The API of the Physical World: Sy Bohy on the Future of Connected Hardware
What if the future of software wasn't building everything from scratch, but orchestrating the physical world through a single API? In this episode of How I Grew This, Amanda and Adam sit down with Sy Bohy, Founder and CEO of Seam, to explore why IoT integration is becoming the next critical infrastructure layer for developers, how AI is democratizing automation for prosumers and small businesses, and the counterintuitive strategy of knowing when not to build in-house. Whether you're a founder navigating the AI revolution or a builder looking to compete at scale, this conversation reveals why operational knowledge and distribution are now your greatest defensibility—and why the next golden age of discovery is just beginning.
What You’ll Learn:
How to identify defensible moats in a crowded AI landscape
Why calculated early-career risk-taking compounds into founder advantage
The framework for choosing what not to build as a founder
How prosumer markets are reshaping SaaS
Why personalization at scale requires coordinating multiple devices, not just data
The non-negotiable CEO skill in the AI era
About the Guest(s):
Sy Bohy is the Founder and CEO of Seam, a software platform providing a unified API to control any IoT device across thousands of brands and models. With a Stanford background and early experience at Nest—the pioneering company that kicked off the smart home revolution—Sy brings deep expertise in IoT infrastructure, hardware-software integration, and scaling connected device ecosystems. In this episode, Sy shares insights on how AI and agentic workflows are democratizing device automation for prosumers and enterprises alike, while offering actionable advice on building defensible companies in the age of AI. His work at Seam is enabling developers and businesses worldwide to coordinate the physical world at scale, making this conversation invaluable for founders, product leaders, and anyone building the infrastructure of tomorrow's autonomous environments.
If you enjoyed this episode, make sure to subscribe, rate, and review it on Apple Podcasts, Spotify, and YouTube Podcasts. Instructions on how to do this are here.
Episode Highlights:
[00:04:10] Take Calculated Risks Early in Your Career When You Have Financial Freedom
Sy shares a transformative lesson from a Stanford professor who advised graduates to prioritize calculated risk-taking while young, rather than accepting high-paying corporate jobs that will always be available later in life. This insight is critical for founders and builders because the compounding advantage of early exposure to cutting-edge teams (like Nest's iPhone/iPod veterans) creates irreplaceable domain expertise that money cannot buy. Most early-career professionals face pressure to maximize salary, but this framework reframes that choice as a trade-off between short-term security and long-term founder advantage. The practical step is to identify high-impact teams solving important problems, even if compensation is dramatically lower than alternatives, and commit 2–4 years to mastery. Sy's decision to join Nest at $39,000/year instead of a $100k+ tech job directly enabled his later success in IoT—he developed operational knowledge and conviction about the space that competitors lacked. For ambitious builders, choosing learning over early wealth compounds into defensibility when you eventually start your own venture.
[00:11:49] Developer-First Product Strategy Captures AI-Driven Market Expansion
Sy shares that Seam initially experimented with selling to non-technical buyers but discovered the highest leverage came from building for developers, a choice that proved prescient when AI coding agents emerged and democratized programming. This insight matters because the recent 5–10x explosion in API key creation mirrors GitHub's massive growth in code commits—AI is not just a feature, it's expanding the entire addressable market of people capable of building software. Rather than chasing traditional enterprise buyers, Seam positioned itself as infrastructure for the next wave of builders, including non-technical founders and SMB operators coding their own solutions. The strategy is to build with the assumption that your primary user will soon have superhuman coding assistance, making ease-of-integration and complete API coverage the primary competitive moats. For Airbnb hosts automating their properties or small businesses orchestrating smart devices, Seam's API became essential because these operators could now write their own workflows instead of waiting for SaaS vendors. For any infrastructure company, betting on the developer audience early, before AI tools mainstream that audience, provides a massive timing advantage—you'll have market dominance when demand explosively expands.
