The advent and era of the prompt engineer
The rise of the prompt engineer has been a fascinating development in the world of AI art. These individuals are responsible for crafting the prompts that guide the creation of AI algorithms, resulting in some of the most impactful and thought-provoking pieces of art we've seen to date.
But what exactly is a prompt, and how does it influence the output of an AI algorithm? A prompt is essentially a set of instructions or suggestions that guide the AI's artistic process. It can be as simple as a few keywords, or as complex as a detailed description of a scene or concept. The goal of the prompt is to provide the AI with a starting point from which it can generate its own unique art.
The role of the prompt engineer is to carefully study the subject matter of a given area, such as art, and use their expertise to create prompts that will produce the most striking and meaningful pieces of AI art. This requires a deep understanding of the underlying principles and techniques of the subject, as well as an ability to anticipate the ways in which the AI might interpret and respond to the prompts.
A 'prompt' is worth a thousand words... or more
There is no other technology that has ever eclipsed the invention of writing in the existance of humanity. A strong statement and although without this truth almost everything that has allowed the pace of evolution and technological progress would not have been possible. With the clever use of language and some understanding of subject matter in the creation of prompts. It is clearly a skill where both subject matter and the ability to clearly and effectively express ideas through the written word are key in being able to consistently and effectively leverage these AI services.
If we're going to interact with AI through language. Then effective and expressive writing will be a highly sought-after skill.
Subject matter is important but might not be the main ingredient in creating something relevant or even award winning content anymore. If we're going to interact with AI through language. Then effective and expressive writing will be a highly sought-after skill. A skill currently undervalued and in most cases severly underpaid. AI could change the current status quo in this area drastically.
With a well curated prompt books, art, music and more could practically be churned out in moments. In addition to this reailty the quality, volume and style of these pieces would all fundamentally be influenced by how well designed the prompt is. OpenAI dedicates an entire section in their documentation to 'prompt design' within the @link1@. Good documentation and a standardised list of common requirements vs common solutions(read prompts) is a key indicator of good engineering practice, i.e best practice. So here we're starting to see how a new set of career fields focused within specific subjects has emereged. Text Completion Guide
Take the example of Jason Allen's "Théâtre D'opéra Spatial," which took first place in the digital category at the Colorado State Fair. Now instead of thinking about the process of brush stroke, layers, colour pallete specifically. Much of the focus would have been on artistic expression through the prompt. Mr Allen is still an artist, of course, taste, style, approach and even the number of prompt revisions and iterations allowing him to fine tune the idea before the model attempts to paint. What's happening here is the prompt becomes the art and the tool. As the same prompt wouldn't render the exact same results.
Jason Allen’s A.I.-generated work, “Théâtre D’opéra Spatial,” took first place in the digital category at the Colorado State Fair. Credit...via Jason Allen
What we could be looking at here in the way of AI ultimately removing the need for manual input in fields that require heavy specialisation is one of the last steps towards a truly thaught based economy. The final step being the introduction of Brain Controlled Interfaces into the mainstream. Both topics we've covered in previous articles here at Neutron Salad quite extensively.
There is a large amount of reductionism when looking at language models being ready to replace an entire industry. But very little in the way of value creation from a purpose perspective. How does humanity collaborate with these modals to create something valuable. I believe we're missing a trick here as a majority. A 'prompt engineer' is but one of the key areas where the barrier to entry might have even been lowered in some areas allowing those who are great at expressing themselves but not so great at writing code for example would be the next era of problem solvers. A big and possibly controversial ideas, yes, but not as severe from either persepctive as destroying humanity a result of needing to produce more paperclips, surely.
- Jonas Oppenlaender. (Wed, 20 Apr 2022), A Taxonomy of Prompt Modifiers for Text-To-Image Generation, - arXiv.2204.13988
- Stephen H. Bach; Victor Sanh; Zheng-Xin Yong; et al. (2 February 2022), PromptSource: An Integrated Development Environment and Repository for Natural Language Prompts (PDF), - arXiv:2202.01279v3
- Witteveen, S. and Andrews, M., 2022. Investigating Prompt Engineering in Diffusion Models. - arXiv:2204.13988v2