Generative AI is not new. With a few notable exceptions, most of the technologies we’re seeing today have existed for several years. However, the convergence of several trends has made it possible to productize generative models and bring them to everyday applications. The field still has many challenges to overcome, but there is little doubt that the market for generative AI is bound to grow in 2023.

Down the road, however, the real power of generative AI might manifest itself in new markets. Who knows, maybe generative AI will usher in a new era of applications that we had never thought of before.

Generative models were first presented as systems that could take on big chunks of creative work. GANs became famous for generating complete images with little input. LLMs like GPT-3 made the headlines for writing full articles.

But as the field has evolved, it has become evident that generative models are unreliable when left on their own. Many scientists agree that current deep learning models — no matter how large they are — lack some of the basic components of intelligence, which makes them prone to committing unpredictable mistakes.

As Douglas Eck, principal scientist at Google Research, said at a recent AI conference, “It’s no longer about a generative model that creates a realistic picture. It’s about making something that you created yourself. Technology should serve our need to have agency and creative control over what we do.”

Product teams are learning that generative models perform best when they are implemented in ways that give greater control to users.

The past year has seen several products that use generative models in smart, human-centric ways. For example, Copy AI, a tool that uses GPT-3 to generate blog posts, has an interactive interface in which the writer and the LLM write the outline of the article and flesh it out together.

Product teams with seasoned machine learning engineers can use open-source generative models such as BLOOM and Stable Diffusion. Meanwhile, teams that don’t have in-house machine learning talent can choose from a wide variety of solutions such as OpenAI API, Microsoft Azure, and HuggingFace Inference Endpoints. These platforms abstract away the complexities of setting up the models and running them at scale.

Adobe is also preparing to integrate generative AI in its video and graphic design tools. And Google also has several generative AI products in the works.

Source: How 2022 became the year of generative AI


©RGLN3, LLC 2022

KOLOR // Artificial Industrialist

KOLOR Looks Forward to Building With You!


Log in with your credentials

Forgot your details?