A compelling analysis from John Gruber's Daring Fireball, which has ignited fierce debate on platforms like Hacker News, posits a simple but profound idea: artificial intelligence is a technology, not a product. This framework suggests that the vast majority of standalone AI applications are destined to be absorbed as features into larger, existing platforms. This perspective forces a critical re-evaluation of strategy for thousands of startups and the tech giants they compete with.
The 'Electricity' Analogy
The core argument, as laid out by Daring Fireball, compares AI to foundational technologies like electricity or TCP/IP. Consumers don't buy "electricity"; they buy refrigerators, televisions, and light bulbs that are powered by it. Similarly, the long-term value of AI may not lie in selling access to a model, but in seamlessly integrating its capabilities to make existing products smarter, faster, and more useful.
This perspective challenges the business model of many startups built around a single AI-powered function, such as a summary generator or an image editor. While novel today, these tools risk becoming mere features in a future version of Google Docs or Adobe Photoshop. The technology is the enabler, but the product is the established software with the existing user base.
Why Incumbents Have the Upper Hand
If AI is a feature, then the companies with the most successful products and largest distribution channels are positioned to win. Tech behemoths like Apple, Microsoft, and Google have a profound, perhaps insurmountable, advantage. They can bundle powerful AI capabilities into the operating systems and applications that billions of people already use daily.
Consider the strategic advantages for these incumbents:
- Massive Distribution: They can push AI features to billions of users overnight via software updates to iOS, Android, and Windows.
- Existing Workflows: They can integrate AI directly into established products like Microsoft Office or Google Workspace, enhancing productivity without requiring users to adopt new tools.
- Vast Data Access: These companies possess enormous, proprietary datasets to train and fine-tune their models, creating a powerful competitive moat.
- Brand Trust: Users are more likely to trust AI features from established brands than from unknown startups.
For developers and founders navigating this complex landscape, understanding these strategic moats is essential. To stay ahead of platform shifts and market consolidation, subscribe to the AI Breaking Wire newsletter for weekly analysis delivered to your inbox.
The Peril for Standalone AI Apps
The "AI as a feature" model poses an existential threat to startups building thin-veneer products on top of large language models. These companies face the constant danger of being "Sherlocked"—the term for when Apple builds a startup's core functionality directly into its operating system, rendering the standalone app obsolete.