A Dose of Hope for the Future

Kadhir Mani

(7 minutes)

The conversation around artificial intelligence often gravitates toward apocalyptic scenarios. Are we hurtling toward societal collapse as machines replace human purpose and productivity? I've been thinking about these scenarios a lot. They reflect concerns about rapid technological change and its impact on our lives, livelihoods, and communities. But after talking to countless people working in the space, I now feel that we're asking the wrong question. I think the "AI endgame" isn't an ending at all, but rather the foundation for something unprecedented — a world where the physical frontier becomes our next great challenge, where alignment between people and AI defines what we can accomplish, and where the problems that have always mattered most finally come within reach. <h2 id='e2617a33-30e5-40d6-b4fb-84cee3d63c51'>Software as a solved problem</h2>I have a new perspective: AI is fundamentally "solving" software in much the same way that previous technological revolutions solved their respective domains. Consider electricity. Once a marvel that required specialized knowledge to generate, distribute, and utilize, it has become an invisible utility that powers nearly every aspect of modern life. We don't think about electricity as a problem to solve anymore, we simply expect it to work. The same transformation occurred with telecommunications and networking. The internet, once a complex system requiring deep technical expertise to navigate and utilize, is now a foundational layer that we take for granted, enabling everything from global commerce to instant communication. This doesn't mean innovation in these domains has ceased. Electrical grids continue to evolve with renewable energy integration and smart distribution systems. Telecommunications advances with 5G and beyond. Networking protocols improve continuously. But the fundamental problem, how to reliably deliver electricity, connect people across distances, or transmit data between computers, is solved. These technologies have become stable, predictable foundations upon which entirely new categories of innovation are built. I think that software is undergoing this same transformation through AI. When I say AI will "solve" software, I mean that the creation, maintenance, and modification of software will become automated, accessible, and reliable enough that it transitions from being a specialized craft requiring years of training into a ubiquitous utility available to anyone who needs it. <h2 id='66dba802-ac36-4c49-924b-502ad05b44e9'>The consequences</h2>Building in the digital world is becoming cheap, fast, and accessible to anyone with a problem worth solving. The most significant consequence of this, to me, is that software is losing its alpha. For decades, the complexity of building software was a moat. Big tech companies built products in ways that were nearly impossible for competitors to replicate without massive engineering teams and years of runway. That complexity premium is disappearing, and with it, the margin structure that the entire industry has taken for granted. This has a cascading effect. Goods and services that remain purely digital — software that exists only as software — face brutal pressure toward commoditization. Users will expect dramatically more for their money because they'll increasingly understand just how cheap it is to produce. But this isn't a new cycle. Every technological revolution went through this exact transition, from scarce craft to invisible foundation. Electricity ceased to be a competitive differentiator the moment it became a utility. The same happened with telecommunications, and then the internet. In each case, the complexity premium evaporated, the industry panicked, but then something remarkable happened: <callout icon="rocket" color="#3E63DD">The world built something entirely new atop what was now taken for granted.</callout>I think software is next in that lineage. And if history is any guide, the real story isn't what we're losing, it's what becomes possible once that foundation is laid. <h2 id='467d578c-8710-4572-ab6d-da9e42ea5b6d'>The next frontier</h2>When anyone can build anything digital, I think the new frontier that opens up is the physical world. Let me explain. There's a version of this story that frightens people: AI takes the jobs, and nothing fills the void. But I find myself increasingly unconvinced by that version. Capitalism has a self-correcting mechanism that gets overlooked amid the panic: you cannot sell goods and services to people with no income. The more compelling story is that we will discover something new once the constraint of "this requires ten engineers and two years" is lifted. We always find new ways to stay busy. And when you ask people what they'd build if building were free? They don't say another SaaS product. They say they want to fix climate change. They want better healthcare for their parents. They want to understand why diseases go underfunded, why food gets wasted while people go hungry, why drug discovery moves at a glacial pace when lives are on the line, etc. These are the problems that have always mattered most, and they've always lived in the physical world, often so daunting, beyond our reach. I'll