Centralization and IT-ification has made flouting difficult. There’s one common course site on the institution’s learning management system for all sections where assignments are distributed and collected via upload dropbox, where grades are tabulated and communicated.
So far, it’s still possible to opt out of this coordinated model, and I have been. But I suspect the ability to opt out will soon come under attack (the pretext will be ‘uniformity == fairness’). I never used to be an academic freedom maximalists who viewed the notion in the widest sense, but I’m beginning to see my error.
I teach math at a large university (30,000 students) and have also gone “back to the earth”, to pen-and-paper, proctored and exams.
Students don’t seem to mind this reversion. The administration, however, doesn’t like this trend. They want all evaluation to be remote-friendly, so that the same course with the same evaluations can be given to students learning in person or enrolled online. Online enrollment is a huge cash cow, and fattening it up is a very high priority. In-person, pen-and-paper assessment threatens their revenue growth model. Anyways, if we have seven sections of Calculus I, and one of these sections is offered online/remote, then none of the seven are allowed any in person assessment. For “fairness”. Seriously.
Ok, now I understand your reply. In retrospect, you were being perfectly clear all along and my confusion was due to me not appreciating the precise definition of a CRDT. Thanks!
This is a natural point of confusion. The true (IMO) primitive concept here is the probability measure. Probability measures on the real line are in canonical bijection with CDFs, the latter being axiomatizable as càdlàg functions (see https://en.wikipedia.org/wiki/Càdlàg) asymptotic to 0 (resp. 1) at minus infinity (resp. infinity). On the other hand, not every probability measure has a density function. (If you want the formalism of densities to capture all probability measures, you need to admit more exotic generalized functions à la Dirac.)
The list in the above comment isn’t a summary — it’s a precise definition. It can and must be carefully explained with lots of examples, contrasts with other languages, etc., but the precise definition itself must figure prominently, and examples and intuition should relate back to it transparently.
Is NVIDIA's JIT-based approach here similar JAX's, except targeting CUDA directly rather than XLA? Would like to know how these different JIT compilers relate to one another.
Number of job posts itself seems like a better measure of the state of the job market than mentions or mention-density. What aspects might mentions capture that number of postings wouldn’t?
The article does not claim that smoking determines outcomes for any individual. It gives statistical evidence that outcomes (measured brain activity while performing certain tasks) for weed smokers are worse than for non-smokers __on average__, and that this pattern persists in various subpopulations determined by demographic and lifestyle factors.
Cryptanalysis relies on deep conjectural heuristics in analytic number theory. These conjectures becoming theorems wouldn't affect cryptanalysis at all, because their validity is already baked in. If, however, any of these conjectures turn out to be false, there would be ramifications.
Correct. HTMX doesn’t deal with client-side state and updating UI in response to client-side interactions. React, on the other hand, is all about that stuff.
One advantage of (Named)Tuples over dataclasses or SimpleNamespaces is that they can be used as indices into numpy arrays, very useful when you API is returning a point or screen coordinates or similar.
I totally get why people are infuriated by rationalizations like "inflation rates are now good". Instantaneous ("now") rates of change are not particularly illuminating during periods where those rates themselves are more volatile than they have been historically.
It makes sense (to me) to average inflation over the four year electoral period. The average inflation over the Biden years 2021-2024 was 5.3%, versus 1.9% over the Trump years 2017-2020 [1]. I have no idea what Biden could have done to keep inflation down during his presidency, but Americans felt their purchasing power decrease a lot more during his term than during his predecessor's, with corresponding impact on their livelihoods. They have every right to be pissed off. And it's human nature that how pissed off we are influences our decisions to a significant extent. Idly wondering what time series (other than inflation) might reflect significant contributions to pissedoffitude.
This is good intuition for why ensembling overparametrized is a good idea. Doesn’t speak to why ensembles of tree-structured estimators in particular perform so well compared to ensembles of other nonparametric estimators.
I echo your experience. Effexor/Venlafaxine are absolute poison. I tapered off of it slowly over the last two months under doctor’s supervision. The day after I took my last dose — small thanks to the tapering — I’ve had constant brain zaps, wobbly vision, horrible perpetual nausea, cold sweats, and inability to sleep more than an hour at a time. It’s been two weeks so far of lying in a dark quiet room waiting for things to resolve so I can go back to my wife/kids/job/life. Still no noticeable improvement. On sick leave for the first time in my life.
Which problems people work on is dictated to a large extent by the need to publish to keep your job. There is a lot of incentive to work on publishable low-hanging fruit problems. Hence the abundance of “write-only” journals in mathematics.
I don’t think there is by any means a shortage of hard, interesting problems. But working on them directly comes with significant career risk.
I suppose you get performance benefits if the the time it takes to start up a nodejs process dominates the execution time of the script. This is probably the case for a decent proportion of “serverless function” type scripts.
It varies, but it often comes down to deep expertise combined with creativity, years of toil, and standing on the shoulders of giants. Cf. Fermat’s Last Theorem, bounded gaps between primes, the Weil conjectures, the Poincaré conjecture, etc.
In my mind, the justification for referring to RH as the “biggest” open problem in mathematics is its importance in numerous fields of mathematics. It’s hard to take three steps in analytic number theory without running into Riemann. Zeta functions — and with them RH — have transcended number theory and are fundamental objects of study in other mathematical fields, most prominently representation theory and the Langlands Program.