patagurbon

14 ms·

31.
by patagurbon 1y ago
None of the other universities have agreed yet. What do you mean by laser focused? Do you have specific policies to address this? If not then this is a natural part of the unfocused nature of knowledge work, and the natural weakness of huma
32.
by patagurbon 1y ago
There is a fundamental issue with this kind of federalism though in that it increases strife and could easily lead to civil war. Let’s say we get rid of Medicare, Medicaid, social security, and research funding at the federal level. What ha
33.
by patagurbon 1y ago
What are “they” doing now that is not conducive to scientific knowledge? > they produce a lot of anti-knowledge Regardless of the hilarity of that phrase being associated with universities over the attention driven news and internet that
34.
by patagurbon 1y ago
> whatever the mechanism Is straining to carry the rest of this sentence. I responded to an assertion that a policy by certain limited universities was attempted and deemed unfit by the natural process was still active. This is how insti
35.
by patagurbon 1y ago
Sure, that’s why I said the process is an approximation of truth and only then in the limit. This is known in the scientific, philosophical, and research communities. It is a reality that is only solved by the slow inexorable application
36.
by patagurbon 1y ago
> universities pushing restorative justice Universities? Can you point to a university whose explicit policy is restorative justice? That’s not how universities generally work. You’re probably taking issue with an individual department
37.
by patagurbon 1y ago
I am not in favor of mandated DEI statements outside of basic respect for students and colleagues, I’m not sure why that’s relevant or why you responded to me about it. Most universities have moved away from those.
38.
by patagurbon 1y ago
They are not reasonable. Ideological “balance” has no place in any of the sciences. Science is about truth not balance.
39.
by patagurbon 1y ago
An eye for an eye is not how scientific progress is made.
40.
by patagurbon 1y ago
This push to force ideological “balance” on universities is incredibly dangerous. The pursuit of truth is difficult and has its pitfalls but it naturally leads to the dominance of certain viewpoints, which hopefully approximate the truth. W
41.
by patagurbon 1y ago
It does seem fairly arbitrary. It’s not a list of schools with big labs, since Hopkins or Berkeley aren’t there. It doesn’t seem to be private vs state schools, it’s not Ivy League or blue state only. One potential reason to select a divers
42.
by patagurbon 1y ago
Zygote hasn’t been the “canonical” autodiff library for some time now, the community recognized its problems long ago. Enzyme and Mooncake are the major efforts and both have a serious focus on correctness
43.
Parting ways with our Julia simulation after 100M miles(youtube.com)
1 points by patagurbon 1y ago|1 comments
44.
by patagurbon 1y ago
If you write Julia similar to Fortran, with explicit argument types and for loops and avoiding allocations it shouldn’t be too far off. Fortran IIRC has a few semantics which might make it more optimal in a few cases like aliasing But ind
45.
by patagurbon 1y ago
I didn’t say it was illegible. I said legibility is paramount, and I don’t think it makes the right trade offs in that regard to be a great teaching language It’s far more legible for numerics than a lot of languages, maybe except Julia and
46.
by patagurbon 1y ago
The post dismisses Julia quite quickly, especially since it is a language essentially purpose built to teach numerical linear algebra. Numerical methods is taught in Julia in at least a dozen universities I'm aware of, including MIT. U
47.
by patagurbon 1y ago
It’s likely ticks per second. Fairly common terminology in video game servers
48.
by patagurbon 1y ago
AI (apparently) needs much lower precision in training and certainly in inference than gaming requires though. A very very large part of the die on modern datacenter GPUs is effectively useless for gaming
49.
by patagurbon 1y ago
Only in the abstract / around election time, much in the same way they run on cutting taxes for the poor and middle class. There are Republican run states that do better on certain diminished state capacity things like zoning (Texas) o
50.
by patagurbon 1y ago
There is a 3rd point, that craftsmanship, machining skills, and logistics experience don’t fall from trees. As the US moves certain “lower” tier manufacturing overseas it can risk losing the experience needed to create tooling for productio
51.
by patagurbon 1y ago
Julia is effectively hard locked to LLVM. Large swathes of the ecosystem rely on the presence of LLVM either for intrinsics, autodiff (Enzyme) or gpu compilation. Nevermind Base and Core. The compiler is fairly retargetable, this is an acti
52.
by patagurbon 1y ago
You would have to have a very complete audit trail for the LLM and ensure the paper shows up nowhere in the dataset. We have rare but not unheard of issues with academic fraud. LLMs fake data and lie at the drop of a hat
53.
by patagurbon 1y ago
Impeachment of Kristi Noem could be more likely to succeed though.
54.
by patagurbon 1y ago
Unlike a lot of government spending research spending provably increases revenues by more than expenditures.
55.
by patagurbon 1y ago
I would counter your anecdata with the 5 friends I have, all of whom are whiter than printer paper and 3 of whom are deeply conservative, who received GRFP. Your failure to get GRFP had nothing to do with the diversity statement.
56.
by patagurbon 1y ago
They’re great for teaching. But I agree for actual work.
57.
by patagurbon 2y ago
The US is not really in trouble because of maintaining global hegemony. It’s in trouble because of repeated tax cuts by Republicans that are too popular for Democrats to fully unravel, and deficit spending largely caused or enacted by Repub
58.
by patagurbon 2y ago
The US was very good for a very long time at integrating immigrants. It should continue that tradition and work even harder at it. I believe it was some Republican president who said something to the effect of “if you move to Germany you ma
59.
by patagurbon 2y ago
No system will optimally allocate resources. However projects are typically funded under competitive grants and that process is fairly good at moving slowly but methodically in the right direction. Even when it doesn’t, it is training resea
60.
by patagurbon 2y ago
US research funding is not what you want to cut though. It is among the most productive funding possible and there is evidence aplenty that it pays for itself many times over. University bureaucracy is by and large fairly small for researc

More ›