be honest. My gut reaction to believing that everything will be solved by unlimited software is still hesitation. But, that's more hope than I've felt in years! I don't know if we'll "just" solve climate change, cure cancer, or eliminate preventable suffering. But for the first time, the instinctual answer for me isn't "that's too hard." If we can truly build anything in the digital world with just a thought, maybe, just maybe, some of these problems aren't as insurmountable as they once were. Something fundamental has shifted. The scales are tipping. And the next chapter of human ambition might be pointed at the problems that deserve it most. <h2 id='d3dd4744-9868-425b-a643-7a4d62402ae7'>Why I'm at ProductNow</h2>So if the next great frontier is the physical world, why am I here, building a product tool? It's a question I've sat with honestly. And the answer I keep landing on is this: almost every hard problem I've run into in my career wasn't ultimately a technical one. The hardest problem, the one that quietly kills more good ideas than any engineering challenge, is getting a group of people to agree on what to do and work together. Alignment. It sounds simple until you spend a decade watching brilliant teams fail at it. Teams working on the world's most important problems still struggle to decide what to build next week. Organizations with unlimited funding lose months to decisions that were never clearly made, to small misunderstandings that compound into major course changes. The problem isn't talent or ambition. It's that humans are bad at maintaining shared understanding at scale. And that problem is only getting harder as AI becomes part of the team. More actors, more context, more chances for things to quietly fall out of sync. To me, this is a deeply human and deeply physical problem. Alignment isn't a software bug you can patch. It requires people (with different backgrounds, incentives, and blind spots) to agree. That process is nuanced and messy. ProductNow is working on exactly this. What we're building is infrastructure for shared understanding: a place where the thinking behind decisions lives alongside the decisions themselves. Gathering signal, setting the direction, and execution stay in contact. Not through another meeting or a Slack thread that gets buried, but as a natural layer of how the work is documented and shared. What we're building is natural alignment between humans and agents to drive better outcomes, faster. <h2 id='15ff89e4-435f-48c1-be2e-4f22323ea6e4'>Teams of tomorrow</h2>The teams that tackle the hardest problems in the physical world will need to understand, to think, and to move together. <callout icon="target" color="#3E63DD">Alignment is upstream of everything. You can't cure a disease without a team that agrees on the approach. You can't build a city without people who share a vision of what it's for.</callout>This is why the moment we're in feels so exciting to me. AI is making it possible to build software that actually understands context: software that helps a team see what's been decided, why, and by whom. We've always known alignment matters. Now we might actually be able to build tools worthy of that problem. The teams of tomorrow will be mixed: people and AI working side by side, pursuing the same goals, and needing to stay in sync. But here's the uncomfortable truth: the ways we work today weren't designed for this. The rituals, the handoffs, the review cycles, the whole rhythm of how a team takes an idea from conception to shipped product. All of it was built around human pace, human communication, human judgment, and human accountability. Autonomous agents don't fit neatly into any of it. This is the challenge that keeps me up at night, and the one I'm most excited to work on. The future of team productivity isn't about plugging AI into today's processes. It's about rethinking those processes from the ground up. I want to build the tools that make that collaboration possible. Tools that help any team, however it's composed, to think clearly, decide well, and move in the same direction. <feedback data-feedbacknodeid="b4e2d3c5-f6a7-8901-bcde-f23456789012" data-title="Does this resonate with you?" data-datatype="SINGLE_SELECTION" data-metadata="{"options":["👍 Yes","🤔 Maybe","👎 No"],"type":"SINGLE_SELECTION"}"></feedback><feedback data-feedbacknodeid="7475be05-48a3-47e6-9683-717fb653f810" data-title="What are your thoughts?" data-datatype="TEXT" data-metadata="{"type":"TEXT"}" data-description="Tell us what you think!"></feedback>

Published

May 26, 2026

Keep reading

The silent tax of bad communication

Bad communication isn't a soft-skill problem — it's a quiet management failure that creates rework, duplicated effort, and meetings that lead to more meetings. Why writing things down first is the unlock for better thinking, faster alignment, and stronger memory in the AI era.

Product

Write for human download time

The real problem isn't AI — it's slop. Why human download time, the gap between encountering information and actually understanding it, is the metric that matters most now that content is cheap.

Product

Teams that coordinate effectively will win.

As AI expands execution capacity, the next bottleneck will be coordination..

ProductEngineering