The evidence for physics is mediated through structures inhabiting the other layers - think what is involved in your cognition of some result from CERN.
The propositions of physics carry an implicit Galilean ideal that they should be indifferent to context and so to speak hold indifferently across space and time. You are coding this surprisingly fruitful desideratum into the (btw basically theological) conception of ‘law’. With this, other discourses end up destructured by considerations alien to them and the element of specifically physical epistemology goes out the window.
Is there a known headcount for OpenAI influences v Anthropic influencers?
On threads like this, this site seems to be made of nothing but boosters for one or the other, with their emphatic professions of faith, all based on inscrutable, unverifiable inner experience. When no one bothers even to reflect what conditions a proper assertion would require, the discourse is pure faith propositions.
The objections that one treatise makes to another will typically be of type violation, category mistake, nonsense - which is typical of compilers. I don't need to tell you which discipline invented the ideas of type and type error and was even largely organized around them in early and mid- 20th c, even across wildly different theoretical views.
When I was a young man I took three philosophy courses from a very old man who, when he was a young man, was the dissertation director of Alan Turing. The latter, by the way, was an habitué of the seminar of Wittgenstein.
The experts on hospital ethics panels are trained internally by medical schools, which are like little universities inside universities. There has not been philosophical access for about thirty years.
It’s actually pretty bad at it. I think there just isn’t enough literature to get a good effect from the LLM approach. Good luck with verifiable rewards when the target discourse is effectively pure self-criticism. Maybe ‘disputable rewards’ …
I have found that with proper framing I can get good help from Claude and ChatGPT on questions of translation of haute German philosophy and, to my amazement, Ancient Greek. An immediate ‘translate this passage’ request is a cataclysmic disaster. The nexus of sentences differs from other forms of discourse.
I would have thought so too, a priori, but at this point three former colleagues are working for Anthropic; the most extraordinary case, one of the brainiest people I have known, was announced this week.
A universal document converter knows what document it is working with and what to do with it once it has it. 'What a document is' is an AST that has resulted from a few thousand years of literate civilization. You can detect the outline of this AST - or AAST as you might call it - by asking what must be preserved in a different printing of the same, or in a translation.
A universal document converter is 'expected' to admit transformations on the AST of a document. Luafilters do this more or less directly; operations via json representation do it in another.
I never used luafilters before, not knowing lua, but these days use them all the time for simple problems and am getting a clearer picture of the possibilities. This is because claude and codex write luafilters at the drop of a hat.
One simple illustration I have found of use with academic writing published inter alia in html arises from the willful decision of the html bureaucracy never to include a footnote syntax - and thus fall short of ABCs of any document concept however narrow and curtailed - because having said 'o we don't need footnotes, we have hypertext' back in clintontime they are too proud to change. In fact of course html is the format par excellence of footnotes ... as a gander at wikipedia will tell you. Pandoc can't parse them out of html - including its own html - since there is nothing to parse: the reader recognizes them by inspection in the browser. But you can ask claude to write a lua filter e.g. recognizing pandoc's own html footnotes - which are as arbitrary as everyone else's - and generate the structure intended by the author, in which they are footnotes.
Why shouldn't the goalposts move? That it was possible to beat or tie a chess master, if you had enough computational power, was basically the content of a theorem of Zermelo over a hundred years ago. It differs not a whit from tic-tac-toe. Even Eliza was practically passing the Turing test, which seems comically silly now. There's just an incredible amount of computational power so all sorts of things are possible that were formerly unimaginable - like training LLMs on the whole corpus of extant human discourse.
The purpose is the destabilization and discrediting of democracy. The Beijing state, in concert with Putin and ex-Khamenei, has no other purpose but to prove there is no democratic alternative. Sure, it turns out you get free things as e.g. one got free RT America.
You are saying: it turns out that one-shot playable Minecraft clones are actually pretty simple. Maybe it seemed a hard problem to programmers, but why not just say that the verifiable-rewards training has shown that their skill is unusually simple?
Teacher salaries have collapsed since the fifties - when, by the way, we could get the brainuest element of half the population due to institutional sexism. My jr high teachers would be federal judges today. They were worth about 7 times what we paid them then; we pay them less in relation to other highly trained services today.
The propositions of physics carry an implicit Galilean ideal that they should be indifferent to context and so to speak hold indifferently across space and time. You are coding this surprisingly fruitful desideratum into the (btw basically theological) conception of ‘law’. With this, other discourses end up destructured by considerations alien to them and the element of specifically physical epistemology goes out the window.