Robert Gonzalez

I've been away from LinkedIn for a while. Personal time off, some changes at work, and a few genuine unblocks with AI tools — all of it added up, and I just haven't had much time to spend here. I never expected to see almost 300 unread notifications. Yet here we are. If you've messaged me in the last month, I probably haven't seen it. I'll be catching up soon. In the meantime, I want to pick your collective brain. As I move toward a more automated workday, I'm getting through my regular work faster than ever. But I'm also so energized by it that I'm pouring nearly all my waking hours into moving fast. It's a strange trade-off. So, my question: as you do more and produce more with these tools, are you spending 𝘮𝘰𝘳𝘦 time with them, not less? And are you still finding time for yourself in the middle of it all?

I had a conversation yesterday I haven't been able to stop thinking about. I was talking with my colleague in what was an interview format and when she wrapped up her questions, she turned it over to me. Did I have anything for her? I had one question. 𝘏𝘰𝘸 𝘢𝘳𝘦 𝘺𝘰𝘶? Not "how are you doing at work." Not "how are you holding up." Just — how are 𝘺𝘰𝘶, the person, doing right now. It brought her to tears. And I think it brought her to tears because, in the noise of meetings and 1:1s and standups and roadmap reviews, nobody had asked her that — honestly — in a long time. It was in that moment that something clicked for me: I used to open many of my 1:1 meetings with my staff with that exact question. The human version, not the work version. Somewhere along the way, I stopped. The calendar got full. The work got loud. And I let one of the most important things I do as a leader quietly fall off the list. 𝗘𝗳𝗳𝗲𝗰𝘁𝗶𝘃𝗲 𝗽𝗲𝗼𝗽𝗹𝗲 𝗹𝗲𝗮𝗱𝗲𝗿𝘀𝗵𝗶𝗽 𝘀𝘁𝗮𝗿𝘁𝘀 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗽𝗲𝗼𝗽𝗹𝗲 𝘆𝗼𝘂'𝗿𝗲 𝗹𝗲𝗮𝗱𝗶𝗻𝗴. Not with the strategy. Not the deliverables. The people. I'm an empath. The folks on my teams have never been an abstraction to me — they're whole humans with whole lives, and that shapes how I show up. But empathy doesn't have to be innate to be real. If it doesn't come naturally, the path is intention: put the human in front of the role, ask the human question first, and mean the answer. The work matters. Of course it does. But when people feel seen, the work gets better — not as a side effect, as a direct result. This matters more now than it ever has. We are leading through a period of extraordinary change — in our tools, our industries, and the very pace of our work — and that change is taking a real toll on the people we lead. The human question is one of the few things in our day that cannot be automated, abstracted, or delegated. It has to come from us, and it has to be real. So I am reclaiming the habit. I am going to open my 1:1 meetings with the human question, I am going to mean it when I ask, and I am going to give the answer the same attention I would give the most important item on my agenda — because in that moment, that is exactly what it is. When was the last time someone asked 𝘺𝘰𝘶, honestly, how you were doing? When was the last time you asked someone on your team how 𝘵𝘩𝘦𝘺 were doing?

I've said it before, I'll keep saying it: context is king. Jugal Anchalia nails what I've been talking about for a while now. Your UI surfaces a slice of the data and patterns that drive your business. The meaning underneath — why this field maps to that one, why this exception exists, why this customer gets handled differently — lives in a context layer most companies have never seriously invested in. Without it, you're missing the insights, recommendations, and gotchas that turn data into decisions. If you want to win, invest in the context layer with the same rigor you put into the presentation layer. In the next 12–24 months, the presentation layer is going to be commoditized. What's left standing is the institutional knowledge you've captured — or the knowledge you've let walk out the door.

