Reminder: no one knows how to continuously improve LLMs without exponential increase in compute resources beyond what is physically possible.

Ofc, that was before HCNNs & visual transformers outperformed handcoded models in predicting visual system responses. But prediction≠understanding. We neither fully know the algorithm of trained-up ANNs nor which of the predictively equivalent ANNs best approximates actual neural computations

I agree with that too, but my question was not about implementation. My question was about which mathematical description best approximates actual system operations underlying human cognitive capacities. Seems to me that for many capacities the most compact computational description is unknown.

Do you think the same is true for computational questions? It seems to me there is enough unknown about the brain to spark a lot of fruitful research into computational mechanisms

I stopped playing after high school, and the latest games weren’t even new then 🤣 (tbh also played some later versions but they weren’t imprinting on my soul that deeply)

I‘d be curious how you’d respond to this kind of argument?

I recently gave a talk to philosophers of AI about energy use in brains and AI and was pointing out the enormous differences in efficiency used for computation. The general response was something like ”well computation is substrate independent and also the LLM on my laptop uses very little energy“

about a decade ago i attended a talk by a (very famous) chip guy & he said that we needed to make neuromorphic chips in order to facilitate energy-efficient AI. he said given the interdisciplinary audience, he'd particularly like some tough & stimulating questions i didn't have one, but ...

I've been collaborating with the Norwegian Consumer Council for a while now; they not only a quite interested in working with us as philosophers, and do great legal work taking on the big shots, but also: they're funny. If you liked Norsemen, you might like this.

Sounds like they’re exploiting the ambiguity of „human extinction“ strategically

Google: “Why reviewer 2 unhelpful and mean?“ Maybe I should stop talking about concepts and brains and join the promising field of international foot and ankle research 🤔

Not sure who created this, but how fitting

Amazon's own balance sheet presents the most compelling evidence that we are stuck in the Enshittocene, the era in which everything is turning into a pile of shit, because the worst ideas of the worst people now make the most money. 1/