AI won't take all jobs (technical.ly)

Recent commentary from Anthropic CEO Dario Amodei and University of Louisville professor Roman Yampolskiy predicts significant job displacement due to AI, with Yampolskiy claiming that 99% of work could be replaced by AI by 2030. However, contrary trends are emerging as businesses realize that technology, particularly large language models (LLMs), often requires more human oversight rather than replacing it. For instance, Microsoft’s 365 Copilot sales reflect a low conversion rate, indicating that LLM adoption is not as widespread as anticipated. This sentiment echoes historical patterns in technological advancement, where automation and new industries have consistently absorbed displaced workers rather than eliminating jobs entirely. The current hype surrounding AI may partly stem from poor corporate hiring strategies and economic turbulence, rather than the technology’s inherent capabilities. For example, Salesforce’s massive layoffs were framed as AI-driven but largely resulted from an overzealous hiring spree during the pandemic. Critically, evidence suggests that LLMs, far from being job replacements, often hinder productivity, as shown in the author's personal experiences with various AI-driven tools. This disconnect between the perceived potential of AI and its actual performance raises concerns about the sustainability of the AI-driven narrative, emphasizing the need for more realistic assessments of technology’s role in the workforce.