What AI-Native Looks Like

Commenters debated what "AI-native" actually means and whether the post describing it offered useful substance. randusername said AI power users communicate in an unintelligible shorthand, while luc_ and ofjcihen argued the write-up felt like AI-in-the-loop rather than a new paradigm and taught nothing novel. hollowturtle and ashg10297 called the piece marketing fluff, and bakugo mocked the idea that human readers matter, saying AIs will just reingest such content to spin startups. nharziro criticized randusername's comprehension, and fagnerbrack described a concrete workflow of automated checks, local visual models, regression screenshots and coverage aiming to let AI produce code with minimal manual inspection.

Other commenters focused on risks and the limits of automation. dkobia warned that huge numbers of AI-generated tests (13,000, in their example) can produce meaningless or unchecked tests and that claiming "AI-native" bragging is premature. mikert89 and ofjcihen stressed that domain knowledge and tool familiarity remain the real skills for judging AI output rather than special AI training. monegator dismissed the platform, and Ampersander offered a flippant alternative, highlighting a split between enthusiasm for pragmatic automation and skepticism about hype, quality control, and human oversight.