We often hear about the dichotomy between quantity and quality. Our intuitive reaction is typically to prioritize quality over quantity, but this isn't always the correct approach. Observing the evolution of tooling reveals that the opposite can be true, albeit with a caveat. Quantity, when combined with discernment, can be significantly more valuable than quality alone. Let me elaborate.
Several years ago, a team working for a highly esteemed author (whom I greatly respect) released a statement. They revealed that the author would simply unload raw ideas onto them, and it was their job to sift through these ideas, connect the dots, curate them, and present them in the engaging format that readers appreciate. Obviously, with due respect to the editorial team, an author is an author due to the ideation.
The Diminishing Need
With the advent of advanced tools, the landscape of content creation has undergone a significant transformation. Ideation remains a cornerstone, yet the entire process of editorial curation has become streamlined. The journey of writing my first book spanned 9 months, whereas the second edition was completed in just 6 months. Presently, I am able to compose a newsletter within approximately an hour, enabling the daily publication of new editions. ChatGPT and similar technologies act as my editorial team, albeit with a unique twist. It typically requires 20-40 revisions before I deem the content ready for publication. Absent ChatGPT, this process would extend to at least a week, even with support – a luxury typically reserved for well-known authors. With ChatGPT, the processes of ideation and curation are expedited to about an hour. Naturally, delving into the realm of generative AI is an ongoing endeavor. My content creation is a direct outcome of my explorations in this field, rather than the opposite.
RoostGPT exemplifies the transformative power of AI in software testing, serving as a virtual test case author. This innovative product navigates through a myriad of test case scenarios, conducting thorough assessments of each. Despite the complexities and detailed requirements for crafting high-quality test cases, RoostGPT's foundational strategy is to leverage a wide array of viable options. It meticulously sifts through these options, enhancing and honing the output to spotlight the most appropriate and impactful solutions. This approach highlights a significant paradigm shift towards an environment abundant with choices, enabling us to extract the most favorable outcomes from an extensive pool of possibilities. This not only accelerates the testing process but also elevates the quality and effectiveness of software testing to unprecedented levels.
In the scenarios I've discussed, employing numerous human editors or test-case authors is not only financially unsustainable but also impractical. Additionally, the motivation of human contributors can wane if their efforts don't contribute to the final product. In contrast, Generative AI or Foundation agents, lacking human concerns, don't face such motivational issues. They tirelessly iterate and refine without the constraints of cost and practicality that limit human involvement.
In conclusion, as automation becomes increasingly sophisticated, the perceived value of human expertise in routine tasks is set to diminish. To justify human involvement in activities that can be efficiently executed by Foundation Agents—capable of performing tasks more economically, swiftly, and effectively, without any fuss—a person would need to possess nearly superhuman abilities. However, the quintessential human qualities such as ideation, pure creativity, and attributes related to higher-level consciousness will remain uniquely human domains.