Changing Roles in Software Development

A year ago, this role barely existed. Now I'm seeing it everywhere. I call them 𝗽𝗿𝗼𝗱𝘂𝗰𝘁 𝗯𝘂𝗶𝗹𝗱𝗲𝗿𝘀. They're not product managers. They're not engineers. They're not designers. They're something new: people who can take an idea from concept to working software, end to end, in hours rather than weeks. I became one by 𝗮𝗰𝗰𝗶𝗱𝗲𝗻𝘁. I hadn't written production code in ten years. Then I started building thredspan with AI tools. Now I plan features, write PRDs, write code, run tests, do QA, and handle basic design decisions. All as one person and none of it is hand-written. The thrill of planning a feature and seeing it fully working, demoable, in under a couple hours is hard to describe. This isn't just happening to me. Satya Nadella took over product management for Copilot himself. Not as oversight. Hands-on, in the details, running weekly meetings with engineers. Shopify now requires AI usage in prototyping. PMs are expected to build rough versions themselves. The boundaries between roles are dissolving. And yet most companies are still hiring like it's 2019. Separate job specs for product managers and engineers. Clean divisions that made sense when building software required specialised teams at every step. Those divisions are becoming expensive fictions. The companies that figure this out will move faster than their competitors. The ones that don't will wonder why their roadmaps take quarters when others ship in weeks. If you're a founder, look at your org chart. If you're a PM who can't prototype, or an engineer who doesn't think about product, the gap is going to get uncomfortable. This shift is happening faster than most people realise. 📣 Next month we will be putting out a role to hire our second Product Builder (to work with me 👋). Follow us on thredspan to be notified when new roles come up.

Okay, 14 days overdue, but I think very important and helpful: Human Vibes is published. Draws on data from my ongoing study of software engineering across 23 firms, co-written with a literal dream team: Steve Yegge, Brendan Hopper, Jon Hassell. https://lnkd.in/gAciCjcZ The punchline: AI use isn't killing software engineering—it's causing mitosis. We're seeing rapid polarization into 3 tiers: the apex (strategic AI orchestration, $250K+), the hybrid middle (engineering + domain expertise), and the shrinking automatable tail. Choose wisely. Most striking finding: developers using genAI often complete tasks faster, BUT the "comfortable middle" of routine coding is vanishing. The twist? This creates MORE interesting work, not less. Small teams of 2-3 will soon commonly ship what used to take 20 people. We name new roles emerging RIGHT NOW: Platform Designers (1/3 PM, 1/3 designer, 1/3 engineer), Agent Experts (domain experts who tune agents), and my favorite: Fleet Supervisors ("air traffic controllers for bots") - I found it in robotics 8 years ago! No LinkedIn categories yet. Here's what we think actually matters for thriving: Code scrutiny velocity (10x more reading than writing now), productive skepticism (our "rule of three" for AI validation), and what we call "optimal delegation"—knowing when to give stretch assignments to AI vs humans. Personal revelation from the work: Your unconnected domain expertise is your NEW superpower. That tax attorney who learned prompt engineering and is conversant w code? They're now irreplaceable. The message is clear: AI amplifies deep human skill, it doesn't replace it. For those asking "what should I do?"—we provide specific action checklists by career stage. But the meta-lesson: continuous learning isn't optional anymore. The tools are free, documentation infinite, communities welcoming. The key scarce resource? Insight * agency. Read this doc, then go get it! I'm assigning it to my master's students this fall... Steve is at Sourcegraph, writes prescient hot takes on SWE. Written 1m+ lines of code, read 2m. Brendan is CIO of AI at Commonwealth Bank. Grew up a hacker. Jon is Content Director at O'Reilly pps: thank you, Gene Kim, for getting us together and publishing!

