DeepSeek Engineer Prepares Own Replacement with AI

DeepSeek engineer Liu Shengyue, who wrote the core attention kernels for DeepSeek V4.1, published a viral WeChat essay stating that AI is already reading GPU code (CUDA, PTX, SASS) and optimizing operators on its own, estimating the full transition of this work to AI in 6–12 months.

Compiled by Sergey KostenchukPublished 2026-09-18Updated 2026-09-18

2026-09-18 Research OpenAI

🤖 DeepSeek Engineer Prepares Own Replacement with AI

Liu Shengyue, author of the DeepSeek V4.1 attention kernels, published an essay on WeChat titled "I Have No Choice but to Bury My Talent in Yesterday": over the past year, AI has learned to read GPU code (CUDA, PTX, SASS) and optimize operators on its own, and in 6–12 months it will be writing kernels at the author's level.

🌍 Demand is shifting from those who write kernels manually to those who make AI do it faster. Liu does not trust Anthropic and OpenAI to lead in AGI and advocates for open access to advanced models.

👤 Developers are now valued for orchestrating AI and understanding the entire stack, not just manual coding. Business Insider confirmed the authorship, Liu confirmed it, but declined to comment.

Source 1: https://www.businessinsider.com/deepseek-engineer-viral-post-inside-china-ai-race-2026-9 Source 2: https://www.geopolitechs.org/p/why-a-deepseek-engineer-is-helping

Sources