Cold LinkedIn outreach almost never lands with me. But a recent message absolutely did. Ellie Daw reached out a couple of weeks ago to talk about the future of engineering in the AI age. Twenty-five minutes into our call yesterday I had two thoughts running in parallel: 𝗳𝗶𝗻𝗮𝗹𝗹𝘆, 𝘀𝗼𝗺𝗲𝗼𝗻𝗲 𝘄𝗵𝗼 𝗴𝗲𝘁𝘀 𝗶𝘁. And — more tech leaders need to be talking to her. If you run a software company and you are still measuring your AI investment by how fast it ships code, stop reading this and go find Ellie at gjalla. The rest of this post is for people who already know that is the wrong yardstick and want to know what the right one looks like. Here is what she has built that most leaders have not: A view of the AI-native SDLC where 𝘁𝗵𝗲 𝗼𝗿𝗰𝗵𝗲𝘀𝘁𝗿𝗮𝘁𝗶𝗼𝗻 𝗹𝗮𝘆𝗲𝗿 𝗶𝘀 𝘁𝗵𝗲 𝗽𝗿𝗼𝗱𝘂𝗰𝘁. Every step in the workflow carries its own specs, guardrails, and attestations. The "review" we used to hand to a senior engineer is replaced by a chain of attestations that, taken together, become the sign-off. And the engineering org is not the only thing being accelerated — because she sees, clearly, that accelerating engineering alone just relocates the bottleneck into GTM, customer success, and ops. I have been evangelizing the idea of a Product Development Orchestrator for almost a year. I have been describing a tiered context layer where every handoff produces an attestation that compounds into the final delivery. I have been telling anyone who will listen that AI-native is a business-wide reset, not an engineering-only one. 𝗘𝗹𝗹𝗶𝗲 𝗵𝗮𝘀 𝗯𝗲𝗲𝗻 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗶𝘁. If you are a CTO, a VP of Engineering, or a software founder and you still cannot articulate where your AI value lives outside the IDE, you have a gap. Ellie can help you see it before it costs you. And even if you walk away from a single conversation with her without doing anything else — your thinking will be sharper for it. Go talk to her... ASAP. gjalla

I got a one shot! Pinpoint #735 | 1 guess 1️⃣ | 100% match 📌 lnkd.in/pinpoint.

A few weeks ago I posted that 𝗻𝗼𝗯𝗼𝗱𝘆 𝗰𝗮𝗿𝗲𝘀 𝗮𝗯𝗼𝘂𝘁 𝘆𝗼𝘂𝗿 𝗯𝘂𝗶𝗹𝗱. The takeaway: apps are cheap, disposable, and the durable play is contextual intelligence inside the tools people are already married to. Today I want to back that up with receipts. In the last three months, here are the applications I have practically stopped opening: - Jira - Confluence - GitHub - Jellyfish - VS Code - IntelliJ - PowerPoint - Excel - Word - To-Do - Notion Eleven apps. All replaced by Claude — either through Cowork projects reading from and writing to the outputs I used to open those apps for, or through Claude Code spinning up small, short-lived interfaces that surface exactly what I need in the moment I need it. If it hasn't yet, I hope this clicks for you: 𝗔𝗽𝗽𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 𝗮𝗿𝗲 𝗯𝘂𝗶𝗹𝘁 𝗳𝗼𝗿 𝗮 𝗴𝗲𝗻𝗲𝗿𝗮𝗹𝗶𝘇𝗲𝗱 𝘂𝘀𝗲 𝗰𝗮𝘀𝗲. 𝗠𝘆 𝗔𝗜 𝘁𝗼𝗼𝗹𝘀 𝘁𝗮𝗶𝗹𝗼𝗿 𝘁𝗵𝗲𝗺𝘀𝗲𝗹𝘃𝗲𝘀 𝘁𝗼 𝗺𝘆 𝘀𝗽𝗲𝗰𝗶𝗳𝗶𝗰 𝘂𝘀𝗲 𝗰𝗮𝘀𝗲 — 𝗮𝘁 𝘁𝗵𝗲 𝗺𝗼𝗺𝗲𝗻𝘁 𝗼𝗳 𝗻𝗲𝗲𝗱. That asymmetry is not a feature gap. It is a category shift. While every SaaS company on earth is bolting "AI" onto their product line, AI tools are quietly eating their lunch. And the SaaS companies who actually survive this are not the ones shipping more UI. They are the ones opening their context graphs and knowledge graphs to agents — so an AI tool can pull exactly what is needed, exactly when it is needed, in the form it is needed. 𝗬𝗼𝘂𝗿 𝗨𝗜 𝗶𝘀 𝗯𝗲𝗰𝗼𝗺𝗶𝗻𝗴 𝗮 𝗹𝗶𝗮𝗯𝗶𝗹𝗶𝘁𝘆. 𝗬𝗼𝘂𝗿 𝗱𝗮𝘁𝗮 𝗹𝗮𝘆𝗲𝗿 𝗶𝘀 𝗯𝗲𝗰𝗼𝗺𝗶𝗻𝗴 𝘁𝗵𝗲 𝗺𝗼𝗮𝘁. If context is king, your app's interface is a throne nobody needs to sit in anymore. As a builder, where does your app still beat the agent? As a user, where do you still reach for the UI? Tell me where I'm wrong.

