Generative AI: A Creative New World - The Difference Between Amateurs and Professionals

Generative AI: A Creative New World - The Difference Between Amateurs and Professionals

In the rapidly evolving field of generative AI, we are witnessing the dawn of a new creative world. The potential of generative AI is vast, promising to bring down the cost of creation and knowledge work towards zero, leading to increased labor productivity and economic value. This, in turn, translates into significant market potential.

Let's take a closer look at the different waves that have shaped the landscape of generative AI. In the first wave, which occurred before 2015, small models were considered the pinnacle of language understanding. However, as we entered the second wave, the race to scale began. Google Research introduced transformers, a new neural network architecture for natural language understanding. These transformers proved to be superior in generating high-quality language models, while also being more parallelizable and requiring less training time.

As AI models grew larger, they started surpassing major human performance benchmarks. However, these models posed challenges. They were large and difficult to run, requiring GPU orchestration. Additionally, they were not easily accessible to the broader public, often limited to closed beta or unavailable altogether. Furthermore, the cost of using these models as a cloud service was high. Despite these limitations, the early applications of generative AI started emerging.

Now, in the fourth wave, we are witnessing the emergence of killer applications. With the platform layer solidifying, models becoming better, faster, and cheaper, and access to models trending towards being free and open source, the application layer is ripe for a creative explosion. Text generation has made significant progress, but the challenge lies in getting it right and ensuring quality. Code generation, on the other hand, has the potential to greatly impact developer productivity, as exemplified by GitHub CoPilot. Images and speech synthesis have also gained traction, showing promise in areas such as cinema, gaming, VR, architecture, and product design. Even video and 3D models are quickly catching up.

Among the various applications, there is a growing need for personalized web and email content to fuel sales and marketing strategies, as well as customer support. Language models are perfect for this task. Additionally, vertical-specific writing assistants are gaining popularity. Most writing assistants available today are horizontal, but there is a demand for specialized assistants tailored to specific industries or domains. Another exciting application is code generation, which turbocharges developers and enhances their productivity.

Generative AI apps are built upon large models like GPT-3 or Stable Diffusion. As these applications gather more user data, they can fine-tune their models to improve performance for specific problem spaces and reduce costs. These apps act as a UI layer and a "little brain" that sits on top of the "big brain" of large general-purpose models. Currently, generative AI apps exist as plugins in existing software ecosystems. However, as models become smarter, driven by user data, these apps are expected to improve until they are good enough to serve as the final product.

To succeed in the generative AI space, companies must create a sustainable competitive advantage by leveraging the flywheel between user engagement/data and model performance. This entails exceptional user engagement, turning that engagement into better model performance through prompt improvements, model fine-tuning, and user choices as labeled training data, and using the improved model performance to drive more user growth and engagement.

Now, let's shift our focus to the difference between amateurs and professionals. One crucial aspect is mindset. Amateurs often stop when they achieve something, considering it a goal accomplished. Professionals, on the other hand, understand that the initial achievement is just the beginning. For professionals, it's not about achieving a specific goal; it's about following a process.

Consistency is highly valued by professionals. They understand the importance of maintaining a consistent approach to their work. Amateurs, on the other hand, tend to focus on identifying their weaknesses and improving them. Professionals, in contrast, focus on their strengths and seek out others who excel in areas where they are weak.

While amateurs believe that knowledge is power, professionals go beyond knowledge. They pass on wisdom and advice, sharing their insights and experiences with others. Amateurs are often fixated on being right, whereas professionals prioritize achieving the best outcome. They understand that being right is not as important as the end result.

Amateurs may engage in tearing others down, while professionals focus on making everyone better. Professionals understand the importance of collaboration and cooperation. They see disagreements as an opportunity to learn and grow, rather than as threats.

Another key difference lies in how amateurs and professionals approach ideas. Amateurs tend to go with the first idea that comes to mind, while professionals recognize that the first idea is rarely the best one. They know that better outcomes come after several iterations and are willing to put in the necessary effort and time.

Lastly, amateurs often hold the belief that the world should conform to their desires. Professionals, on the other hand, understand that they need to adapt and work with the world as they find it. They embrace the reality and find ways to navigate through it.

In conclusion, generative AI is opening up a whole new world of possibilities, offering immense potential in various domains. As the field evolves, it is essential to adopt a professional mindset that focuses on continuous improvement, collaboration, and adaptation. To thrive in both generative AI and any other field, remember these three actionable pieces of advice:

1. Embrace a process-oriented approach: Instead of fixating on specific goals, prioritize following a consistent and iterative process that allows for continuous improvement.

2. Collaborate and leverage strengths: Recognize the strengths of yourself and others, and seek out collaborations that enable you to excel in areas where you may be weak.

3. Embrace feedback and iteration: Understand that the first idea is rarely the best one. Be open to feedback and iterate on your work to achieve better outcomes.

By incorporating these principles into your work, you can navigate the world of generative AI and any other domain with professionalism and excellence.