</pubDate>
            <atom:updated>2026-05-13T08:49:01.826Z</atom:updated>
            <content:encoded><![CDATA[<h3>Expectation</h3><p>Add documents to any commercial LLM, publish on a website.</p><h3>Reality</h3><pre>**Channel**<br>  - Web<br>    - Adapter                        -- channel-specific bot behavior<br>    - Auth                           -- OAuth, Auth0, etc<br>    - Domain / SSL mgmt              -- looks irrelevant until it fails<br>    - UI<br>      - UX, GUI                      -- the best &quot;design&quot; is unnoticed<br>      - Progress streaming           -- elevator music, can get important<br>  - Admin<br>    - Billing                        -- zero friction, recurring, compliant<br>    - Alerts and Analytics           -- alerts, question and feedback stats<br>    - Settings                       -- roles, labels, style settings<br>    - SaaS admin                     -- alerts, performance, backups<br>  - IM (planned)<br>    - Adapter                        -- channel-specific bot behavior<br>    - Auth                           -- depends on the service provider<br>    - UI<br>      - UX, GUI                      -- the best &quot;design&quot; is unnoticed<br>      - Voice Interface (TTS/STT)    -- PaaS or edge computing<br><br>**Harness**<br>  - Security                         -- must have, despite auth<br>    - Throttling                     -- limiting requests / second<br>    - Sanitization                   -- preventing code injection<br>  - Cost control                     -- limiting token usage<br>  - Roles                            -- auth &gt; role &gt; data (enforced)<br><br>**Orchestrator**<br>  - User prompt evaluation           -- relevant? in scope? specific? ...<br>  - User prompt rewrite              -- Massive answer quality lift<br>  - Intent routing                   -- makes the bot &quot;feel&quot; smart<br>  - Tool use logic                   -- connecting calendar, booking, etc<br>  - Context management               -- feeding the whole convo back is dumb<br><br>**Knowledge**<br>  - Retrieval logic<br>    - Reasoning lookup logic         -- design choice, RAGs were too dumb<br>    - Keyword lookup logic           -- complements the index-reasoning well<br>    - Similarity lookup logic        -- needed for a future advanced feature<br><br>**Ingestion**<br>  - Integration and GUI              -- Upload &amp; manage GUI for end users<br>  - Ingestion API                    -- allowing bot integration<br>  - Source update event handling     -- source change management<br>  - Source scope and expiration mgmt -- start and end date, role-scope<br>  - Quality gates and error handling -- clear instructions, when docs fail<br>  - Source &quot;feature engineering&quot;     -- 80% of the value is created here<br>  - Corpus index build agentic loop  -- retrieval quality depends on this<br><br>**Infra**<br>  - end-to-end observability         -- OTelemetry-ish, every module streams<br>  - SaaS maturity (7 dimensions)     -- great tool to track project maturity<br>  - data isolation and security      -- LLM can not be trusted, go oldschool<br>  - scripted deployment (IaaC)       -- dev/uat/prod, regression tests<br>  - monitoring / alerting            -- all module exposes a /health enpoint<br>  - HA / disaster recovery           -- backups, recovery drills<br>  - multi-tenant enablement          -- isolated: data, services, logs, cost<br><br>**Compliance**<br>  - GDPR                             -- user data handling by the policy doc<br>  - AI Act                           -- EU is strict about what AI can do<br>  - Local tax                        -- billing and reporting compliance</pre><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=37d8849ffd5a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Working with AI is draining you]]></title>
            <link>https://blog.fastrevops.com/working-with-ai-is-draining-you-01147502d7f3?source=rss-32b4c99e2825------2</link>
            <guid isPermaLink="false">https://medium.com/p/01147502d7f3</guid>
            <category><![CDATA[ai-augmented-development]]></category>
            <category><![CDATA[burnout]]></category>
            <dc:creator><![CDATA[Lorinc "Lorenzo" Nyitrai]]></dc:creator>
            <pubDate>Tue, 24 Mar 2026 16:17:12 GMT</pubDate>
            <atom:updated>2026-03-25T10:20:38.497Z</atom:updated>
            <content:encoded><![CDATA[<h4>because you are clinging to an obsolete measure of value</h4><p>I used to be objectively awesome at what I do. But keeping up with AI feels like drinking from a firehose.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*bJctwGl3cnQkcExj.jpeg" /></figure><p>And I’m not alone. Even OpenAI founder Andrej feels the same way.</p><iframe src="https://cdn.embedly.com/widgets/media.html?type=text%2Fhtml&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;schema=twitter&amp;url=https%3A//x.com/karpathy/status/2004607146781278521%3Fs%3D20&amp;image=" width="500" height="281" frameborder="0" scrolling="no"><a href="https://medium.com/media/3db80a48e3188ea31bdd1a9cd6d095b4/href">https://medium.com/media/3db80a48e3188ea31bdd1a9cd6d095b4/href</a></iframe><p>The most annoying thing is that this shift should favor people like me. Applying new knowledge without derailing projects is my mojo.