Articles by Derek
-
Why EQ and IQ Alone May Not Be Enough
Why EQ and IQ Alone May Not Be Enough
Since coming to Credera I’ve had the opportunity to reach back into my network and re-connect with former clients and…
1 Comment
Activity
2K followers
-
Derek Knudsen shared thisAlways a good reminderHarness vs Loop vs Graph engineering; Your Transformation Has Too Many KPIs!Harness vs Loop vs Graph engineering; Your Transformation Has Too Many KPIs!
-
Derek Knudsen shared thisIs AI Hollowing Out Our Ability to Reason? I’ve been reflecting on Addy Osmani piece regarding Comprehension Debt—the widening gap between the systems we produce and the logic we actually understand. Addy describes human code reviews as a "productive bottleneck" for comprehension. I’ll be the skeptic here: In reality, many reviews are already surface-level or biased "rubber-stamping" exercises. The bottleneck was already leaky. The real danger? AI doesn't just bypass this gate; it obliterates it. To his point, when we can generate output faster than we can critically audit it, we trade systemic correctness for surface correctness. We stop even trying to understand because the machines make it look "good enough." This isn't just a developer problem—it's a biological one. The human brain is evolutionarily wired for cognitive ease; we naturally gravitate toward the path of least resistance to conserve metabolic energy. When an AI presents a solution that looks correct, our brains are biologically incentivized to stop interrogating and simply accept the "fluent" answer. The bottom line? Understanding complex, difficult things is a practiced skill. By defaulting to AI-driven shortcuts, we aren't just losing track of our "codebases" — we are atrophying our neural plasticity for deep reasoning. If we optimize for volume over depth, we lose the very ability to comprehend hard things. https://lnkd.in/gd48tCZCComprehension Debt - the hidden cost of AI generated code.Comprehension Debt - the hidden cost of AI generated code.
-
Derek Knudsen posted thisEveryone is talking about speed in generative AI. Almost no one is talking about the user. A lot of what you’re hearing right now about generative AI in software is centered on velocity. “We’re shipping faster than we ever have!” “Our ProdEng teams are 10x more productive!” “We can finally build out all of those ideas that have been sitting in our backlog!” Sure, the narrative is directionally true. Generative tools ARE removing real constraints in the SDLC. But it’s also incomplete. Because it largely ignores where the constraint is actually moving — to the user. In SaaS, feature adoption has always been a challenge. Multiple studies over the years (Pendo, Standish Group, Productboard) have consistently shown that somewhere in the range of 60–80% of features are rarely or never used. That reality is one of the reasons movements like Lean Startup and Lean Product gained traction in the first place and I have been drinking that Kool-Aid for years. They weren’t about building faster. They were about building LESS — but with more precision. Because more features didn’t equal more value (despite what we like to tell ourselves). If anything, in B2B environments, the problem is even more pronounced. It’s not uncommon for customers to ask us (vendors) to slow down. At Notarize, we made the decision at one point to intentionally pace feature delivery because customers weren’t struggling with a lack of functionality—they were struggling to fully adopt what already existed. In the last year alone, I’ve had multiple enterprise customers ask for the same thing in different ways: “I know you’re excited about what’s new, but can you help us get more value out of what we already bought first?” Generative AI now changes the supply side of software in a potentially masive way. It dramatically reduces the cost and time required to build. What it does not change is the demand side. Users don’t suddenly gain more time. They don’t absorb new workflows faster. They don’t continuously relearn products without friction. So the constraint is shifting. Historically, it lived primarily in engineering capacity. Now it’s moving in two directions: - Upstream to product judgment and prioritization - Downstream to user adoption and value realization Which 1) makes product judgment a first-order problem - Generative AI doesn’t just increase output but also amplifies the consequences of poor prioritization - and 2) reframes the opportunity. The question is no longer: “How much can we build?” It’s: “How much can our users actually absorb and turn into value?” Generative AI will absolutely allow us to ship more — and it already has. But unless we account for the user as the limiting factor, we risk accelerating output while diluting impact. And we’ll find ourselves in a familiar place— Shipping more features…that fewer people use.
