Samo Burja: AI Automates 'Complicated Scripts,' China Poised to Match Latest Advancements

Political scientist and founder of Bismarck Analysis, Samo Burja, recently offered a "brutally honest" assessment of artificial intelligence, challenging prevailing narratives about its current capabilities and future trajectory. His critique, highlighted by Marko Jukic on social media, suggests that while AI is impressive, its impact will be shaped more by political and economic realities than by a rapid ascent to general intelligence. Burja's insights delve into the true nature of AI's automation and the competitive landscape.

Burja contends that much of what AI currently achieves involves automating "complicated scripts" and "median white-collar professions." He argues that many jobs, even without advanced machine learning, could be automated with sufficient traditional programming, implying that the economy is "shockingly primitive." This perspective suggests that AI's strength lies in processing vast datasets and performing repetitive tasks reliably, rather than exhibiting human-like general intelligence.

A significant aspect of Burja's analysis focuses on the political economy of automation. He posits that professions with "guild-like lock on credentials," such as medicine, law, and finance, will likely resist full automation due to their political power and protection. Instead, he suggests these roles could become "100% self-serving," with AI handling the substantive work while human oversight is mandated, leading to an "extractive class" that could hinder economic growth.

While acknowledging AI's potential to increase productivity and "make society richer," Burja expresses hope that it will empower new productive jobs and help "beat the bureaucracies." He anticipates a "Cambrian explosion of different forms of intelligence" rather than a singular AGI. Crucially, he noted in a social media post, "> I know today everyone is riding on wave of Gemini 3 enthusiasm, and it is warranted. Yet expect to see Chinese lab to match or exceed the performance in 2-3 months at lower costs and at lower compute once again," underscoring the rapid global competition and the potential for non-Western entities to quickly rival leading AI models.