</p><p>Generalists with cross-domain expertise in QA / Product / UX, who are very technical — just not SW Engineers yet — are uniquely positioned to deliver great products with agentic coding.</p><p>So — why am I not enjoying this? <strong>Why is this so hard?</strong></p><p>Because by the time you finish a six-weeks “project,” architectural and design choices become obsolete — probably they were already outdated the day you made them. I used to recharge by looking at what I’d built with watery eyes and pride. Not anymore — the sense of “I know my shit” is gone.</p><h4>So what’s now?</h4><p>Enter Ray, who is worth $15B, and tells you EXACTLY what <strong>you must tell yourself </strong>to get back in the game: <em>“</em><strong><em>I’m a dumbsh*t,</em></strong><em> and I know only a small percentage of what I need to know”… “so </em><strong><em>my success comes from knowing how to deal with NOT KNOWING.</em></strong><em>”</em></p><p>I love this guy. You’ve got to be a retired billionaire to have the luxury of this honesty.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FFdGeJWzy6HI%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DFdGeJWzy6HI&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FFdGeJWzy6HI%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/99ee19f9d684cf9c6fb04e2c970e3841/href">https://medium.com/media/99ee19f9d684cf9c6fb04e2c970e3841/href</a></iframe><h4>What does this all mean, if you build stuff?</h4><p>If you can <strong>ship products with AI </strong>that work, meet UX, security, performance, reliability, scalability, and cost-management standards, you are awesome. You <strong><em>are</em></strong> the AI revolution.</p><p>Everyone and their dog tells you that they vibecode. And they do.</p><p>But have you seen actual products? Do you think they were secure in any sense? Or had tested disaster recovery mechanisms? Or LLM harness with managed ambiguity-thresholds? Or conversational UX design?</p><p>If you do well enough, by oldschool development / UX standards, you are ahead of 99.99% of all vibe coders. And vibe coders are less than 1% of humanity. Do the math.</p><p>Yes, you could have done better. But — actually you could not. No human can anymore.</p><blockquote>The goal post of “perfect” is accelerating beyond human capabilities. It sucks to be a perfectionist today. You might as well give up.</blockquote><h4>If you work hard and smart, you can still stay proudly good enough.</h4><p>This is how:<strong> </strong>try to build an AI development harness that you keep improving and automating. Try to incorporate “old” ideas — quality gates, observability, isolated environments, proper Git hygiene, UX, security, availability… you know the drill.</p><p>The gap between vibe coders and <a href="https://trigger.dev/blog/10-claude-code-tips-you-did-not-know">coding agent orchestrators</a> is widening.</p><p>You are not only you anymore. <strong>Your value is how you and your rig work together. </strong>This is not how we imagined the first cyborgs, but here we are.</p><p>What a sight. Hairy, fat, stressed, middle-aged and 100% uncool cyborgs.</p><p>But we ship products, while the cool kids do their podcasts.</p><h4><strong>My new identity: I ship, at AI speed, by pre-AI standards.</strong></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*22pnmvS40ec5k3AR" /><figcaption>Photo by <a href="https://unsplash.com/@barbarazandoval?utm_source=medium&amp;utm_medium=referral">Barbara Zandoval</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=01147502d7f3" width="1" height="1" alt=""><hr><p><a href="https://blog.fastrevops.com/working-with-ai-is-draining-you-01147502d7f3">Working with AI is draining you</a> was originally published in <a href="https://blog.fastrevops.com/">FastRevOps</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Step #0: Engineering Traction]]></title>
            <link>https://lorinc.medium.com/step-0-engineering-traction-83b45bec5055?source=rss-32b4c99e2825------2</link>
            <guid isPermaLink="false">https://medium.com/p/83b45bec5055</guid>
            <category><![CDATA[education]]></category>
            <category><![CDATA[operations]]></category>
            <category><![CDATA[integration]]></category>
            <category><![CDATA[llm]]></category>
            <dc:creator><![CDATA[Lorinc "Lorenzo" Nyitrai]]></dc:creator>
            <pubDate>Sun, 08 Feb 2026 22:57:47 GMT</pubDate>
            <atom:updated>2026-02-09T09:07:24.313Z</atom:updated>