-
Derek Knudsen shared thisLove what Ines Lourenço did here - building a full operating environment specific to your roles "work to be done". Just makes so much sense versus a bunch of one off skills/plugins/etc.Derek Knudsen shared thisStop prompting. Start building. I made a single file that makes Claude build your entire PM operating system — 38 files, 6 AI skills, 7 frameworks — in 2 minutes. Claude Cowork isn't a chatbot — it's an 𝗲𝗻𝘃𝗶𝗿𝗼𝗻𝗺𝗲𝗻𝘁. So I built a "PM Workspace Wizard" that makes Claude build its own operating system for you. And I'm giving it away for free. Here's how it works ↓ → Download the .skill file from my Featured section 📌 → Open Claude Desktop, switch to Cowork mode → Drop the file into your skills folder → Select a folder for your workspace → Type "set up my PM workspace" → Answer 6 clickable questions about your context → Wait ~2 minutes while the Wizard builds everything One file in, 𝟯𝟴 𝗳𝗶𝗹𝗲𝘀 out. What it builds ↓ → 38 files across 25 folders — a complete PM system → 6 auto-loading skills (PRD Writer, Socratic Challenger...) → 9 templates + 7 frameworks (incl. B2C behavioral psych) → 5 configured workflows — Weekly Reviews, Discovery Sprints, more Once it's done, say "Write a PRD for [feature]" and the system pulls your templates, your frameworks, your context. 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗰𝗮𝗹𝗹𝘆. Inspired by 🥞 Carl Vellotti & Aakash Gupta — plus years of PM muscle memory turned into files. Works for junior PMs. Senior PMs migrating from ChatGPT. Directors who want their team on one system. The .skill file is in my Featured section 📌 Grab it, drop it in your Cowork folder, and NO — you don't have to comment or DM or do anything else. Just go get it 🧙♂️ What's the most time-consuming part of your PM workflow you wish was automated?
-
Derek Knudsen shared this"I see you spending a lot of money on GenAI tools but can't see what value we are getting from them." Using GenAI tools in the workplace can be an expensive proposition. When on one platform it's not terribly difficult to track and optimize usage; when people start expanding usage across different CLI tools, IDEs, etc., it gets harder to not just track usage but ROI. While you can build custom tooling for this (I did with Claude) it is great to see Factory integrating this into their already great platform. If you haven't looked at them I strongly recommend you take a look. https://lnkd.in/gmKuKsWc
-
Derek Knudsen shared thisLegacy SDLC is dying. I am a fan of what Aakash Gupta writes regarding Product and a big fan of his posting on this subject. I have been approached by many companies trying to sell me on “agentifying” the SDLC but their vision is same old legacy process just with agentic capabilities integrated in the seams. I think there is a market for that now and acknowledge that the Claude team is a bit of an anomaly but the IMO the reality is this will be the new norm in a year at most. Every company should be incubating these new models and redefining their way of working. https://lnkd.in/gfmvB8NMThere's a New PM Skill. It's Called Taste at SpeedThere's a New PM Skill. It's Called Taste at Speed
-
Derek Knudsen shared thisHaving to hand curate Github Action Claude PR review manifests is a pain so was fired up about this until I read the fine print: "Code Review optimizes for depth and is more expensive than lighter-weight solutions like the Claude Code GitHub Action. Reviews are billed on token usage and generally average $15–25, scaling with PR size and complexity". https://lnkd.in/ensT3yK4
-
Derek Knudsen posted thisAgentic AI is amazing. It’s also exhausting. I've been noticing something. At the end of the day most days I was mentally wiped in a way that felt different. An eight-hour day started to feel like a twelve-hour day. When I mentioned it to a few developers, the reaction was the same: “Yeah… I’ve been feeling that too.” I figured out why. I’m a big lean process guy. One of the core ideas in lean is reducing work-in-progress. Historically, less WIP means faster flow because humans are terrible at juggling cognitive threads. Agentic tools are quietly changing that dynamic. When working with CC/Codex/etc., a lot of the workflow now looks like this: prompt… model thinks… prompt… model thinks. The human is effectively in a waiting state while the agent works. So what do we all naturally do? Start something else of course! If you’re like me, pretty soon you’ve got multiple agent sessions open. New feature idea? Sure! Architecture exploration? Why not?!? Another experiment? Of course! Two sessions become three. Then five. Then TEN! “Look how productive I am!” But there’s a cost. What we were all describing is something I’ve started calling Agentic Switching Fatigue. Context switching has always had a cost. Cognitive research shows switching tasks can reduce effectiveness by 20–40% and significantly increase mental fatigue. Agentic workflows amplify this cost. When you return to a normal task, you