The goal post moving by anti-AI people is getting ridiculous.

A viral critique from a long-time AI observer highlights a pervasive pattern in the industry's discourse: rapid technological normalization. The post traces the journey from 2017's pixelated AI images to 2024's Sora generating studio-quality video, and from models struggling with coherent sentences to Claude 3 Opus solving Olympiad-level math. Each monumental leap—like AI transitioning from writing simple functions to managing entire codebases—is instantly absorbed into the expected baseline. The conversation then immediately shifts to the next frontier of imperfection, such as a 99% accurate code assistant being dismissed for the 1% it misses.

This 'shifting goalpost' phenomenon is attributed to a wave of users who entered the field post-ChatGPT, for whom conversing with a large language model is mundane. They take for granted capabilities that would have seemed like science fiction just five years ago, like generating in minutes what required months of animation studio work. The core argument is that this constant re-baselizing of progress creates a distorted public narrative, where exponential gains in generative AI, multimodal reasoning, and agentic systems are underappreciated. The implication is profound: if the current pace of dismissing achievements continues, the industry risks misrepresenting one of the fastest technological adoptions in history as a plateau, rather than the ongoing revolution it is.