ebrso
7d ago
Executing drug dealers on the spot is an unmitigated good for society.
What does this even mean, "on the spot"? I assume it's something like, immediately at the place of action, i.e., as the drugs themselves are being dealt (or very shortly thereafter).
Surely, abandoning centuries of criminal procedure would mitigate some of the good for society. (I would still think this even if I were in favor of execution as punishment for drug dealing, which I'm not.)
ebrso
7d ago
What I was taught was that in the late 19th century experimental physicists did what was known as an "ether drift experiment" and famously found that light seems to travel at the same speed regardless of the motion of the reference frame. And that subsequently Einstein started with this assumption and worked out the seemingly bizarre theoretical implications.
I also recall being taught in school that Einstein's development of the postulates of special relativity (particularly that the speed of light is equivalent in all inertial reference frames) was influenced by the results of the Michelson-Morley Experiment, which failed to identify expected changes to the measured speed of light when the Earth's motion changed relative to the position of the sun.
However, this discussion suggests that Einstein's development of special relativity was probably not significantly influenced by the MM experiment, and was mainly influenced by assymetries in classical electromagnetism.
ebrso
7d ago
Research programs have supposedly been gutted by these results. "Careers ruined," as the narrative goes. I think that is an exaggeration or overreaction. There are thousands of mathematicians and their interests are very wide ranging. The odds of someone's research program or problem being "solved" by AI are very small.
I suspect you're wrong; I get the sense that the majority of the projects worked on by today's most productive research mathematicians have been very significantly disrupted by OpenAI's public releases. You can get a sense of the scope of these results by skimming the collection of mathematicians' reactions compiled here.
I don't see anything changing much. The status quo of journals and academia will remain intact even with rapid progress in AI, save for the occasional breakthrough in an unsolved problem.
Again, I suspect you're wrong; I expect the most important math journals will all pivot now to publishing primarily expositions of computer-generated results, which is a major shift. Soon, the expositions themselves will be computer-generated, too. As far as academia, the entire systems of reward, hiring, and promotion for research math will be unsettled (each of these systems is currently focused on the ability of individuals to develop proofs of mathematical statements).
A math PhD program, especially from a good school, teaches or guides future researchers towards productive areas of research that are of interest to the "math community," rather than results that are merely "personally interesting". AI cannot recreate such communities or networks.
I think you're making the point here that research math isn't solely about problem solving (proof development), and that exerting social influence over (e.g.) area focus and direction is similarly important. That's clearly true. Gauss wrote: "I confess that Fermat's Last Theorem, as an isolated proposition, has very little interest for me, because I could easily lay down a multitude of such propositions, which one could neither prove nor dispose of" - which I take as suggesting that the prominence of Fermat's Last Theorem owes as much to historical and social factors as it does to any inherent importance of the statement. But so what? Yes, the emphasis placed on particular problems emerges as a consensus across the work of mathematicians operating as a community. But this emphasis has no real bearing on whether humans or machines are better-positioned to tackle such problems, which was previously only ever the work of individual humans.
I suspect many of these successful proofs were prompted by experts who knew what to look for, rather than just naively trying. What is touted by the media as an AI breakthrough, may involve surgically precise prompting in the background by an expert in said field.
I think this is a fair point. At least right now, a lot of the success of the models in solving major math problems may owe to effective interactions with human experts in the form of seasoned prompting. We honestly don't know how significant the human collaboration was to generating these results. Many mathematicians have called for OpenAI to share these details.
Alluding to the "airplane meme," there is also survivorship bias: you don't see all the failed attempts with AI . Sometimes it works great, but other times hit or miss.
I don't understand this point. Certainly we don't deride great (human) mathematicians this way (look at all the problems Gauss couldn't solve). Maybe you're suggesting that humans today can still more effectively resolve certain types of math problems than machines can? If so, what evidence do you think supports this idea?
ebrso
25d ago
The standard print versions of the Talmud, which Jews actually study, specify "mi-Yisrael" or the soul of a Jew. And this gets to the heart of the problem of left-wing antisemitism, especially with notorious left-wing commentators like Hasan Piker. They always denounce antisemitism in their diatribe against Israel. When in fact the right-Wing antisemites are correct that this is not only a problem with Israel, it is actually a problem with the way Jews politically and culturally organize in the United States. It leads to wildly inappropriate and disparate treatment of Jews in which they get all sorts of public funding and support for their dedicated spaces, combined with unlimited cultural veneration and educational propaganda. [Emphasis added]
Can you help me understand what the pronoun this represents above? Does it refer to Jews systematically favoring the interests of their fellow Jews over the interests of non-Jews as part of the observance of an explicit religious obligation? Is it Jews hiding their general perfidy by systematically misrepresenting their religious beliefs to non-Jews? Something else entirely?
Also, what is the educational propaganda you have in mind here? You've called (many times) for a rejection of the broad historical consensus that several million European Jews were systematically murdered in WWII - is that what you mean? Are you talking about disinformation surrounding the historical settling / foundation of Israel and Israel's conduct in its historical / current wars? Propaganda about Israel's treatment of non-Jewish residents within its 1967 borders and the occupied territories? Propaganda about Jewish religious beliefs and obligations? Something else?
What is the evidence that current models are poor "conjecture and definition makers" relative to humans?
An alternative hypothesis is that this is a case of the "what happens when you post a real Monet and say it's AI?" phenomenon, in which human purported experts systematically downplay the strength of AI capabilities for complicated social / psychological reasons.
More options
Context Copy link