DeepSeek AI engineer slams Anthropic, OpenAI over ‘pacing’ calls, invokes Nazi Germany
DeepSeek engineer Liu Shengyu criticized Anthropic and OpenAI for their closed-source approach to AI development. He specifically stated his distrust of Anthropic mastering advanced artificial intelligence or AGI. Liu Shengyu argued that cutting-edge AI should be open and affordable. This stance contrasts with DeepSeek's open-source model strategy.
The DeepSeek engineer's comments are not just about open versus closed source. They reflect China's broader anxiety about US dominance in advanced AI. Liu Shengyu's direct challenge to Anthropic and OpenAI signals a growing frustration within China's tech community. They see calls for "pacing" AI development as a move to consolidate power.
This tension directly impacts Asian AI companies. Chinese firms like DeepSeek prioritize open-source models, aiming to democratize AI access. This strategy could give them an edge in markets where proprietary US models face regulatory or trust hurdles. It also forces a choice for developers in Southeast Asia: align with US closed models or Chinese open-source alternatives.
The thing to watch is whether China's open-source AI models gain significant traction outside the country. If DeepSeek's V4.1 models achieve widespread adoption, it would validate their strategy. This would also intensify the global competition for AI talent and market share between China and the US.
Related reading
6 stories
China is beating the US in consumer AI adoption, thanks to super apps: Morgan Stanley
China's AI strategy, prioritizing open-source models, is already yielding results in consumer adoption.
AI slowdown calls prompt chip selloff
Calls for an AI slowdown, as discussed in the article, have already impacted the chip market.
People’s Daily rejects US claims of malicious AI distillation, warns of countermeasures
Databricks to Invest Over US$350 Million in Singapore, Double Its Workforce
ByteDance’s AI-enhanced short-drama app eclipses China’s Netflix rivals combined
