When code gets cheap, look beyond the screen
Greg Isenberg, a sharp observer of the startup landscape, frames AI resistance not as immunity, but as a question of defensibility. While AI makes launching digital products easier than ever—allowing founders to "press a button and boom"—a product that is easy to build is also, inherently, easy to copy. This core tension means that true competitive advantage shifts from mere creation to uncopyable execution.
Founders should look beyond the screen, distinguishing between pixels and atoms. Software businesses, operating entirely in the digital realm, face declining margins as AI commoditizes code. In contrast, "atoms" businesses—those rooted in the physical world—still require real-world delivery. Think plumbers, HVAC technicians, hospitality services, or local repair shops; AI cannot physically fix a leak or brew a coffee.
AI can certainly optimize the back office of an atoms business, streamlining scheduling, customer service, or bookkeeping. However, founders must not mistake this automation for their core moat. Physical businesses remain exposed to labor shortages, capital expenditure, and the myriad risks of real-world execution. The value is not in the software, but in the tangible work AI still cannot perform.
Build a moat AI can’t generate: permission and trust
Permission, not pure innovation, forms a robust AI moat. While AI accelerates digital product development, it cannot instantly grant the structural advantages of licensing, regulation, or specialized credentials. Consider law firms or bookkeeping, where jurisdiction-specific educational requirements and certifications legally restrict who can operate. AI can assist with research or drafting, but it cannot sit for the bar exam or earn a CPA.
Enterprise partnerships offer a similar, if less formal, barrier. Becoming a preferred vendor for giants like Salesforce or Microsoft isn't about having a functional product; it's about established relationships, proven credibility, and years of trust. These distribution moats take time and human capital to build, making them impervious to overnight AI replication.
Founders should identify the real-world gatekeepers in their target niche. AI excels at automating research and outreach, but it cannot forge a strategic deal or earn a buyer’s confidence. Focus on cultivating the relationships and navigating the compliance essential to your business; these are the non-generative assets that will protect your venture.
Turn attention and timing into compounding assets
Attention is not merely a marketing metric; it’s a compounding asset. While AI tools like Opus 5.5 can democratize content creation—making newsletters, short videos, and "meme pages" more accessible to produce—output alone won't build a moat. A durable media business requires a distinct point of view, a specific audience, and credibility that inspires loyalty, creating an "irreplaceable trustworthy brand."
Greg Isenberg argues that the real defensibility comes from establishing a trusted niche brand, which lowers customer-acquisition friction and provides distribution for future products. This is why "people follow people," and even in an AI-saturated landscape, human-led content, amplified by AI, can still generate hundreds of thousands of views and build valuable communities. For more insights from Isenberg, visit the Greg Isenberg Official Website.
This leads to the crucial advantage of timing in emerging micro-niches. Monitoring trends with tools like Grokbot or Claude Duo allows founders to be first movers. The initial product might be easily copied, but the enduring value lies in the audience, insights, data, and brand relationships gathered. These assets create a defensible position that AI cannot replicate overnight, transforming early attention into a long-term strategic advantage.
Enjoying this? Get one like it in your inbox each morning.
one email a day · unsubscribe in two clicks · no third-party tracking
Make belonging, liquidity, or data the real product
Beyond permissions and attention, some businesses thrive by making network effects their core product. People pay for shared goals, accountability, and expertise, whether in virtual communities or in-person experiences. A strong community delivers value not just from the host, but from meaningful member-to-member interaction.
Marketplaces, for example, connect supply and demand, but their true defensibility lies in liquidity: enough trusted participation on both sides to facilitate transactions. While a distinct niche or sales mechanic can help, an audience alone won’t guarantee a viable marketplace. Building that critical mass is difficult for any competitor to copy overnight.
Finally, proprietary data can be a powerful moat. Curated or hard-to-access information holds value for customers and even model developers, provided founders have legitimate access, clear usage rights, and a compelling use case. The practical test for founders: what valuable asset remains if a competitor perfectly clones your software? If the answer isn't belonging, liquidity, or unique data, your business might be more vulnerable than you think.
Frequently Asked Questions
What makes a business AI-resistant?
No business is completely AI-proof, but physical operations, trusted relationships, licensing, proprietary data, and strong distribution can make it harder to copy or replace.
Are software businesses still worth building?
Yes. Software can still succeed when it connects to defensible advantages such as a trusted brand, exclusive data, a community, or a marketplace with real liquidity.
How can AI help founders build a more durable business?
Use AI to reduce operating costs, spot niche trends, produce media, and automate routine tasks while investing in advantages AI cannot instantly reproduce.
Why can licenses and partnerships protect a business?
Licenses create barriers to entry, while established vendor relationships can provide access and trust that a new AI-powered competitor cannot gain overnight.