            <content:encoded><![CDATA[<h3>Driving an AI project in the wild</h3><h4>Episode #1 — Building traction, onboarding Stakeholders</h4><p>I was asked to build an LLM interface for a school’s policies. Both for administrators (creation, maintenance) and for consumption (parents asking, if they missed a deadline.)</p><h3>I expect a challenging start</h3><ul><li>IT projects are hard, LLM integration into legacy tech/org is harder.</li><li>The client has probably never seen an IT project.</li><li>The client’s IT infrastructure is a black box, managed by a single person.</li><li>I do this in a barter — meaning flat, recurring, low fee → low commitment, low visibility, perceived low value, chance of explosive retrospective evaluation (i.e. “we have done THAT in 2 months??!”).</li><li>Hype vs. reality — YouTube is flooded with “build an automated business in a weekend” content, pushing expectations even further from reality.</li></ul><h3>Brainstorm, then distill. Then expand, then narrow.</h3><p>This is an awesome technique for building a framework for a problem. Works with people and AI too. Collect a broad, unfiltered swath of inputs, then look at them, and categorize them, not in one, but many ways.</p><p>This gives you a lot of different lenses to look at the problem through, and a few of these will just feel right.</p><p>I did exactly that. I called three veteran PM friends, told them my situation, and <strong>asked for guidance. Purposefully unspecified. </strong>They said the following:</p><ul><li>Do your research. Know the infrastructure, systems, capabilities well. Do not build something that someone has already done.</li><li>Do your research. Know tribal knowledge, undocumented processes, pain points, frustrations, admin chores better, than anyone else.</li><li>Know your KEY stakeholders, keep them happy. Attract stakeholders into participation with short videos showing awesome features.</li><li>Control the narrative at all times, create a framework for collaboration, that is rewarding for everyone involved.</li><li>Do not overwhelm your clients with analysis or management tasks.</li><li>Keep weekly progress demos, use mind maps to keep complexity / backlog / scope / progress digestible and obvious for everyone.</li><li>Keep a white list of features. People expect LLMs to auto-magically know and do everything — managing expectations is importance #1.</li><li>Do not expect your clients to differentiate between needs and solutions (features and JTBDs). You are the Product Manager and the Business Analyst.</li></ul><p>Looking at these awesome insights, I realized — I’m fairly confident about most of the processes, except one. The foundation.</p><h3>Getting the collaboration framework right.</h3><p>Planning in uncertainty can be overwhelming — until you apply Agile. Have generic plans and goals for the uncertain future, then be VERY specific about the next few steps.</p><h4>My broad collaboration goals</h4><p><strong>I needed this.</strong> If I just copied, what everyone else is doing, I’d have slapped the client in the face with a fat proposal document with audit and bilateral SLAs. Or jumped straight into coding. Instead, now I follow a customized, lightweight, solid framework that works with the client’s culture.</p><ul><li><strong>Promoting participation </strong>(Project Manager)<strong>:</strong> This is a fight for sustained attention. I need to draw in, then keep active, interested, involved participants from all user groups, reliably spending some time on the project.</li><li><strong>Channeling involvement </strong>(UX Designer)<strong>:</strong> I need these participants doing, what they do best: represent their user archetypes’ true needs, challenges, JTBDs and problems. Apply “<a href="https://www.momtestbook.com/">Mom Test</a>” principles.</li><li><strong>Anchoring perception </strong>(Business Analyst)<strong>:</strong> Anchoring the future is expectations management. Anchoring the past is record keeping. Everything looks trivial in retrospect without the “this was our understanding” context.</li></ul><h4>My immediate steps</h4><ul><li>Find the most active users in all user groups</li><li>Make them feel heard, seen, excited and involved</li><li>Meanwhile map the tech, processes and JTBDs well</li></ul><p>The best thing I can do for the project NOW is getting these right. And I know, these steps look trivial. But in the “you can do anything” search space, knowing, why you do, what you do is awesome and liberating.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*c78xbpppLTNZDciyLbdqjA.jpeg" /><figcaption><em>I needed a cover image, do not think too much about it.</em></figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=83b45bec5055" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How My LLM Projects Do Not Fail]]></title>
            <link>https://blog.fastrevops.com/how-my-llm-projects-do-not-fail-b3635b6804c8?source=rss-32b4c99e2825------2</link>
            <guid isPermaLink="false">https://medium.com/p/b3635b6804c8</guid>
            <category><![CDATA[lessons-learned]]></category>
            <category><![CDATA[llm-applications]]></category>
            <category><![CDATA[best-practices]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Lorinc "Lorenzo" Nyitrai]]></dc:creator>
            <pubDate>Wed, 12 Nov 2025 15:51:11 GMT</pubDate>
            <atom:updated>2025-11-13T09:15:38.842Z</atom:updated>