"Staging was designed for humans moving at human speed, one change at a time. Agentic development works in parallel." That one line is the whole game right now. 𝗬𝗼𝘂 𝗰𝗮𝗻𝗻𝗼𝘁 𝗹𝗮𝘆𝗲𝗿 𝗻𝗲𝘄 𝘁𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆 𝗼𝗻𝘁𝗼 𝗼𝗹𝗱 𝗽𝗿𝗼𝗰𝗲𝘀𝘀𝗲𝘀 𝗮𝗻𝗱 𝗲𝘅𝗽𝗲𝗰𝘁 𝗶𝘁 𝘁𝗼 𝗵𝗼𝗹𝗱. It's like putting new parts on an old car — upgrade the engine, and the transmission can't keep up. Upgrade the transmission, and the brakes become the weak link. Every time you unlock one piece, the rest of the system reveals where it was always going to fail. We are watching this play out in real time. First it was code review. Coding agents started producing more code than any human could reasonably review. So review agents picked up the slack, and the pipeline flowed again — straight into staging. Now staging is the bottleneck. 𝗪𝗵𝗮𝘁 𝗰𝗼𝗺𝗲𝘀 𝗮𝗳𝘁𝗲𝗿 𝘀𝘁𝗮𝗴𝗶𝗻𝗴? Because something will be next. Here's what I've been sitting with for a while: 𝗮𝗽𝗽𝗹𝘆𝗶𝗻𝗴 𝗔𝗜 𝘁𝗼 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 𝗶𝘀 𝗻𝗼𝘁 𝘁𝗵𝗲 𝘀𝗮𝗺𝗲 𝗮𝘀 𝗮𝗽𝗽𝗹𝘆𝗶𝗻𝗴 𝗔𝗜 𝘁𝗼 𝘁𝗵𝗲 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀. If we only accelerate engineering, we are just relocating the bottleneck — and the relocation is usually downstream, where the unlock is harder, the friction is worse, and the bottleneck is only visible because it's already costing you. If ideas still reach engineering the old way, high-performing AI-native teams get starved. If engineering ships faster but GTM, product ops, and customer success still run at human speed, work piles up downstream and the feedback loop slows to a crawl. 𝗦𝗽𝗲𝗲𝗱𝗶𝗻𝗴 𝘂𝗽 𝗼𝗻𝗲 𝗰𝗮𝗿 𝗼𝗻 𝗮 𝘀𝗹𝗼𝘄 𝗵𝗶𝗴𝗵𝘄𝗮𝘆 𝗱𝗼𝗲𝘀𝗻'𝘁 𝗺𝗼𝘃𝗲 𝘁𝗵𝗲 𝘁𝗿𝗮𝗳𝗳𝗶𝗰. 𝗜𝘁 𝗰𝗮𝘂𝘀𝗲𝘀 𝗮𝗰𝗰𝗶𝗱𝗲𝗻𝘁𝘀, 𝗯𝗼𝘁𝘁𝗹𝗲𝗻𝗲𝗰𝗸𝘀, 𝗮𝗻𝗱 𝗰𝗹𝗼𝘀𝘂𝗿𝗲𝘀. The bet most leaders are making — that faster engineering equals a faster business — is only true if the rest of the business is moving with you. Otherwise you are just building pressure against the next weakest seam. So I'm curious: 𝗵𝗼𝘄 𝗮𝗿𝗲 𝘆𝗼𝘂 𝗮𝗽𝗽𝗹𝘆𝗶𝗻𝗴 𝗔𝗜 𝗯𝗿𝗼𝗮𝗱𝗹𝘆 — not just inside engineering — to make sure you aren't creating bottlenecks, crumple zones, or complete closures across your workflow? Are you seeing these pain points in your AI-native operations today?

I just had this exact conversation. Your fancy new app is one $200 Claude Max subscription away from being replicated by a Monster-addled teenager at 3:00 AM. So what's actually worth building? Not apps. Not tech. Not data. 𝐂𝐨𝐧𝐭𝐞𝐱𝐭. The graph underneath your app has to outlive the app itself — or the business goes with it. Unless Anthropic or OpenAI shows up with eleventy billion dollars and a hunger for a new feature. In which case… congrats. See you in next month's Claude Desktop release.

