automation

Gary Marcus critiques Matt Shumer’s viral blog post with nearly 50 million views, calling it hype with misleading claims about AI’s coding capabilities. He argues that Shumer’s assertions lack data and ignore issues like error rates and user experiences. Marcus highlights five points of concern, including potential user burnout and security risks of AI-generated code, while cautioning against overconfidence in AI’s reliability. He concludes that despite some improvements, Shumer’s perspective is overly optimistic and unrealistic.

https://garymarcus.substack.com/p/about-that-matt-shumer-post-that

AI is rapidly advancing, transforming how work is done. Major breakthroughs, such as model GPT-5.3, have made AI increasingly competent in tasks that once needed human expertise. Many tech workers already feel AI’s impact, observing that it can outperform them in coding and other areas. This shift is not gradual; it’s accelerating, with predictions that AI can replace up to 50% of entry-level white-collar jobs within 1-5 years. The implications are vast, affecting fields like law, finance, writing, and customer service, as AI develops decision-making capabilities similar to human judgment. To adapt to these changes, individuals should actively engage with AI tools, understanding their potential and leveraging them to enhance productivity. Continuous learning and an openness to change are critical to thrive in this evolving landscape. The next few years will be crucial for those willing to embrace AI as it reshapes various industries.

https://shumer.dev/something-big-is-happening