            <content:encoded><![CDATA[<h4><em>Don’t build a system. Build a streamlined testbed.</em></h4><p>I’m an ex-Google Technical Product Owner and have spent the last decades on Data Engineering, and the last few years on AI/ML projects.</p><p>I have learnt that neither software-development nor ML-development pipelines work on projects that utilize LLMs — especially when you need accuracy and predictability.</p><p>Foundation models, like Gemini or GPT are astronomically complex statistical models, and — by definition — any human oversight of their training is superficial. In practice, when you apply them in business, this means you are working with a black box.</p><p>That said, I love working with LLMs. They deliver stellar ROI, and not building competence today is a big strategic mistake.</p><h3>How Building a System with LLMs Is Different</h3><p>In software development, your goal is to maintain 100% control over what your software does: unit, functional, integration, end-to-end, UAT, regression tests. Laborious, but it works.</p><p>In classical machine learning, you start with a machine-learning algorithm — essentially just code. You run it, and it massages a big pile of numbers into a good-enough representation of whatever you’re modelling. You control all the training data and all the buttons and knobs of the ML algorithm. You keep improving it until it’s good enough.</p><p>With pre-trained LLMs, none of this works as-is. You also don’t have the infrastructure (or rationale) to train your own LLM. So you start with inherent uncertainty. The model is more complex than anything humanity has ever created, and it’s only good at replicating the average of what it has learnt.</p><h3>So What’s a Feasible and Practical LLM Project?</h3><p>When used within its Goldilocks zone — “autocomplete on steroids” — an LLM does a faster, more reliable job than any human could. Probably not as good as the best human you could hire, but in ROI, it beats everyone by two to four magnitudes.</p><p>And if you can express the problem you want to solve as a few steps of autocomplete, it can handle that too. Push it away from its comfort zone and start stacking LLM layers, and you create what statisticians call a volatile model — small input changes cause large output changes. And since LLMs usually deal with unpredictable inputs, you’ve got a problem.</p><figure><img alt="AI Agent is constructing Operational Dependency Graph via conversational interface" src="https://cdn-images-1.medium.com/max/1024/1*Xoit5BXuz0AIempMSck5QQ.png" /><figcaption>AI Agent is constructing Operational Dependency Graph via conversational interface</figcaption></figure><h3>Change Your Goal to Deliver LLM Projects</h3><p>As I wrote above, anyone who tries to build an LLM-based automation by walking a straight path will fail.</p><p>The only way this works is if the PO/PM and stakeholders understand that the goal is not development, but finding a good-enough combination of arbitrary inputs and settings in a practically infinite search space.</p><blockquote><strong><em>Experience helps. A ridiculously strict iteration process helps more.</em></strong></blockquote><p>Scrum isn’t enough. The software is just packaging. The real product is the data-preparation pipeline and the hyperparameter tuning. (With LLMs, that’s context management.) You need a system nimble enough to maintain velocity through daily pivots and dead ends.</p><p>The most annoying thing about an LLM project is that, at the end, when you finally find the right combination of input and tuning, you can redo the entire three-month project in a week.</p><p>That’s also the best part. Once it works, repackaging, reusing, and applying it to other domains is fast and easy. And now, with the testbed in place, you can address problems quickly and reliably.</p><h3>How to Build That Testbed</h3><ul><li><strong>Build for iteration speed and resilience.</strong> Your goal is to streamline the problem → idea → prototype → feature → UAT process while keeping everything that works and documenting everything that doesn’t. It’s not a trivial task in a high-speed environment.</li><li><strong>Test the end-to-end UX. </strong>That’s your baseline — the only real metric that matters. Build the whole testing system around this. You want to test any change in the context of the final UX. All other tests exist to give developers visibility.</li><li><strong>Build for extreme observability. </strong>You need a log entry — visible, not buried in a text file on a dev server — for every step, decision, context, input, output, and threshold contributing to your baseline UX. Miss this, and you’ll fail the project. The log is where you catch tiny misalignments and misrepresentations in your LLM steps before they compound into annoyed users.</li><li><strong>Build for modularity. </strong>Nothing works at first. It will look like it does, and then when you scale, regulate, streamline, or chain it, it will break. Be ready to attack the same problem from different angles. If you must refactor every time you realize that something promising early on can’t take you all the way, you’ll give up soon. Make fixing mistakes easy — that’s how you win.</li><li><strong>Treat data, context, and parameters as code.