Generative AI Needs a Business Model, Not Just a Model
Generative AI is a fantastic component .. but it's just a component.
The excitement around Generative AI, particularly with the rise of GPT-3, ChatGPT, and similar large language models, feels unprecedented. Headlines scream about transformative possibilities, from automated creativity and hyper-efficient customer service to groundbreaking research. Yet, we must remember: Generative AI is a powerful component, but still, fundamentally, a component of larger solutions.
To fully understand this, let's reflect on historical technological components that similarly reshaped our world:
Binary Theory
In 1936, Alan Turing formalized a machine representing logic through ones and zeros. His insights did more than just establish computational theory—they formed the foundation of digital computing, which significantly contributed to Allied victory in WWII through cryptography and laid the groundwork for modern computers.
Transistor
The invention by Shockley, Bardeen, and Brattain revolutionized circuits beyond mere miniaturization. It sparked the digital age, making computers smaller, faster, and affordable, fueling a global technology explosion.
Relational Databases (RDBMS)
Standardizing data storage and retrieval fundamentally transformed information management. Oracle, IBM DB2, and later Microsoft Access made data structured, accessible, and actionable at scale.
Data Exchange Protocols
Protocols like SMTP, FTP, and TCP/IP allowed computers to reliably communicate, laying the groundwork for networks and the eventual internet.
The Internet
Tim Berners-Lee's HTTP and HTML shifted knowledge access from local storage to a global interconnected network, profoundly decoupling information from physical location and dramatically increasing accessibility.
Visual Web Browsers
With Mosaic and Netscape Navigator, the internet became user-friendly. Browsers made digital exploration intuitive and attractive to the general public.
Object-Oriented Programming (OOP)
Alan Kay and his team at XEROX PARC introduced a number of innovations, including Smalltalk, their personal favorite (though the world was more tuned to the graphic user interface). Later, C++ and Java reorganized logic into reusable, easily manageable components, paving the way for complex, scalable software development.
Ubiquitous Cheap Storage
Advancements from HDDs to SSDs enabled mass content creation, local caching, and a “save everything” mindset, fundamentally changing our interaction with data.
Open Source Software
GNU, Linux, Apache, and GitHub democratized software creation, removing barriers to infrastructure and fostering unprecedented collaboration.
3G/4G/5G & Wi-Fi
Mobile networks and wireless technology unshackled computing from desks, embedding real-time interaction into everyday life and movement.
APIs & Web Services
SOAP, REST, and JSON transformed how systems integrated, enabling components to effortlessly snap together like digital Lego blocks.
Mobile Computing & App Stores
The iPhone and App Store revolutionized personal computing, making powerful applications accessible everywhere and redefining human-computer interactions.
Cloud Infrastructure
AWS, Azure, and Google Cloud abstracted hardware complexities into a flexible utility, democratizing access to vast computing resources.
Search Engines
Google’s PageRank algorithm reshaped how knowledge was accessed, filtered, and prioritized, transforming information retrieval into an intuitive daily activity.
Containers & Virtualization
VMware, Docker, and Kubernetes decoupled software from hardware dependencies, standardizing deployments and reducing friction in development processes.
Recommendation Engines
Amazon, Netflix, and TikTok didn't just respond to demand; they began actively shaping consumer behavior, driving personalized content and commerce at unprecedented scales.
Graphic Processing Units (GPUs)
What started as a better way to render graphics, typically for high-end gamers, morphed into a new type of chip for parallel processing and eventually gave us transformer tech (chatty AI).
Attention is All You Need
Google set out to improve search by better understanding meaning through transformers. It worked but when applied in reverse, to write instead of read, it gave us Generative AI.
Generative AI (LLMs)
Transformers and GPT-based models have dramatically reshaped our approaches to ambiguity, narrative creation, content synthesis, and human-computer interactions. They're already changing industries from education to healthcare to entertainment.
Yet, even with such profound impacts, these innovations—including Generative AI—are still components. Powerful, yes; transformative (pardon the pun), undoubtedly. But isolated, they achieve little. Just like the printing press — arguably the most significant technological breakthrough of its time — it was a knowledge component. Without authors, scribes, ideas, and societal structures prepared to absorb and disseminate printed texts, Gutenberg's press would have merely been an impressive mechanical novelty.
The printing press didn't just press paper or print pages: it propelled humanity from the Dark Ages into the Renaissance. However, it required a synergy of other components: literacy, paper, ink chemistry, and socio-political changes to reach its full potential.
Similarly, Generative AI will reach its greatest potential as a part of broader vertical solutions, paired with robust data systems, intuitive user interfaces, strategic human oversight, and ethical frameworks. Its role is to enhance and elevate these elements, not replace them entirely.
Ultimately, understanding that Generative AI is a powerful yet integral component can guide better decision-making. It reminds us to thoughtfully integrate it into larger frameworks and avoid trivializing its potential … like reducing it merely to generating cat memes.
Speaking of those cat memes, we shouldn’t ignore that people aiming for one things can spawn another. When the GPU was created, people really just wanted better rendering for the Call of Duty sessions. Instead, we can draw a direct line from that to where we are now (and we still have better Call of Duty graphics). People building GPUs just wanted better gaming graphics … and accidentally ushered in the next AI revolution.
Let's harness Generative AI's immense potential responsibly, ensuring it contributes meaningfully to larger, valuable solutions rather than fleeting novelties.