Science fiction
Baldur Bjarnason blogs about my post on the foundational question of language models —
So: I think Baldur misreads me a bit, but/and on balance his consideration is sharp, and I’m very grateful for it. Now, there’s a part of his post that I want to emphasize.
Regarding the vision of super science empowered by advanced language models, Baldur writes:
This is science-fiction. There is no path that can take the current text synthesis models and turn them into super-scientists.
You don’t get super-science by modelling text that is itself an inaccurate model of reality. That’s not how any of it works.
First: yes, it is precisely science fiction. Three years ago, the vision of a fluent, formidable software correspondent was science fiction, too. Even if these systems are flawed —
Second: I don’t think it’s possible to say, with airless certainty, that “there is no path” from language models as we know them to super scientists or super scientist-enablers. Neither Baldur nor anyone else —
It’s totally possible these models won’t ever be useful for science, at least of the super variety; I can think of many obstacles, some having to do with physical reality, others with the social process of science itself. But none of them seem to me to rise to the level of like, epistemological showstopper. I don’t know, man! If Paul Dirac can sit down and think his way into the positron …
In my post, I laid out my counter-scenario: if, in a few years, there aren’t any signs of meaningful contribution to science from Claude’s successors, I’ll reconsider my position. It would be interesting to know if there’s any potential evidence that might make Baldur reconsider his.
To the blog home page