</strong> As I said earlier, the real product isn’t the software but how your team gets the black-box LLM to do what you want. Keep it under source control. I’d keep interim outputs there too — you’ll want to revisit them at some point. If your input is massive, that’s a challenge, but keep something that can restore that state later under source control.</li><li><strong>Track assumptions, not plans. </strong>Planning for a learning process instead of a feature factory will save you a lot of effort. Why plan architecture and class representation two epics ahead? By the time you get there, nothing will be the same — and you’ll have broken the first rule. Instead of big fat upfront planning, have a meticulously maintained <strong>assumption ledger</strong>. This is where you track the lessons. This what we assumed, this is how we tested, this is what worked, this is what did not work. We drop this, we keep this. Not only it is magnitudes easier to maintain, it is also a great springboard for onboarding new members or even telling your AI co-pilot, what do do.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b3635b6804c8" width="1" height="1" alt=""><hr><p><a href="https://blog.fastrevops.com/how-my-llm-projects-do-not-fail-b3635b6804c8">How My LLM Projects Do Not Fail</a> was originally published in <a href="https://blog.fastrevops.com/">FastRevOps</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[RevOps Expert Interview Framework]]></title>
            <link>https://blog.fastrevops.com/revops-expert-interview-framework-174b1a23634d?source=rss-32b4c99e2825------2</link>
            <guid isPermaLink="false">https://medium.com/p/174b1a23634d</guid>
            <category><![CDATA[inbound-marketing]]></category>
            <category><![CDATA[revops]]></category>
            <category><![CDATA[interview-preparation]]></category>
            <dc:creator><![CDATA[Lorinc "Lorenzo" Nyitrai]]></dc:creator>
            <pubDate>Tue, 08 Apr 2025 21:19:22 GMT</pubDate>
            <atom:updated>2025-04-25T18:15:27.178Z</atom:updated>
            <content:encoded><![CDATA[<p><em>This is the interview framework that I use to extract work experience in a structured manned from my collaborators / partners. The purpose of this framework is material sourcing for hyper-targeted inbound marketing content.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8XVxTRTY1fNW5qJbUmKmtg.png" /></figure><h3>1. <strong>Format &amp; Expectations</strong></h3><p><strong>Purpose: </strong>Begin the interview by outlining the session format and clarifying the expectations with the interviewee. Emphasize that the discussion will focus on one specific problem from their work experience, covering every step from initial challenges to final outcomes.</p><h4><strong>Key Points:</strong></h4><ul><li><strong>Explain the structure</strong>d segments (Background, Challenges, Solutions, Takeaways).</li><li><strong>Highlight the goal:</strong> to generate multiple pieces of content (manuals, playbooks) addressing specific issues for scale-up founders and C-level operatives.</li></ul><h3>2. <strong>Background</strong></h3><p><strong>Purpose: </strong>Establish context by gathering information about the specific problem encountered during the work experience. This phase should illuminate the environment, organizational context, and the stakes involved.</p><h4><strong>Key Points:</strong></h4><ul><li><strong>Discuss the interviewee’s role</strong> and their relevance to revenue operations and growth strategy.</li><li><strong>Detail the specific problem faced:</strong> what was happening in the organization, and why it was critical to address. Clarify the scale-up environment, including any unique operational or technical constraints.</li></ul><h3>3. <strong>Challenges</strong></h3><p><strong>Purpose: </strong>Delve into the specific challenges, both expected and unexpected, that surfaced during the pursuit of a solution. Focus on the complexities and potential pitfalls encountered.</p><h4><strong>Key Points:</strong></h4><ul><li>Outline the strategic, operational, and technical flaws that contributed to the problem.</li><li>Describe the hurdles faced during the initial stages of problem identification and the path-seeking process.</li><li>Discuss the dead ends or alternative approaches that were considered but ultimately not pursued, providing insight into the decision-making process.</li></ul><h3>4. <strong>Solutions</strong></h3><p><strong>Purpose: </strong>Analyze the process and reasoning behind the chosen solution. Explore the decision criteria, the implementation journey, and how various alternatives were weighed.</p><h4><strong>Key Points:</strong></h4><ul><li>Detail why a particular solution was ultimately chosen over others.</li><li>Examine the process of implementation — steps taken, modifications made along the way, and any iterative adjustments.</li><li>Identify any expected versus unexpected factors encountered during the rollout, providing a nuanced view of the solution’s development.</li></ul><h3>5. <strong>Takeaways</strong></h3><p><strong>Purpose: </strong>Extract actionable insights and key learnings that can be translated into targeted playbooks. These takeaways should provide clear, checklist-like guidance for similar challenges faced by scale-ups.