Gareth Cokell

The shine is starting to wear off vibe coding. I know, because I’ve been there. One moment I’m in Cursor, impressed by what a large language model can produce. The next, I’m staring at the clock, realising I’ve lost hours because it couldn’t navigate the complexity of the codebase I’m working on - and I find myself back in PyCharm, questioning the value of the detour. One answer that’s emerging is "Spec-First Development". You don’t toss vague instructions at an AI and hope for magic. Instead, you craft the clearest, most rigorous specification you can - and let the agent build from that. Better still, treat specifications as code artefacts: versioned, under source control, and continually refined with compacted, curated context. For me, this isn’t a radical shift. I’ve long worked with test-driven development, where the unit test is the specification. And LLMs thrive on this. Unit tests aren’t just notes in prose; they’re executable specifications - formal, computable, and unforgiving in all the right ways. That’s the deeper lesson: when you hand a spec to an LLM, the more formalisation, the better. Natural language is fine, but structure wins. Code loves tests. Maths loves notation. Business, however, has no equivalent - no unit test for strategy documents, compliance rules, or process design. Enter ontologies and knowledge graphs. They give us a way to formalise business semantics, capturing domains in rigorous, computable detail. Paired with an LLM, they don’t just guide the generation - they also validate it. The future of agentic coding - and of LLMs in business more broadly - won’t be built on “vibes.” It will be driven by how well we can formalise intent: transforming ambiguity into something structured, testable, and executable. ⭕ Spec-kit: https://lnkd.in/dPZCgUeq ⭕ What is an Ontology: https://lnkd.in/ePS7ha8z ⭕ What is a Knowledge Graph: https://lnkd.in/e5ed_f8g

394

58 Comments

He Built an Enterprise-Level Software for over a Year. Zero Users. Last week, I sat in a coffee shop with a brilliant friend who showed me something incredible. An enterprise software he'd been building for over a year. Beautifully designed. Thoughtfully architected. Genuinely solves a real problem. I asked him: "How many users do you have?" "None yet. I'm still working on it." "Does anyone know this exists?" "...No." I looked at him and said, "Your first 10 users could've come from your network. But nobody knows your solution exists."? Here's what kills me about this: We resist showing up because we're afraid of crickets. "I don't get engagement." "I don't know what to post." "Nobody cares what I have to say." Meanwhile, someone in your network has the exact problem your solution solves. The numbers don't lie: My second open-source project? The first 5 users came directly from my network. Now it has: • 26 active users • 5 countries • 426 actions taken (times someone actually used a feature) You know I love sharing receipts, but none of that happens if I stay invisible. Let's talk about what 3 likes actually means: In an attention economy where everyone is drowning in content, 3 people stopped scrolling, read your entire post, and chose to acknowledge it. That's not failure. That's three humans who gave you their most precious resource: attention. Leave metrics, focus on connections Don't know what to post? Just be yourself and tell people what you care about. The project you're building. The problem you're solving. etc You don't need profound wisdom. You need an authentic presence. Start small if you need to: → Comment thoughtfully on others' posts → Document what you're building etc Visibility isn't all-or-nothing. It's showing up consistently in whatever way feels natural. What showing up has done for me: • People recognise me from LinkedIn at events and random places • I get DMs with opportunities • The last 3 speaking gigs came from my network • Referrals I've never applied to speak at any event. Visibility did that. The voice in your head: If something inside you is saying, "I should be out there," stop ignoring it. Your ideas, your work, your perspective have value. Quieting that voice doesn't protect you. It just keeps you invisible. The cost of invisibility? Zero impact. "Your silence doesn't serve anyone, including you." Focus on value, not clout The authority comes from consistent value, not chasing viral moments. You don't need permission to be visible: You don't need thousands of followers. You don't need perfect content. You don't need to go viral. You just need to stop being invisible to the people who are already looking for what you offer. It's perfectly fine if you choose not to show up. Not everyone needs to be online, constantly sharing. But if that voice is telling you that you DO need to be out there, then get out there

472

55 Comments

Leading a technical team is VERY different to being a Senior Developer, don't you think? Reflecting on more than 30 years of hindsight, there are a few things that I wish I had understood better and learned more quickly back in the day. So I thought I would share them and maybe help others learn to be better Technical Leaders, more quickly. That's why I've put together this guide, that pulls that experience together and you can get your copy for FREE ➡️ https://lnkd.in/ekRVdi6T In my experience, the best leaders in software engineering are the servant leaders who clear obstacles, create safe environments for experimentation, and focus relentlessly on enabling their teams to succeed. Leadership is about recognising that the success of the organisation depends on the people closest to the work having the freedom and support to do their best. If you want high-performing teams, your job as a leader is to build trust, remove friction, and help brilliant engineers thrive. ❗ See more of my advice how to be a great tech leader in the guide here ➡️ https://lnkd.in/ekRVdi6T

