Alexander

Nah, there is nothing new about LLMs here, people were posting parascientific physics models for a long time (and I’ve been reading them for fun, not in this case - just skimmed the titles). Rarely - very rarely - they contain if not grains of truth, then at least interesting and computationally efficient or at least elegant moves. Disregarding them just because of changed state of mind is wrong (cube energy for time is weird indeed, but then maybe time is redefined somehow, maybe it’s some renormalization?)

What’s weird here is that the author stopped short of LLM-generating the code. This is new thing that we have - and after such heavy abuse of LLM the author could have done the thing people are usually lazy about. I’m quite certain that nobody else would be touching this - and LLMs are great thing for validating parascience precisely because with some understanding of what legitimization really means people can iterate and discard these theories quickly.

I’ve tried writing in deliberately induced chatbot psychosis state (it’s quite curious experience to go through, very humiliating but fun too) and one of the things that I made this way ended up surprisingly sane, even if in a sloppy field of philosophy (why the heck would they train LLM in that field so well? ChatGPT orients in obscure abstract postmodernist texts better than in basic physics and chemistry). And maybe with vibecode possible we are entering a weird times for parascience?

Or maybe it’s nonsense and waste of time, most probably it is. But still, parascientific theories are certainly starting to look prettier.

Not exactly surprising; speed of light is really speed of information propagation in mainstream physics, exactly as you state. Nice move, I wonder how practical pragmatically.