DeepSeek engineer says AI may soon catch up in kernel development work

News Editor

2026-09-15 06:57:45

DeepSeek kernel engineer Liu Shengyu said AI is rapidly closing in on parts of his job, including low-level operator development and optimization. In a recent post, he said he wrote the main Attention kernel for DeepSeek V4.1, but at the current pace of progress, AI will likely match or exceed his kernel-writing ability within six months to a year. He said that a year ago, AI could mainly help with documentation lookup, code reading, and bug finding. Now it can read CUDA, PTX, and SASS, analyze instruction stalls, and optimize kernels on its own. Liu said the next step may be AI systems that design scheduling plans, evaluate performance, and complete implementation independently. He added that he is unlikely to lose his job outright, but may be forced to shift roles, moving from writing kernels by hand to operating AI agents. He also explained why he remains at DeepSeek, saying frontier AI should be offered openly and cheaply to everyone, and warning against a future in which the most powerful AI is controlled by a small number of companies.

DeepSeek kernel engineer Liu Shengyu said AI is having a direct impact on his work, arguing that the technology is moving fast enough that it will likely catch up with, or even surpass, his ability to write kernels within six months to a year.

Liu said he wrote the main Attention kernel for DeepSeek V4.1. He added that a year ago, AI could mainly help him look up documentation, read code, and find bugs. Now, he said, it can read CUDA, PTX, and SASS, analyze the stall time of each instruction, and optimize kernels independently.

AI is moving deeper into optimization work

Looking ahead, Liu said AI may soon be able to design scheduling plans, evaluate performance, and complete implementation by itself. He said he understands clearly that the better he optimizes kernels, the faster DeepSeek models can train and run inference, which also means AI may catch up with him more quickly.

Even so, he said stopping now would not change the broader race because models at other companies are not standing still. For that reason, he plans to keep pushing kernel work as far as possible, even if the competition could eventually replace the kind of work he does today.

From writing kernels to operating agents

Liu said he probably will not become unemployed, but he may be forced to change roles, shifting from writing kernels by hand to becoming what he described as a "mecha pilot" operating agents. What is harder for him to accept, he said, is not losing a livelihood, but the possibility that he may no longer get the chance to do the work he actually loves. As he put it, "I have no choice but to bury my talent in yesterday."

Why he is staying at DeepSeek

Liu also explained why he remains at DeepSeek. He said frontier AI should be made available to everyone in an open and inexpensive way, and said plainly that he does not believe Anthropic or OpenAI will do that.

He added that he worries the most powerful AI could end up concentrated in the hands of a small number of companies, turning technological gaps into wider gaps in power and social class.

This article was originally published by Bit.Fan. For more cryptocurrency news and market insights, visit www.bit.fan.

5600

Disclaimer:

The market information, project data, and third-party content displayed on this platform are for industry information sharing only and do not constitute any form of investment advice or return commitment.

Cryptocurrency trading carries high risks. Users should fully assess their risk tolerance and make independent decisions. All profits, losses, and legal responsibilities are borne by the users themselves.