Experts Dispute Claims of Imminent AI Automation of Knowledge Work

Viral blog post by AI influencer Matt Shumer overestimates the pace of AI disruption in fields beyond software development, critics say.

Published on Feb. 12, 2026

Got story updates? Submit your updates here. ›

A recent viral blog post by AI influencer Matt Shumer warning that AI will automate most knowledge work within 1-5 years has been criticized by experts as exaggerated and based on flawed assumptions. While AI is making rapid progress, particularly in automating software development, experts argue that fully automating other professional fields like law, finance, and medicine will take much longer due to the lack of clear quality metrics and the need for reliability, governance, and auditability that enterprises require.

Why it matters

Shumer's dire predictions, if taken at face value, could cause unnecessary panic and anxiety among knowledge workers about the future of their jobs. Understanding the limitations of current AI capabilities in domains beyond software development is important to have a more realistic assessment of the pace and scope of AI's impact on the workforce.

The details

Shumer's blog post argues that the rapid progress in AI's ability to autonomously write code is a harbinger of what's to come for other knowledge work professions. However, experts point out that coding has quantifiable metrics of quality, like whether code compiles and passes unit tests, that don't exist in other domains. Assessing the quality of legal analysis, medical diagnoses, or marketing campaigns is much more subjective. Additionally, enterprises require reliable, auditable AI systems, which are still lacking compared to the automated workflows possible in software development. Critics also say Shumer exaggerates the current capabilities of large language models, ignoring their well-documented flaws and limitations.

  • On February 5, 2026, OpenAI and Anthropic released new AI models that sparked Shumer's viral blog post.

The players

Matt Shumer

An AI influencer who wrote a viral blog post warning of imminent AI-driven automation of knowledge work.

Gary Marcus

An emeritus professor of cognitive science at New York University and a leading skeptic of current large language models.

Connor Boyack

The president of the Libertas Institute, a policy think tank in Utah, who wrote a counter-blog post arguing that technological revolutions historically create more jobs than they eliminate.

OpenAI

An American artificial intelligence research company that released the GPT-5.3 Codex model cited in Shumer's blog post.

Anthropic

An American artificial intelligence research company that released the Opus 4.6 model cited in Shumer's blog post.

Got photos? Submit your photos here. ›

What they’re saying

“He gives no actual data to support this claim that the latest coding systems can write whole complex apps without making errors.”

— Gary Marcus, Emeritus Professor of Cognitive Science, New York University (Fortune)

“No AI system can reliably do every five-hour long task humans can do without error, or even close, but you wouldn't know that reading Shumer's blog, which largely ignores all the hallucination and boneheaded errors that are so common in every day experience.”

— Gary Marcus, Emeritus Professor of Cognitive Science, New York University (Fortune)

“He likes to sell big. That doesn't mean we should take him seriously.”

— Gary Marcus, Emeritus Professor of Cognitive Science, New York University (Fortune)

The takeaway

While AI is making rapid progress, the full automation of knowledge work beyond software development will likely take much longer than the 1-5 year timeline suggested by Shumer's viral blog post. Enterprises require reliable, auditable AI systems, which are still lacking compared to the automated workflows possible in coding. Experts caution against exaggerated claims about AI's current capabilities and the need for a more nuanced understanding of its impact on the workforce.