113

5 Comments

I once had an engineering manager say to me, “Matt, do you know what you get when you hire a new engineer?” I said, “I’m hoping we get a bunch of great new features shipped. “ “No, Matt.” He said, “You get a new engineer.” He makes a fair and accurate point that it takes time for new hires to get up to speed and contribute at the level that drives the business. The problem is that the attrition rate of new employees is high and has been getting worse every year for the past five years. That means that over 30% of the time, new employees are leaving before they truly hit their stride. And then the cycle begins again. Poor retention is a significant drag on the momentum of teams. In sales, in engineering, and everywhere else. A new report published today by Revelio Labs presents a compelling argument that new hires are leaving at a higher rate than in the past, mainly due to a lack of career opportunities. To quote the report “What aspects have reportedly worsened most for early leavers amid the decline in new employee retention? Compared to 2022, when new employee retention was last stable, career opportunity ratings have dropped the most (again controlling for business outlook ratings and seniority)” BOTTOM LINE? Companies should pay more attention to talent development, which includes providing career progression clarity and skill-based training to promote growth. CONSIDER THIS... 52%+ of employees say there is little to no transparency into their career path at their organization. (Grant Thorton, State of Work in America 2024) 54% of employees felt their employer did not provide adequate opportunities to learn skills helpful to their career. (PwC - Hopes & Fears Workforce Survey 2024) ONE MORE IMPORTANT POINT The Revelio study also showed that newer employees stay longer when their more tenured colleagues are getting promoted. This correlation suggests that a culture that emphasizes internal mobility and growth improves retention for everyone. 9 in 10 employees feel stuck, and 84% say a clear career path would boost their commitment. (Dayforce/Ceridian, Pulse of Talent)

43

2 Comments

AI won't make your team faster on its own. Better developer experience will. I sat down with Laura Tacho, CTO at DX, to talk about what actually improves engineering team performance and what just looks like it does. We covered: 1. Why most metrics fail (and how to use them properly). 2. What developer surveys can tell you that GitHub stats never will. 3. How to track DevEx without a fancy tool, even at a 10-person startup. 4. The real impact of AI on engineering workflows (spoiler: it’s not replacing your team). The big takeaway? Developer performance isn’t about shipping faster. It’s about removing friction. Whether it’s bad tooling, unclear goals, or noisy metrics, the teams that win are the ones that make it easier to do great work.

42

5 Comments

Stop stuffing your Quarto repo with figure files like an overpacked suitcase... If you're storing generated figures in your Quarto project's main repository, you're creating a mess. Slow clones, bloated repos, unhappy teammates. Git submodules = lean main repo + organized assets + happy developers See it working in Quarto projects: 📂 Main repo: https://lnkd.in/gUtitFyW 🖼️ Figures repo: https://lnkd.in/gHViUFCH Implementation guide: https://lnkd.in/gzgf3tMg #quarto #git #submodules

62

3 Comments

The sequence matters: 1. Build with no-code 2. Find paying customers 3. Hit limitations 4. Then hire engineers Not: 1. Search for CTO 2. Spend 6 months pitching technical co-founders 3. Build something nobody wants 4. Wonder why it failed The market doesn't care about your tech stack. It cares if you solve a problem worth paying for.

41

10 Comments

𝗧𝗵𝗲 𝗳𝗮𝘀𝘁𝗲𝘀𝘁 𝘄𝗮𝘆 𝘁𝗼 𝘀𝘁𝗮𝗹𝗹 𝘆𝗼𝘂𝗿 𝗰𝗮𝗿𝗲𝗲𝗿 𝗶𝗻 𝗽𝗿𝗼𝗷𝗲𝗰𝘁 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁? Becoming the “yes to everything” person. It starts small: • “Can you grab notes?” • “Can you chase Jira action items?” • "Can you just put the deck together real quick?”    You’re good at it. So you become the default. And suddenly even a 𝗣𝗠𝗣-𝗰𝗲𝗿𝘁𝗶𝗳𝗶𝗲𝗱, 𝗣𝗠𝗕𝗢𝗞-𝘁𝗿𝗮𝗶𝗻𝗲𝗱, 𝗮𝗴𝗶𝗹𝗲 𝗽𝗿𝗼𝗷𝗲𝗰𝘁 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 𝗲𝘅𝗽𝗲𝗿𝘁 is reduced to human duct tape. ❌ Always helpful → Never seen as strategic ❌ Always available → Rarely trusted with impact ❌ Always busy → Invisible when it matters ✅ Protect focus → Seen as a leader who sets direction ✅ Drive outcomes → Trusted with responsibility that matters ✅ Practice 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗰 𝘃𝗶𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆 → Recognized for influence, not busyness The more time you spend in coordination mode... The less anyone sees you as a leader. And the harder you work to hold it all together... The easier it is to overlook you. You’re not duct tape. You’re a leader. The best PMs, whether they sit in a PMO, run complex agile programs, or lead multimillion-dollar initiatives,  don’t prove their worth through exhaustion. They prove it through: ✅ Systems ✅ Influence ✅ 𝗦𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗰 𝗹𝗲𝗮𝗱𝗲𝗿𝘀𝗵𝗶𝗽 𝘃𝗶𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆. Strategic PMs don’t say yes to everything. They protect their energy for the decisions that actually drive value. 𝗜𝗳 you feel stuck in the “glorified secretary” trap, ask yourself: ✅ What impact am I driving? ✅ What value am I protecting? ✅ What would a 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗰 𝗽𝗮𝗿𝘁𝗻𝗲𝗿 𝗶𝗻 𝗹𝗲𝗮𝗱𝗲𝗿𝘀𝗵𝗶𝗽 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 do here? Because visibility doesn’t come from saying yes. It comes from 𝗹𝗲𝗮𝗱𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝗰𝗹𝗮𝗿𝗶𝘁𝘆, 𝗰𝗿𝗲𝗱𝗶𝗯𝗶𝗹𝗶𝘁𝘆, 𝗮𝗻𝗱 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗰 𝘃𝗶𝘀𝗶𝗯𝗶𝗹𝗶𝘁𝘆. ____ ♻️ Repost if you are ready to take control of your career 🔔 Follow Elizabeth Dworkin for more on strategic visibility

