An unsolicited guide to being a researcher [pdf]
Summary
An unsolicited guide to being a researcher argues that true scientific advancement requires visionary AI that enables paradigm shifts rather than mere prediction within existing frameworks. Using Borges’s map allegory, historical examples (Maxwell, Einstein, Darwin), and modern AI trends (AlphaZero, foundation models), the piece cautions against hypernormal science and advocates designing AI that helps create new vocabularies and cross-disciplinary analogies. It proposes metascience experiments and practical principles to steer AI development toward transformative science rather than incremental gains.