# A Turning Point in AI Perception: When "Best Humans Still Outperform" Becomes News > Coverage of lessw-blog **Published:** April 17, 2026 **Author:** PSEEDR Editorial **Category:** platforms **Tags:** Artificial Intelligence, Benchmarks, Societal Impact, LLMs, Foundation Models, LessWrong **Canonical URL:** https://pseedr.com/platforms/a-turning-point-in-ai-perception-when-best-humans-still-outperform-becomes-news --- A recent post from lessw-blog examines the psychological and societal shifts in how humanity benchmarks itself against artificial intelligence, highlighting a fascinating turning point in our collective reaction to rapid AI advancements. In a recent post, **lessw-blog** discusses a subtle but profound shift in how we talk about artificial intelligence. The author points to a 2023 academic journal headline-"Best humans still outperform artificial intelligence in a creative divergent thinking task"-as a defining moment in the history of human reaction to AI capabilities. By analyzing the phrasing and the underlying assumptions of this headline, the post captures a critical transition in our collective understanding of machine intelligence. As foundation models and large language models (LLMs) continue to scale, the benchmarks used to evaluate them have increasingly encroached on domains traditionally considered exclusively human. Creativity, divergent thinking, and nuanced problem-solving were long thought to be safe havens from automation, acting as the ultimate firewall against machine parity. However, the rapid public adoption of generative tools like ChatGPT and Stable Diffusion has forced a societal reckoning. The fact that an academic paper felt the need to explicitly state that the _best_ humans still outperform AI in a specific creative task reveals a significant lowering of the bar for what we expect machines to achieve. It implicitly suggests that average, or even above-average, human performance may already be matched or surpassed by these systems. This dynamic is critical for anyone tracking the platform layer of AI, as it dictates how benchmarks are perceived and how future models will be evaluated. The lessw-blog post argues that this framing represents a fundamental "attitude shift" and an early sign of societal "cope." Rather than assuming human superiority by default, the narrative has shifted to finding specific edge cases, highly specialized tasks, or elite human percentiles that still hold a measurable lead. The author contextualizes this phenomenon not as a sudden, unpredictable break in technology, but as part of a continuous lineage of AI development that has finally crossed a psychological threshold for the general public. While the specific details of the creative divergent thinking task or the exact architectures of the models tested in the cited journal are not the primary focus, the post successfully captures the anxiety and recalibration occurring across industries. It highlights how the goalposts for human exceptionalism are being rapidly moved. For professionals tracking the trajectory of AI benchmarks and foundation models, this analysis offers a crucial lens on the human element of technology adoption. Understanding how society reacts to these milestones, and how we frame our remaining advantages, is just as important as tracking the technical capabilities of the models themselves. This psychological landscape will inevitably influence future AI development, regulatory approaches, and enterprise adoption strategies. [Read the full post](https://www.lesswrong.com/posts/weD4iST2FvSo5D9T4/best-humans-still-outperform-one-turning-point-in-the) to explore this fascinating perspective on our evolving relationship with artificial intelligence. ### Key Takeaways * A 2023 academic headline celebrating that the 'best humans' still beat AI in creativity signals a major shift in societal expectations. * The necessity of this claim implies that average human performance in certain creative tasks is already being challenged or surpassed by AI. * This shift represents a psychological 'reckoning' and a form of societal 'cope' as AI capabilities encroach on traditionally human domains. * Public perception of tools like ChatGPT and Stable Diffusion marks a psychological threshold, even though the underlying technology follows a continuous developmental lineage. [Read the original post at lessw-blog](https://www.lesswrong.com/posts/weD4iST2FvSo5D9T4/best-humans-still-outperform-one-turning-point-in-the) --- ## Sources - https://www.lesswrong.com/posts/weD4iST2FvSo5D9T4/best-humans-still-outperform-one-turning-point-in-the