The last few days have been noisy for a reason. Tools like Claude’s coworker-style plugins didn’t just add features — they shifted expectations. They hinted at a future where software no longer waits for instructions, but anticipates work, coordinates tasks, and completes flows end to end. That rattles the IT industry because it challenges a long-held assumption: that complexity guarantees human relevance. Some of the sharpest minds in tech have been quietly preparing us for this moment. Satya Nadella has repeatedly said that the real value of AI is not intelligence alone, but its ability to reshape workflows. Jensen Huang frames AI as a new computing layer, not an app. Dario Amodei speaks about systems that act with intent, not just output text. Different voices, same signal: the unit of work is changing. Over the next five years, IT will move from “software that supports work” to “software that does the work.” Ticket handling, test creation, infra monitoring, report generation, even parts of design and architecture will compress. Not vanish — compress. What took teams will take systems plus a few sharp humans. This is where many professionals feel cornered. But the real risk is not AI replacing humans. The risk is humans staying static while the interface between clients and machines collapses into a single layer. The future role sits in the middle — not as a checker, but as a shaper. Human-in-the-loop is too small a phrase. What’s emerging is the human-in-the-judgment role. Someone who knows when to trust the system, when to override it, how to guide it, and how to explain its choices to a client who cares about outcomes, not models. Clients will not ask, “Was AI used?” They will ask, “Can I trust this result, and who stands behind it?” That “who” still matters. The professionals who thrive will do three things well. First, they will understand systems thinking — how tools connect, fail, and scale. Second, they will build deep context in domains, not just code. Third, they will act as translators between business intent and machine execution. This is not a pessimistic future. It is a narrower one. Fewer roles, yes — but sharper, more accountable, more human ones. AI may run the engines. Humans will still decide the direction. And that middle space — between client trust and machine capability — will be the most valuable seat in the room. DC*

A decade ago, the boundary between Product Management and Engineering was very clear. Product managers focused on requirements, roadmaps, customer conversations, and prioritization. Engineers focused on system design, architecture, and building software. There was some overlap, but it was thin and deliberate. That separation made sense at the time. In today’s AI-driven world, that boundary is fading fast. With modern AI tools and vibe coding workflows, getting a working POC no longer requires weeks of detailed handoffs. Ideas can move from concept to something tangible in days, sometimes hours. In the past, a typical flow looked like this. A product manager wrote a PRD. Engineers interpreted it. The first real output appeared after multiple sprints. Feedback loops were slow and expensive. Today, the workflow is very different. Using AI-assisted coding, agents, and scaffolding tools, I can explore ideas end to end. I can think through the customer journey, define feature behavior, prototype logic, and validate feasibility early. Many assumptions get tested before formal engineering cycles even begin. This is completely changing the nature of the role. Product managers are no longer limited to conceptual ownership. They are increasingly shaping solutions at a technical level. Engineers, in parallel, are deeply involved in product decisions from day one. This is how Product and Engineering roles are blending into a Product and Engineering role. From my own experience, the technical depth I can reach today in AI product work is far deeper than before. I still need to understand product vision, customer journeys, and core product management fundamentals. But I also need to engage with architecture, model behavior, orchestration patterns, and system-level tradeoffs. AI tools make this possible. They compress learning curves and shorten feedback loops, but they also raise expectations. Staying shallow is no longer an option. Looking ahead, I see the intersection of Product and Engineering growing significantly. Over time, we may end up with thinner layers of dedicated Product roles and dedicated Engineering roles, with a much larger core where both blend together. I write about #artificialintelligence | #technology | #startups | #mentoring | #leadership | #financialindependence   PS: All views are personal Vignesh Kumar

Figma just dropped a study that explains why the edges of your role keep dissolving. They surveyed 1,199 U.S. product and marketing professionals this year. ↳The results show something big is happening behind the scenes: People are doing more, faster, with less clarity around where one role ends and another begins. "As the market adapts to new tools that enable faster iteration cycles, product teams are wrestling with evolving role boundaries that bring traditional titles into question." Workflows are more fluid and cross-functional, blurring lines and speeding up delivery. ↳Role shift: ➤Most professionals now identify with ~3 roles. ➤64% wear two or more hats. ➤PMs, marketers, and project managers are skewing generalist; developers and researchers stay deep and narrow. ➤The design process? Everyone’s in it now. 56% non-designers are doing at least one design-centric task. That’s good for speed. But it creates more overlap, more ambiguity, more need for coordination. Context switching is brutal. ↳Main Takeaways: ➤ AI is the #1 driver of change (72% say AI tools reshaped their role). ➤ Time spent collaborating with AI rose from 11% → 19%. ➤ Multi-hat reality: responsibilities are up 19% YoY; only 36% identify with a single role; average scope spans 2.75 roles. ➤ 57% of developers now prototype moderately to significantly. ➤ Collaboration > handoffs: 56% report more cross-functional collaboration; teams work in parallel, not linear phases. ➤ Tool sprawl is real: product builders juggle 7.6 tool types on average. ↳Leadership's playbook: ➤ Redesign roles for generalist collaboration + specialist depth; update career ladders to match reality. ➤ Codify "AI as colleague” (prompt playbooks, sources, etc.). ➤Move from handoffs to parallel workflows on a shared stack; reduce tool friction. ➤ Use time saved to shift into high-value work (vision, roadmaps); 57% already are. ➤ Build capability: fund AI + craft upskilling; stand up guilds/rotations; keep a single source of truth (design system + shared backlog). The game has changed. This is your edge.

