Purpose
This article is an opinion piece intended to explore the rapid development of AI image synthesis technologies (such as DALL-E 2, Imagen, Midjourney) and their impact on the creative industry, and it proposes a vision for the future form of prompt engineering tools.
Main Points
- Technological breakthrough: AI image synthesis has crossed the threshold of reliability and quality and can replace a broad range of creative work, such as product photography, icon design, logo design, and magazine illustration.
- Prompt engineering as a new skill: Using AI models requires a new professional skill—prompt engineering or "prompt whispering." This requires understanding the dynamic behavior of AI and being able to iteratively adjust prompts to obtain desired results.
- Tool requirements: The author believes that the ideal prompt engineering tool should be a "boring" application, similar to a workbench, focused mainly on workflows and parameter exposure. It should support multi-project management, color management, custom style fine-tuning, visual exploration of variants, a prompt builder, queue management, and automated batch generation.
- Business model: Because computational costs are high, the tool itself may be free, but it charges through compute top-ups (with a 5% markup), and commercial licensing also requires payment.
- Industry impact: The creative industry will not be immediately replaced by software; rather, agencies will use software to increase productivity 100-fold. Creative personnel will still be needed, but the content of the work will change, with less drawing and more prompt whispering.
- Future outlook: AI capabilities may grow exponentially ("AI overhang"), and in the future real-time, conversational image generation may emerge, requiring tools that are as practical and reliable as office or CAD software.
How to Use
This article is not a tool or service, but an analytical article. Readers can learn about the development trends of AI image synthesis and the concept of prompt engineering by reading it.
Applicable Scenarios
Suitable for readers interested in AI image generation, the future of the creative industry, and prompt engineering, including designers, creative workers, product managers, and technology enthusiasts.
Limitations
This article is based on the state of technology in June 2022, and subsequent developments may have gone beyond what is discussed in the article. The functions and availability of the specific products mentioned (such as DALL-E 2, Imagen, Midjourney) may have changed. The tools envisioned by the author have not yet been realized and are only concepts.