just reload your context. When you return to an agent session, you have to figure out what happened while you were gone. What did it change? What assumptions did it introduce? Did it drift from the original intent? What exactly am I looking at here? So the switch isn’t just rehydrating context. You’re rehydrating, interrogating, and reconstructing the state of the work. Do that across five or ten parallel sessions all day and you start to feel it. Agentic Switching Fatigue. Academically, the cleanest solution is obvious—and a little uncomfortable: remove the human from the loop entirely and let agents run end-to-end systems. But until that world fully arrives, we’re still in the loop. So here are a few things I’ve started dogfooding every day to manage the fatigue: 1. Before leaving a session, I have the agent produce a quick state summary—what changed, key decisions, and open questions. 2. I limit how many agent threads I run. Lean still applies. Just because ten agents can run doesn’t mean I should track ten contexts. I try to keep it around three. 3. I have the agent maintain notes or checkpoints so I’m not reconstructing the full state of the work every time I come back. Humans have always shown an incredible ability to adapt to new tools and ways of working. I suspect this shift to agentic workflows will be no different, and it will be fascinating to see the practices and habits that emerge.
-
Derek Knudsen shared thisI’ve spent the last few years building on these models, and this update feels less like a heel‑turn and more like "capitalism cashing in its veto". Once infra and capital commitments get this large, “we’ll stop ourselves” becomes a story the balance sheet can’t support and not just for Anthropic. Will be very interested to see how Anthropic navigates the DoW ultimatum, it's next ethical challenge. Perhaps foreshadowing of the outcome? https://lnkd.in/g8RVcHZH
-
Derek Knudsen liked thisDerek Knudsen liked thisI'm 38 years old. My kids are in the decade when they'll need me most. My career says these should be my highest-earning years. My body is changing, and my marriage is at the stage that deserves real investment, not what's left over at 9 pm. I do corporate strategy for a living, and if a client handed me a roadmap where every major initiative peaked in the same window, I'd flag it in the first meeting. But I NEVER SAW THIS COMING in my own plan until I was standing in the middle of it, overwhelmed and running on fumes to keep every ball in the air. But here's the thing about strategy... IT CAN CHANGE. Our systems were built for a family structure that no longer exists, but that doesn't mean we have to stay stuck in them. We don't have to subscribe to a life of survival. We can choose to simply look at the roadmap differently.
-
Derek Knudsen liked thisAI can write code, pass the bar exam, and fold proteins. But the thing that's actually going to change your Tuesday? Knowing which duplicate record to keep. That's the unglamorous, hour-eating, second-guessing work nobody puts in a keynote. And it's exactly what our new AI Deduplication experience solves — not just a match score, but the signals behind it, the conflicts flagged, a confidence score, and the reasoning for which record to keep. Your team still makes the call. Cleaner data. Faster decisions. Hours back in your week. The best AI isn't the flashiest thing in the room. It's the thing that makes something frustrating feel effortless. This one does. #HigherEd #AI #Element451 #AIinHigherEdDerek Knudsen liked thisDuplicate records have always been easy to flag. Knowing which one to keep? That's the hard part. After 15 years working in higher ed, our VP of Product, Eric Range, felt that pain firsthand: the hours lost manually comparing conflicting records, the second-guessing over which one was right, and the nagging sense that something still slipped through. So we built a better way. Element451's new AI-powered Deduplication experience gives teams more than a match score. It surfaces the signals behind a potential duplicate, flags conflicting information, assigns a confidence score to each recommendation, and explains its reasoning for which record to keep. Your team reviews the recommendation and stays in control of the final decision. The result? Cleaner data, faster decisions, and a whole lot less time spent untangling duplicate records. See it in action 👇 And read Eric's take on what 15 years of manual deduplication taught him about building AI you can trust: https://lnkd.in/e2JGVkFr #HigherEd #AI #HigherEdTech #DataManagement #Element451
-
Derek Knudsen liked thisDerek Knudsen liked thisHappy International Friendship Day! 🧡 We're grateful to work alongside talented people every day, and even more grateful for the friendships we've built along the way. Here's to the teammates who make work a little more fun, celebrate the wins, support each other through the challenges, and remind us that the best workplaces are built on great people.