69

64 Comments

As promised, I'm going to start covering some of the fantastic talks I attended at Commit Your Code Conference 2025! I had a fantastic time listening to Wayne Jones share his talk on advocating for accessibility. He didn't just give us the "why", he gave us the practical playbook for making it a reality in engineering teams. I learned a ton from his experience! 💡 Core message: Accessibility isn’t a feature, it’s a foundational requirement that’s both a moral and business win. 🔑 Key Takeaways for Every Engineering Team: 🔹 The “Temporarily Able” Motto: We’re all just temporarily able. Building with accessibility in mind today prepares products for the future, including our future selves. 🔹 Start Small, Build Momentum: Don’t try to fix every issue with one massive PR. Instead, pilot a fix on one high-impact component, measure the results (e.g., Lighthouse scores), and share the wins to build trust. 🔹 Combine Audit Methods: Automated tools only go so far, always pair them with manual testing using screen readers and keyboard navigation for full compliance. 🔹 Vendor Code is Not Immune: Building an accessible version from scratch is often necessary when third-party components are fundamentally inaccessible. While a tough choice, it drastically reduces long-term liability and creates a reusable, robust UI solution. 🔹 The Business Case is Solid: When pitching accessibility to stakeholders, frame it around measurable business goals: it expands market reach, reduces the risk of legal action, and improves overall customer satisfaction. Hope you find these takeaways as helpful as I did! I can’t thank you enough, Wayne Jones, for sharing such valuable insights with us! 🙏 #CYC25 #Accessibility #A11y #WebDevelopment #ReactJS #FrontendJobs #SeniorSoftwareEngineer # #Hiring #OpenToWork #JobSearch

33

4 Comments

I just helped a client find their dream developer in a place their competitors would never think to look. Last month I was doing a demo for a company trying to hire Ruby developers. They'd been posting on job boards for 3 months with zero qualified candidates. I asked them a simple question: "Do you know anyone in the Ruby community?" Turns out their CTO had been active in a LinkedIn group for Ruby developers for years, but they'd never thought to tap into it for recruiting. So we helped them reach out to the group leaders. When the group leader vouched for someone's technical abilities, that recommendation carried serious weight—way more than any resume or coding test could. Three weeks later, they had 5 qualified candidates in their pipeline. Most recruiters are sitting on goldmines and don't even realize it. We've been so focused on employee referrals that we're missing these massive professional communities where people already know and trust each other's work. Let me give you the five types of communities that have generated the best results for our clients.   • LinkedIn professional groups - Find the groups specific to your roles and get to know the leaders. They know everyone's actual capabilities because they've seen people's work in action.   • Industry specific networks - We've had healthcare clients absolutely kill it with communities like NurseDash. These professionals know each other's clinical skills and work styles in ways that go way beyond what you can see on a LinkedIn profile.   • Skill based communities on platforms like Reddit - When someone refers from a machine learning community, they're vouching for proven project experience. They've seen this person's code and watched them solve real problems.   • Geographic and local networks - Members meet face to face, collaborate on regional projects, and cultural fit comes built in because they understand the local business environment.   • Alumni networks from universities and bootcamps - These connections last forever and members understand learning styles and career paths in ways that skill assessments completely miss. Relationship strength matters way more than frequency of contact. You might not have talked to someone in 6 months, but if you worked on a meaningful project together, your recommendation still carries real weight. Most companies limit themselves to employee networks and then wonder why their referral programs fail. The companies winning right now are expanding into these broader professional communities where trust and professional credibility already exist. Bottom line is this: 60% of jobs get filled through networking anyway. Why not make it intentional and tap into communities where people actually know each other's work quality? What communities have been most valuable for your sourcing?

40

1 Comment