Hot take incoming. Brace yourselves. Everyone is building something right now. Every company. Every entrepreneur. Every curious person with a laptop and a dream. The. Next. Big. Thing. And honestly? I'm one of them — I'm in the game as much as anyone. But a conversation with a colleague today shifted something in my thinking. 𝐘𝐨𝐮𝐫 𝐛𝐮𝐢𝐥𝐝 𝐢𝐬 𝐩𝐫𝐨𝐛𝐚𝐛𝐥𝐲 𝐚𝐦𝐚𝐳𝐢𝐧𝐠. It probably solves real problems. Opens revenue paths. Eases the daily grind. And I bet you've pictured your target users lighting up the moment they get their hands on it. Here's my hot take though: 𝐧𝐨𝐛𝐨𝐝𝐲 𝐜𝐚𝐫𝐞𝐬 𝐚𝐛𝐨𝐮𝐭 𝐲𝐨𝐮𝐫 𝐛𝐮𝐢𝐥𝐝. That's not cynicism — it's the reality of a technology landscape moving so fast that entire industries spin up, spin out, and spin down in a matter of months. If what you're building isn't wired to a survival need, something else will replace it before the year is out. I'm watching it happen in real time. 𝐁𝐮𝐭 𝐈 𝐬𝐞𝐞 𝐞𝐯𝐞𝐧 𝐦𝐨𝐫𝐞 𝐭𝐨 𝐭𝐡𝐢𝐬 𝐬𝐭𝐨𝐫𝐲 — 𝐚𝐧𝐝 𝐭𝐡𝐢𝐬 𝐢𝐬 𝐰𝐡𝐞𝐫𝐞 𝐈'𝐯𝐞 𝐩𝐥𝐚𝐧𝐭𝐞𝐝 𝐦𝐲 𝐟𝐥𝐚𝐠. Context is king. Data is the substrate of the systems we actually 𝘥𝘦𝘱𝘦𝘯𝘥 on. And the most reassuring thing I know to be true right now is this: there is more data generated and available today than at any point in human history. And tomorrow, there will be more than today. So ask yourself this: with all of that context available, why are we building apps that ask people to 𝘤𝘩𝘢𝘯𝘨𝘦 𝘵𝘩𝘦𝘪𝘳 𝘣𝘦𝘩𝘢𝘷𝘪𝘰𝘳 — to leave the tools they're already married to — instead of bringing relevant insights 𝘪𝘯𝘵𝘰 those tools? You built a widget that summarizes my inbox when I dump my emails into it? Cool — 𝐢𝐟 𝐈 𝐞𝐯𝐞𝐫 𝐥𝐞𝐚𝐯𝐞 𝐦𝐲 𝐢𝐧𝐛𝐨𝐱. You have a dashboard that plugs into my business software and surfaces market trends? Great — 𝐛𝐮𝐭 𝐰𝐡𝐲 𝐢𝐬𝐧'𝐭 𝐭𝐡𝐚𝐭 𝐢𝐧𝐭𝐞𝐥𝐥𝐢𝐠𝐞𝐧𝐜𝐞 𝐚𝐥𝐫𝐞𝐚𝐝𝐲 𝐥𝐢𝐯𝐢𝐧𝐠 𝐢𝐧 𝐦𝐲 𝐩𝐡𝐨𝐧𝐞? Most of us have a short list of tools we use religiously, every single day. We check them before we get out of bed. We check them before we go to sleep. We don't leave them. We don't replace them easily. 𝐀𝐩𝐩𝐥𝐢𝐜𝐚𝐭𝐢𝐨𝐧𝐬 𝐚𝐫𝐞 𝐧𝐨𝐰 𝐩𝐫𝐚𝐜𝐭𝐢𝐜𝐚𝐥𝐥𝐲 𝐟𝐫𝐞𝐞 𝐭𝐨 𝐦𝐚𝐤𝐞, 𝐟𝐫𝐞𝐞 𝐭𝐨 𝐭𝐫𝐚𝐬𝐡, 𝐚𝐧𝐝 𝐟𝐫𝐞𝐞 𝐭𝐨 𝐬𝐩𝐢𝐧 𝐮𝐩 𝐚𝐠𝐚𝐢𝐧. That's exactly why new apps feel increasingly like futility to me. What feels 𝘥𝘶𝘳𝘢𝘣𝘭𝘦 — what feels genuinely valuable — are contextual systems that surface the right insight, at the right moment, inside the tools people are already living in. No behavior change required. Am I being myopic here? Target-fixated? Utterly off base? I genuinely want to know what you think. Drop your perspective below — even if it's to set me straight.