Jonathan SchofieldUsing machine learning to model brains, proteins, materials: ok.
Using LLMs to produce summaries: fucking stupid
Paul Healey@urlyman Others might hold that there is a confluence between the limits of Algorithmic Information; the modelling of digital and analogue processes and Information Algorithms; the reasons for the effective and beneficial processing of properties states. With no proof, are LLMs and quantum computers not just the projection and gaslighting of abstract relations and the reductionist predication of presuppositions? #Models #Proofs #Algorithms #Information #QuantumComputers #LLM #PhilosophyOfAlgorithms
@Dialectician My sense is that, even when we apply reductionist methods, if we remain open, attentive and keep a lid on our hubris, we can get to valuable insights. And as long as, having done the work, we constantly remind ourselves that the map is not the territory.
But that’s hard, especially in a world saturated with transactional incentives.
On a related note, I listened to ‘How Life Remembers: From Metamorphosis to Simulation’ yesterday and found it fascinating:
https://helioxpodcast.substack.com/p/how-life-remembers-from-metamorphosis
