A ChatGPT user asks OpenAI to cut running costs without changing how it talks

What the post asks for

As community.openai.com showed, a user posted a request to OpenAI's leadership and engineering teams. They ask the company to look into whether efficiency techniques demonstrated by organizations such as DeepSeek could substantially lower the hardware and operating costs of OpenAI's existing models.

The author is clear about one limit. They do not want GPT's training, reasoning ability or conversational behavior swapped for another model. The idea is to keep the model and improve the machinery that runs it.

What "inference" means here

Inference is the work of running an already-trained AI model to produce an answer. Training builds the model once. Inference happens every time someone sends a message, so it is a cost that keeps recurring.

Making inference more efficient means getting the same answers from less hardware or less electricity. That is the lever the post is pointing at.

The user's real concern

The author says they value more than ChatGPT's technical skill. They like how it communicates, how it explains hard subjects and how it helps them work through problems. They grant that other AI systems may be capable, but say they do not necessarily offer the same experience.

Their fear is that those qualities could disappear if the technology became too expensive to maintain. They also suggest OpenAI could hire or partner with specialists in efficient inference to improve its long-term financial footing.

How much weight to give it

This is a forum post, not an announcement. The source does not say who wrote it, and it offers no data about OpenAI's actual costs or about how much any technique would save. It is an opinion from a user, written in good faith.

It does not show that OpenAI is changing anything. Its value is as a sign of what some users care about: not only whether a model is smart, but whether its manner of working survives the economics of running it.