Is software engineering dead? No! The narrative of "AI will kill software engineering" doesn't hold up. But the job is changing — and that part is absolutely true. The role is moving up the stack. Routine code generation, boilerplate, and bug hunting are increasingly handled by AI tools like GitHub Copilot, Cursor, and Claude Code. So, what are software engineers doing now? They are being asked to do more sophisticated work, not less work. The type of tasks engineers spend more time on is shifting toward designing and architecting solutions, eliciting requirements, evaluating application performance, and engineering work on systems that operate in more complex environments. AI-related and full-stack roles are ascending — and problem-solving, communication skills, and domain knowledge remain highly valued, consistent with the idea that AI can produce code, but not necessarily handle broader software lifecycle tasks. The engineers who will thrive in the next decade are those who use AI to amplify their thinking — not those who wait to see if the tools replace them. You're not competing with AI. You're competing with engineers who know how to use AI well. Onward!! #hiringtrends #jobmarket #students #college #highschool #candogram

The product team you know today won’t exist in 2028. It’s less about replacing roles, and more about evolving them into new, hybrid ones. For example, a UX Designer may become a Strategic Experience Designer capable of leveraging AI in new ways. In today's familiar structure: → Product managers define direction → Designers shape experiences → Developers build → QA ensures quality → Architects maintain scalability AI is starting to augment all of this, but most teams are still learning how to leverage it. Two major shifts will occur by 2028. These shifts have already started and are picking up steam. 1. 𝗔𝗜-𝗔𝘂𝗴𝗺𝗲𝗻𝘁𝗲𝗱 𝗧𝗿𝗮𝗱𝗶𝘁𝗶𝗼𝗻𝗮𝗹 𝗥𝗼𝗹𝗲𝘀 → Developers with AI copilots coding in tandem → Designers exploring more options in a fraction of the time → QA leveraging agents to test + human intuition → PMs simulating market scenarios in real-time 2. 𝗘𝗺𝗲𝗿𝗴𝗶𝗻𝗴 𝗖𝗿𝗶𝘁𝗶𝗰𝗮𝗹 𝗥𝗼𝗹𝗲𝘀 → AI Product Manager (automating scheduling, tasks, and more) → AI UX Designer (potentially a Strategic Experience Designer) → Prompt Engineer → AI Ethicist / Governance experts What should leaders focus on now? • Building smaller, high-impact teams • Mastering human-AI orchestration • Making trust and transparency competitive advantages • Mapping Subject Matter Expertise (SME) knowledge systems The future product team won't just ship features. It will orchestrate both human and artificial intelligence to create value at a faster pace. The time is now to up-skill and adapt.  Start making the shift to focusing on outcomes as an indicator. Leverage AI in your daily work and learn how to create value from them. Those who do will have an edge. Any thoughts on future roles? Share below 👇 --- ♻️ Share if your team needs to see this ➕ Follow Jason Moccia more insights on product and innovation