</p><h4><strong>Key Points:</strong></h4><ul><li>Summarize the measurable impact of the implemented solution.</li><li>Highlight the key lessons learned, best practices, and actionable steps for others facing similar challenges.</li><li>Offer reflections on the overall strategic approach and recommendations for preempting similar problems in the future.</li></ul><h3>Appendix: Effective Questioning Techniques</h3><p>To ensure a balance between structure and flexibility during the interview, consider the following effective questioning strategies:</p><ul><li><strong>Encourage narratives </strong>and detailed responses by asking questions that require more than a yes/no answer. Focus on “how” and “why” to invite detailed explanations.</li><li><strong>Probing for Depth: </strong>Follow up with targeted inquiries to uncover the underlying reasons behind decisions. Ask for examples, clarifications, and any alternatives that were considered.</li><li><strong>Context-Specific Questions: </strong>Keep the discussion anchored to the specific problem by asking for scenarios and concrete examples related to the revenue operations challenges at scale-ups.</li><li><strong>Iterative Refinement: </strong>Remain flexible by allowing deviations based on interesting tangents. Revisit previous sections if new insights emerge that can provide further depth or clarity.</li><li><strong>Action-Oriented Questions: </strong>Focus on extracting actionable takeaways. Ask how the insights gained can be turned into step-by-step guides or checklist items for others facing similar challenges.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=174b1a23634d" width="1" height="1" alt=""><hr><p><a href="https://blog.fastrevops.com/revops-expert-interview-framework-174b1a23634d">RevOps Expert Interview Framework</a> was originally published in <a href="https://blog.fastrevops.com/">FastRevOps</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to talk / write to an audience to create perceived value]]></title>
            <link>https://blog.fastrevops.com/how-to-talk-write-to-an-audience-to-create-perceived-value-696f0f977418?source=rss-32b4c99e2825------2</link>
            <guid isPermaLink="false">https://medium.com/p/696f0f977418</guid>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[wantrepreneur]]></category>
            <category><![CDATA[communication]]></category>
            <dc:creator><![CDATA[Lorinc "Lorenzo" Nyitrai]]></dc:creator>
            <pubDate>Sat, 27 Mar 2021 00:37:33 GMT</pubDate>
            <atom:updated>2025-11-01T19:34:21.839Z</atom:updated>
            <content:encoded><![CDATA[<h4>DRAFT: This is just notes, not an article yet</h4><h4>A review of the University of Chicago Writing Program lecture</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*8Bd0RcXn21PfY93j.PNG" /></figure><ul><li>You are not communicating to transfer your ideas. You are communicating to change the ideas of the audience. It is not “you → them”, it is “them → them”.</li><li>The goal of your contribution is not to create permanent, absolute truth. It is to progress the conversation.</li><li>No one cares about you. The writing is not about you. If you show up in the text, it is to create relatedness, trust, in order to help the audience in their progressing on the subject. Still not about you.</li><li>Credibility is not coming from the amount of knowledge, but from the contribution of the progress of the field.</li><li>You are not creating novel ideas. You are to create value to the audience. These do not contradict, but are not the same either.</li><li>Creative writing that helps the writer materializing and shaping their idea should not be published. That is writing from a completely different perspective than writing in order to create perceived value.</li><li>You got to know what they doubt; what they want to understand better.</li><li>You got to know the language (code) of that community.</li><li>You got to refer repeatedly to the community to which it matters.</li><li>You got to tell the audience, that they are exceptional, but they are wrong at something. And have an argument. Not an explanation, an argument. Create contradiction, challenge the status quo.</li></ul><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FvtIzMaLkCaM%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DvtIzMaLkCaM&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FvtIzMaLkCaM%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/d5a5a2c64282ce8d7260f936392e2393/href">https://medium.com/media/d5a5a2c64282ce8d7260f936392e2393/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=696f0f977418" width="1" height="1" alt=""><hr><p><a href="https://blog.fastrevops.com/how-to-talk-write-to-an-audience-to-create-perceived-value-696f0f977418">How to talk / write to an audience to create perceived value</a> was originally published in <a href="https://blog.fastrevops.com/">FastRevOps</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[This is how talent feels like]]></title>
            <link>https://blog.fastrevops.com/this-is-how-talent-feels-like-925d40f757a4?source=rss-32b4c99e2825------2</link>