-
Derek Knudsen liked thisDerek Knudsen liked thisWrapping up the Encoura national conference where I got to share more about Encoura Connect and the power of Element451's Bolt AI engine. Always great to share the stage with Matthew Ellis.
-
Derek Knudsen liked thisDerek Knudsen liked thisAfter a lot of thought, and more than a little internal debate, I've decided it's time to begin the next chapter. My last day in my current contractor role will be August 14, 2026. This wasn't an easy decision. The team I've had the privilege of supporting has been nothing short of exceptional. They are talented, collaborative, supportive, and genuinely enjoyable to work with. Working alongside people like this makes stepping away much harder than I ever imagined. As I look ahead, I'm embracing semi-retirement. For me, that doesn't mean I'm finished working. It simply means I'm choosing to work differently. Over the past 30 years, I've had the opportunity to lead complex programs, guide organizations through significant technology initiatives, build high-performing teams, and help solve difficult business and compliance challenges. Those experiences are something I still enjoy sharing. Going forward, I'm interested in opportunities where I can provide strategic advice, trusted guidance, and experienced consulting to organizations that need an extra set of seasoned eyes. Whether that's helping executive leadership shape strategy, advising on CMMC and NIST SP 800-171 compliance, providing program or project oversight, mentoring project managers, evaluating organizational challenges, or serving as a fractional advisor, I enjoy helping organizations navigate complexity and make informed decisions. I'm open to short-term, part-time, remote consulting engagements where experience, perspective, and practical solutions bring value. I'm not looking for another traditional 9-to-5 role. Instead, I'm looking for opportunities where I can contribute in meaningful ways while maintaining the flexibility that this next stage of life offers. To everyone I've had the privilege of working with throughout my career, thank you. Your trust, partnership, and friendship have made this an incredibly rewarding journey. And to the outstanding team I've been supporting most recently, thank you for making this one of the most enjoyable chapters of my career. I'll always be grateful for the opportunity to work alongside such an exceptional group of people. Here's to the next chapter. I look forward to spending more time with family and friends, enjoying life a little more, and continuing to help organizations succeed when my experience can make a difference. If you know of an organization that could benefit from strategic consulting, executive guidance, program leadership, CMMC expertise, or an experienced advisor for a short-term engagement with part-time hours, I'd welcome the opportunity to connect. #SemiRetirement #StrategicConsulting #TrustedAdvisor #ExecutiveAdvisor #FractionalLeadership #ProjectManagement #ProgramManagement #PMO #CMMC #NIST800171 #Cybersecurity #InformationSecurity #Leadership #RemoteWork #Part-Time #Veteran
-
Derek Knudsen liked thisDerek Knudsen liked thisAnother incredible customer milestone! 🚀 Huge congratulations to Urban College of Boston and everyone on our team who helped bring this launch to life. It's always rewarding to see the impact our customers are making for future students. #HigherEd #AI
-
Derek Knudsen liked thisWe’re continuing to grow at Element451 and are looking for great people to join us! If you’re exploring your next opportunity or know someone who might be a great fit - take a look at our open roles. We’d love to connect!Derek Knudsen liked thisWe’re hiring across the board at Element451 We’re the AI-native CRM built for higher education, and we’re growing fast. That means a lot of open seats across the business: • Business Development Representative • Product Manager • Senior Campaign and Content Manager • Senior Front End Engineer • Senior Platform Engineer • Senior Software Engineer, Integrations • Quality Engineering Lead • Team Lead, Delivery Management If you want to build the future of student engagement with a team that moves fast and actually ships, I’d love to talk. Check out all open roles and apply ⬇️
-
Derek Knudsen liked thisDerek Knudsen liked thisHere is the most expensive belief about burnout: that the cure for being depleted is rest. 😴 For some people it is. For most, it is not, and they find this out the hard way. Two real weeks off, and within days of being back they are sitting exactly where they started. If that has ever happened to you, you were not resting wrong. You were treating the wrong problem. #burnout #mentalhealth #highperformance #chronicoverload #capacity #productivity
Experience & Education
Recommendations received
7 people have recommended Derek
Join now to viewExplore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content