4 ms·

Escaping a paradigm is the absolute pinnacle of achievement in theoretical science, yes. I think it would be misguided to go all-in on that goal and forget why

Escaping a paradigm is the absolute pinnacle of achievement in theoretical science, yes. I think it would be misguided to go all-in on that goal and forget why these paradigms matter, though. Everyone else in science who helps build those walls makes meaningful contributions too, to the extent that a paradigm is actually useful (wildly different conversation). If we just scrap every paradigm immediately and move on, that’s sort of the same as going back to tribes and superstition. No one will be able to get very far on their own, not even with a tribe of acolytes. Does that make sense? The biggest tragedy about science is that it usually has a huge delay on its utility. “Brand-new scientific answers” to practical problems of the day are often not much better than answers you get from other methods, both because the paradigms take time to get discovered and mature, and because people are corrupt. This is not the same as saying ballistics isn’t useful, it’s just that it’s remained useful long after it was brand-new science, and most of the other stuff has been forgotten.

nonrandomstring 4y ago
> go all-in on that goal and forget why these paradigms matter Certainly. Rampant iconoclasm and revisionism alone bring chaos. > No one will be able to get very far on their own, not even with a tribe of acolytes. Does that make sense? Absolutely, old paradigms still matter, showing where we've been, how we got here. To totally smash and forget them is to undermine our own foundations. I don't think any successful new paradigm can avoid subsuming the old ones. > a huge delay on its utility. Here's where I imagined the conversation going. However much I dislike Peter Thiel's "control machinations" there's one thing I strongly agree with him on, and it's in that famous discussion with David Graeber; "Where Did the Future Go?" Somewhere, perhaps around the turn of the century, we got really stuck; More. Faster. Bigger. Repeat. We stopped paying attention to the immense social utility possibilities that were (and still are) latent in computer science. I think we've been riding on Moore's law the way we rode fossil fuels. LLM's are the latest round. But I don't think scale is the answer to very much. Maybe it's a bit grand to call it a "paradigm", but I'm claiming the kind of radical thinking that could shake us out of it is a bit thin on the ground these days.

nonrandomstring 4y ago

> go all-in on that goal and forget why these paradigms matter Certainly. Rampant iconoclasm and revisionism alone bring chaos. > No one will be able to get very far on their own, not even with a tribe of acolytes. Does that make sense? Absolutely, old paradigms still matter, showing where we've been, how we got here. To totally smash and forget them is to undermine our own foundations. I don't think any successful new paradigm can avoid subsuming the old ones. > a huge delay on its utility. Here's where I imagined the conversation going. However much I dislike Peter Thiel's "control machinations" there's one thing I strongly agree with him on, and it's in that famous discussion with David Graeber; "Where Did the Future Go?" Somewhere, perhaps around the turn of the century, we got really stuck; More. Faster. Bigger. Repeat. We stopped paying attention to the immense social utility possibilities that were (and still are) latent in computer science. I think we've been riding on Moore's law the way we rode fossil fuels. LLM's are the latest round. But I don't think scale is the answer to very much. Maybe it's a bit grand to call it a "paradigm", but I'm claiming the kind of radical thinking that could shake us out of it is a bit thin on the ground these days.