            <guid isPermaLink="false">https://medium.com/p/925d40f757a4</guid>
            <category><![CDATA[wantrepreneur]]></category>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[talent]]></category>
            <dc:creator><![CDATA[Lorinc "Lorenzo" Nyitrai]]></dc:creator>
            <pubDate>Tue, 16 Mar 2021 18:01:32 GMT</pubDate>
            <atom:updated>2025-11-01T19:32:32.773Z</atom:updated>
            <content:encoded><![CDATA[<h4><em>Step 1: Mindset and beliefs</em></h4><h3>This is how having talent feels like</h3><h4>I often tell this story to my friends, who are either not aware of their strengths or feel embarrassed about being superior in something. It shows, how simple and natural it is, and better yet — <strong>how to accept it without being shy or smug about it</strong>.</h4><p>I’ve seen an interview with a musician, who talked about his experience of working with Prince. At one point he was watching Prince building up a song alone, before the band arrived, accord to accord, in one sitting at the piano. In utter shock, this guy told Prince:</p><blockquote>“Damn man, you just made a million dollar song! How on earth do you do that?!” And Prince answered: “I do not understand, why others can not do this.”</blockquote><p>This is how talent feels like. When something feels so trivial, yet people ask you to do it for them, because they can not do it themselves. And also — this is the right attitude to wear the burden of a talent. Accept it, as is, utilize and build upon it, but stay — you know, normal.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*aHVQ3F-j_hd2Zi0F" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=925d40f757a4" width="1" height="1" alt=""><hr><p><a href="https://blog.fastrevops.com/this-is-how-talent-feels-like-925d40f757a4">This is how talent feels like</a> was originally published in <a href="https://blog.fastrevops.com/">FastRevOps</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to cope with failure]]></title>
            <link>https://blog.fastrevops.com/how-to-cope-with-failure-7d925f8c08a2?source=rss-32b4c99e2825------2</link>
            <guid isPermaLink="false">https://medium.com/p/7d925f8c08a2</guid>
            <category><![CDATA[wantrepreneur]]></category>
            <category><![CDATA[failure]]></category>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[patreon]]></category>
            <dc:creator><![CDATA[Lorinc "Lorenzo" Nyitrai]]></dc:creator>
            <pubDate>Mon, 15 Mar 2021 21:47:50 GMT</pubDate>
            <atom:updated>2025-11-01T19:28:02.061Z</atom:updated>
            <content:encoded><![CDATA[<h4>STEP 1: MINDSET AND BELIEFS</h4><h4>One more mental construct that makes or breaks the aspiring entrepreneur. Two ways of looking at the inevitable and seemingly endless series of failures. One way makes the experience agony. This is what all of us learn as kids. The other, based on science, makes the same events almost like Flow. Yet another mental block from our education. Unlearn it.</h4><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FZf5rKTCMNnU%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DZf5rKTCMNnU&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FZf5rKTCMNnU%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/52ba3d765f55658490aa5ef3b435c694/href">https://medium.com/media/52ba3d765f55658490aa5ef3b435c694/href</a></iframe><p>First — watch this talk. Look at this guy. He is the CEO of Patreon. He is as talented as a human being can be, yet, his talk is about how 99% of every idea he had (and you will have) was an unimpressive trash fire.</p><h4>Okay. The way you cope with failure is the following:</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mk64ucwdpfCNgU8HcC9A9Q.jpeg" /></figure><ul><li>You don’t. You do what Jack has been doing. You collapse, die in agony with every flop, spend months in depression, slowly crawl out of your mother’s basement, then go at it, again. And again.</li><li>You do. You accept, that <a href="https://wantrepreneur.club/meritocracy-is-a-lie-and-how-can-an-educated-person-benefit-from-this-251a75675104">success is mostly luck</a> shining at random on average talent and effort. You accept, that <a href="https://wantrepreneur.club/the-science-of-procrastination-dc4ba4318366">a failure is not YOUR failure</a>, it is the failure of an idea, and it does not affect your value as a human, capable being. You flip and pivot ideas, and enjoy the process.</li></ul><p>Now, you see, the difference between this two is “simple”. <strong>Did you fall in love with their problem, or your solution?</strong></p><ul><li>If the first, then detaching failures from your competence and self-worth and ego is easier, and you just get a smoother ride over the very same shit-hurricane.</li><li>Or you focus on your idea, and you die with it, making every wave very-very personal.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7d925f8c08a2" width="1" height="1" alt=""><hr><p><a href="https://blog.fastrevops.com/how-to-cope-with-failure-7d925f8c08a2">How to cope with failure</a> was originally published in <a href="https://blog.fastrevops.com/">FastRevOps</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Meritocracy is a lie— and how can an educated person benefit from this]]></title>