[00:17:11] Custom Tools Beat Generic SaaS in the Age of Personalized Agentic Workflows
Sy reveals that Seam built its own CRM from scratch despite the conventional wisdom that this is "insane," because off-the-shelf tools like HubSpot are optimized for UI and manual workflows, not for AI agent coordination. This is a major shift in software economics: historically, buying was always cheaper than building, but AI has inverted that calculus for companies willing to sacrifice traditional UIs for agent-native systems. The challenge Seam faced is that agentic workflows need data organized for machine efficiency, not human navigation—and typical commercial CRMs struggle to accommodate this. The practical implementation is to identify where your business processes are fundamentally at odds with how SaaS products structure data and workflows, then build a minimal custom layer (database + integrations) that lets AI agents operate effectively. Sy notes this requires a debugging mindset because it demands continuous refinement, boundary setting, and playbook documentation for the agents. For founders with technical depth, the ROI is massive: a custom internal tool optimized for agents outperforms fighting against a generic platform's constraints. This signals a coming wave where every scaling company will maintain hybrid stacks—generic SaaS for some functions, custom agent-native tools for competitive operations.
[00:22:54] Build Only What Creates Incompressible Advantage; Outsource Everything Else
Sy introduces the "incompressible advantage" framework—the idea that some knowledge (operational playbooks, edge-case handling, real-world reliability) cannot be accelerated by AI or shortcuts, while other work can be outsourced to best-in-class providers. This is essential for founders in the AI era because competitive pressure tempts every company to build everything in-house, which disperses focus and slows progress on defensible work. The mistake Sy sees repeatedly is SaaS companies spending months trying to build custom hardware integrations or webhook infrastructure when they should be investing in their core, irreplaceable capability. The practical application is to audit your roadmap and identify the 1–2 areas where you have an unfair advantage, then ruthlessly outsource or pay for everything else (payments via Stripe, webhooks via an infrastructure provider, etc.). For Seam, that incompressible advantage is four years of operational knowledge about device behavior, manufacturer relationships, and edge-case handling—something no competitor can build quickly. This principle directly applies to any startup: focus your scarcest resource (founder time) on what only you can build, not on work that commoditized providers already solve better and cheaper.
[00:25:10] Personalization at Scale Requires Coordinating Multiple Devices, Not Just Data
Sy reframes personalization beyond digital-only thinking: the future is coordinating dozens of physical devices (thermostats, locks, lights, speakers, appliances) to create frictionless, preference-aware experiences. This is critical for founders because most personalization strategies focus solely on data (recommendation algorithms, segmentation) while ignoring the physical layer where real-world experiences actually happen. In hospitality, the difference between a generic room and a personalized one is whether the temperature, music, lighting, and entertainment are pre-configured to individual guest preferences—something only possible by orchestrating IoT devices seamlessly. The practical implementation is to track user identities across physical spaces and attach preference profiles (Seam introduced this via user identity objects), then use automation to synchronize devices at arrival and reset them at departure. Sy notes that by mapping things like target sleep temperature, the environment can reshape itself automatically, delivering an experience that feels like magic. For marketers and operators, the insight is that true differentiation increasingly lives in physical experience coordination, not in traditional marketing messages. As AI enables businesses to automate device orchestration without custom development, the competitive moat shifts from "can we build this" to "did we build this"—making early adoption of coordinated device platforms a strategic imperative.
[00:29:14] CEOs Must Stay Deeply Technical and Build Constantly, Especially in the AI Era
Sy emphasizes that he still pushes 50–60 pull requests per day as CEO, citing Paul Graham's insight that a CEO slightly too deep in building is vastly better than one who isn't, because deep technical involvement is the only way to maintain conviction about what's possible and what's coming. This is non-negotiable in the AI era because the pace of capability change is so rapid that strategy divorced from hands-on building becomes obsolete. The challenge is that executive pressure pulls founders toward management and hands-off strategy, creating a knowledge gap where leaders make poor decisions because they've lost touch with technical reality. The practical guidance is to carve out working hours for coding, testing, and building, not just delegation—this maintains the conviction needed to make high-leverage bets about where the market is moving. For Seam, Sy's constant coding means he experiences device integration problems firsthand, understands which competitor moves matter, and identifies opportunities in real time rather than through filtered reports. For any founder building in fast-moving spaces like IoT or AI, losing touch with technical execution is equivalent to losing market awareness—you risk becoming a manager instead of a builder, and you will be outpaced by founders who remain in the code.
Episode Resources:
Sy Bohy on LinkedIn
Seam on LinkedIn
Seam Website
Amanda Vandiver on LinkedIn
Adam Landis on LinkedIn
Branch on LinkedIn
Branch Website
How I Grew This on Apple Podcasts
How I Grew This on Spotify
How I Grew This on Simplecast
