Roman Feiman @romanfeiman.bsky.social · 28/09/2026Moreover, if anyone did find X-at-first-encounter in NNs, there still wouldn't be any paradox. It would just show that an NN architecture innately processes X, with no learning needed. 210
Jake Quilty-Dunn @quiltydunn.bsky.social · 28/09/2026let's focus on the conceptual point first. i agree that the notion of innateness that is most important in dev psych would say that if a vanilla NN can deploy a representation of X during the first encounter then it has that representation innately. but this does create a tension: 110
Jake Quilty-Dunn @quiltydunn.bsky.social · 28/09/2026the NN has no info about that domain prior to the first encounter. so: it is learning in the sense of tuning its parameters to experience while also not learning in the sense of showing a classic learning trajectory. this is a purely conceptual point about what we would say if NNs could do this 210