            <link>https://blog.fastrevops.com/meritocracy-is-a-lie-and-how-can-an-educated-person-benefit-from-this-251a75675104?source=rss-32b4c99e2825------2</link>
            <guid isPermaLink="false">https://medium.com/p/251a75675104</guid>
            <category><![CDATA[success]]></category>
            <category><![CDATA[luck]]></category>
            <category><![CDATA[talent]]></category>
            <category><![CDATA[wantrepreneur]]></category>
            <category><![CDATA[psychology]]></category>
            <dc:creator><![CDATA[Lorinc "Lorenzo" Nyitrai]]></dc:creator>
            <pubDate>Mon, 15 Mar 2021 21:23:58 GMT</pubDate>
            <atom:updated>2021-03-18T07:27:57.939Z</atom:updated>
            <content:encoded><![CDATA[<h4>STEP 1: MINDSET AND BELIEFS</h4><h3>Meritocracy is a lie— and how this must be a centerpiece of your mindset</h3><h4>A review of the MIT article <a href="https://www.technologyreview.com/2018/03/01/144958/if-youre-so-smart-why-arent-you-rich-turns-out-its-just-chance/">If you’re so smart, why aren’t you rich? Turns out it’s just chance.</a></h4><p>Read the article, I just highlight and reflect on a single sentence that is extremely relevant to the theme of this blog.</p><h4>“What is the most effective strategy for exploiting the role luck plays in success?”</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/391/1*aUd5YJfhZDU5QYvANYweDg.png" /><figcaption>This is how mathematicians see you and your talents.</figcaption></figure><p>Guess. If mathematical models show, that success is luck, what can an individual do? I’m serious. Use common sense. What do you do if you want a win in a game of luck? You run the game as many times as you can. Right?</p><p>This must be a fundamental principle behind your efforts. <strong>Test ideas in bulk.</strong> Make testing ideas a hobby of yours, and dedicate time, effort and money. Build an idea-testing machine out of your resources. Include friends, tools; add everything available, into the mix, and just spin ideas. Become professional, become effective, become thorough. Become extremely good at learning from these experiments.</p><blockquote>There is no more reliable way to self-made success.</blockquote><p>You will need friends or at least one friend, who understands you. You will need a relatively safe social-economic environment. You will need some money to burn, and you will need time. Time to test, time to learn, time to recover from flops.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=251a75675104" width="1" height="1" alt=""><hr><p><a href="https://blog.fastrevops.com/meritocracy-is-a-lie-and-how-can-an-educated-person-benefit-from-this-251a75675104">Meritocracy is a lie— and how can an educated person benefit from this</a> was originally published in <a href="https://blog.fastrevops.com/">FastRevOps</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The science of procrastination]]></title>
            <link>https://blog.fastrevops.com/the-science-of-procrastination-dc4ba4318366?source=rss-32b4c99e2825------2</link>
            <guid isPermaLink="false">https://medium.com/p/dc4ba4318366</guid>
            <category><![CDATA[procrastination]]></category>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[tedx]]></category>
            <category><![CDATA[motivation]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <dc:creator><![CDATA[Lorinc "Lorenzo" Nyitrai]]></dc:creator>
            <pubDate>Mon, 15 Mar 2021 18:57:04 GMT</pubDate>
            <atom:updated>2025-11-01T19:38:42.431Z</atom:updated>
            <content:encoded><![CDATA[<iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F52lZmIafep4%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D52lZmIafep4&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F52lZmIafep4%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/00cb0b0717324f3018ef7765e3d9234a/href">https://medium.com/media/00cb0b0717324f3018ef7765e3d9234a/href</a></iframe><p>This talk hit me like a truck for the first time. Now I’m not watching it anymore, because it became part of my reality.</p><p>You have no idea, how much time have I spent polishing the perfect idea, the perfect product, the perfect landing page, the perfect logo, the perfect business plan.</p><p>I was a very knowledgeable regular on Startup meetups, at one point I was leading SME workshops. Still — my attempts just faded away, in my head, without actual feedback from the market.</p><blockquote><strong>But I never really did business. Until I have watched this talk.</strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/595/0*rUCMrNe2bmVX54AU.jpg" /><figcaption>Self Worth Theory was one of Carl Rogers’ many contribution to modern psychology.</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=dc4ba4318366" width="1" height="1" alt=""><hr><p><a href="https://blog.fastrevops.com/the-science-of-procrastination-dc4ba4318366">The science of procrastination</a> was originally published in <a href="https://blog.fastrevops.com/">FastRevOps</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>