This is Vellore Abhishek. He’s a PM. And last week, he submitted his first pull request. This is how the role of a PM is changing in an AI-native world. Here is how he went from writing specs → building end-to-end prototypes → submitting a PR! Some context - Abhishek didn’t study computer science. He did his undergrad in business/finance, and then got an MBA. He’s never written code or even tried to learn. But what he does have is curiosity, a bias for action, and a desire to play with AI tooling. Here is how last year unfolded for him - Jan–Sep 2025 He became a power user of Lovable. No more Figma mockups or long product specs - he prototyped ideas end-to-end and played with UX directly. It became his default way to communicate ideas to design and engineering. Oct–Dec 2025 Lovable felt limiting to him (which is insane in itself), so he picked up Cursor. Suddenly, his prototypes had working Salesforce integrations. When things broke, he rolled back commits using Railway. He was saying things like “let me debug the previous push” - while still not knowing how to code. He was learning how to work with AI to build. And it was frustrating sometimes. 14-hour days, no weekends, lots of rabbit holes. But he was hooked, and his prototypes became more & more useful to the engineering team. Jan 2026 A small fix was needed in the production app. Devs were busy. Abhishek made the change himself & submitted his first PR. Since then, in just the past two weeks, he’s also built and shipped an on-demand evaluation agent in Slack that analyzes user conversations in Von, reports bugs, and lets anyone tag it for an eval. Our engineers used to get pings constantly on agent performance & accuracy - now people just tag the agent and move on. I don't think “every PM should start coding”. I’m sharing this because the definition of what roles can do is changing fast and if your workflow still looks the same as it did 12 months ago, you should see how you can become AI fluent. AI is lowering the barrier to creation in a way I've never seen before. Vellore Abhishek just chose to lean in fully. We are lucky to have him as part of the team. Next, I’ll share how our designer (Aniket) went from living in Figma to now living in Cursor + Storybook. That story is just as wild.

The role of a software engineer is changing. The new job is to effectively prompt, review, and manage a team of agents. 90%+ of the code you write in 2026 should be drafted by AI. Artisanal, hand-crafted code is like manually writing assembly - it should only be necessary in rare circumstances. Whenever possible aim for agent-in-the-loop (rather than human-in-the-loop) debugging: give an agent a way to reproduce the issue and let it figure out and verify the solution. If you find yourself spinning wheels, figure out a way to formulate the problem in a way an agent can understand (e.g. MCP servers). Part of this new development process means constantly tweaking and improving a shared library of prompts. Update your team prompt library almost daily - if you run into an issue, add or improve the shared prompts so that others can benefit from it. Build separate prompts for planning, implementing, and reviewing code, and make sure you are clear of the purpose for each prompt. 2026 is going to be a wild ride.

🎯 The Developer Is Now The Orchestra Conductor Four weeks ago, as I became familiar with Claude Code and adopted it as the coding assistant of choice, I came to realize that its evolution would fundamentally shift my role from hands-on-keyboard pair-programmer to agent manager. Possibly, orchestra conductor. This week, July 25 proved that prediction right—Anthropic's official sub-agents launch just made multi-agent development workflows production-ready … almost overnight. 🔧 What I'm seeing in practice: The DEVELOPER → REVIEWER → VERIFIER → GIT-MANAGER process of development workspace compliance I've been refining is now officially supported. Instead of co-authoring code, I'm designing agent personalities. ⚡ The technical breakthrough: Separate context windows per agent have solved the coordination nightmare. • No more context pollution • No more community workarounds • Just clean, specialized AI teams working in parallel 💡 Here's what most miss: This isn't about replacing developers—it's about elevating the developer who can think like an architect and manage the development process. I spend my time now on: ▶ Architecture decisions ▶ Quality gates ▶ Strategic orchestration Meanwhile, my agent fleet handles implementation details. The cognitive load has shifted from syntax to systems thinking. 📊 Real numbers: Anthropic's own teams process hundreds of code additions in minutes using specialized sub-agents. Their dev teams run autonomous loops—code, test, iterate—with human oversight at commit points. 🎯 The nuanced reality: Human involvement is still critical. Someone needs to design the agent personalities, manage the handoffs, and maintain quality standards. That someone is the developer who understands both code and coordination. We're not coding less; we're architecting more. The future belongs to developers who master agent orchestration, not those clinging to individual contribution. Lest anyone consider this a slight on the incredible, cutting-edge work of Reuven Cohen, let me counter that sustained success delivering production code using frameworks like claude-flow, requires the kind of depth of knowledge, experience and skills he and others like Adrian Cockcroft bring to the party. 🔮 What's next?: Within months, job descriptions will shift from "senior developer" to "senior agent-based development manager." The question isn't whether you can code — it's whether you can think in terms of design patterns and architecture, then incorporate your skills in agent management for high-speed software development. Are you ready to put down the keyboard and pick up the conductor's baton? 🎼 #ArtificialIntelligence #TechLeadership #SoftwareDevelopment #SoftwareDevelopment #MultiAgentSystems