Advice about AI in software engineering often assumes that today’s limitations will persist: autocomplete needs human context, agents produce code that requires review, and eventually human expertise is said to matter mainly in understanding customers. As capabilities advance, these reassurances shift, like claims of victory in a losing war moving ever closer to the capital. The same tension appears in debates about superintelligence: people may predict transformative effects while acting as though work and society will remain much the same. The argument is that taking AI seriously does not require accepting confident predictions that AGI is imminent; short timelines are a separate, poorly established claim.
Instead, taking AI seriously means considering it possible that progress will continue at its recent pace. By contrast, confidently predicting that AI will stall or have little social impact is difficult to justify: evidence for a slowdown is weaker still, and AI already affects software, mathematics, cybersecurity, writing, and image-making. A story by Borges about an invented world gradually replacing reality offers a metaphor. Rather than debating how likely the transformation is or measuring its economic effects, people should recognize that it could happen. The key is openness to a consequential possibility, not certainty about when it arrives.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.