Optimizing Resource Usage

Hello OpenAI Team,

As a regular user and strong supporter of efficient AI technologies, I would like to suggest a critical optimization idea that could substantially improve your services’ performance, scalability, and environmental impact.
This suggestion is not only technical, but also strategic for the sustainable future of AI development.
Currently, every time a user requests a minor modification to a generated image (or a text/code output), the system performs a full regeneration rather than modifying the existing output.

This causes:

  • Unnecessary energy and computing resource consumption;
  • Increased load on data centers;
  • Longer response times for users.

Suggested Improvement:

  • Implement a temporary session-based storage for generated outputs (images, texts, codes);
  • Allow direct modification of the stored output during subsequent user interactions;
  • Trigger full regeneration only when essential (major transformations).

Expected Benefits:

  • Significant reduction in server workloads and electricity consumption;
  • Faster and smoother user experience;
  • Greater scalability with lower environmental footprint;
  • Alignment with sustainable AI development goals.

This approach seems particularly obvious for image editing workflows but would be equally beneficial for text, documents, and coding assistance.

Thank you for considering this idea. I am convinced that optimizing resource management at this scale is essential for AI’s future impact — and OpenAI could once again lead the way in responsible innovation.

Best regards,
Gérard

1 Like

Hi Gérard,

I love your idea for optimizing resource usage. I’m very curious to know if it’s possible to achieve the goal while allowing direct modification. The only way I can see it happening is by having a temporary session that would consume end user computing resources. Would still be better than triggering a whole new image in my opinion.

Jorge

1 Like