MITSardine

12 ms·

91.
by MITSardine 1y ago
Necessity is the mother of invention! Another way to look at it, is only those who could write very optimized code could get anything meaningful running on those machines. Hence all (?) software from far enough back was optimized down to de
92.
by MITSardine 1y ago
So, the CRAY-2 built in 1985 was rated at 1.9 GFLOPS. The M2 the author uses seems to be benchmarked at around 6GFLOPs per core [1]. So this 6% of the author's CPU (assuming charitably this is a single core) corresponds to about 20% of
93.
by MITSardine 1y ago
The app taking "only" 6% of a $3k laptop's CPU to show two lines of text and a little 2D animation also gave me pause, but maybe this is just how it is. Still I wonder how people were squeezing pacman out of prehistoric machi
94.
by MITSardine 1y ago
They mention 1.5 million CPU hours https://journals.aps.org/prx/abstract/10.1103/PhysRevX.15.02...
95.
by MITSardine 1y ago
Load bearing print statement! I've had a few of those, I think they're usually a symptom of stack corruption.
96.
by MITSardine 1y ago
Gradient descent is not a learning algorithm. It is a simple iterative algorithm that goes from one point to the next. It doesn't even have memory of previous steps (caveat, the authors used BFGS which approximates the Hessian with p
97.
by MITSardine 1y ago
Shape optimization can be done under constraints of manufacturability, which would avoid exotic shapes that can't be worked with. Why would that be the case, though? It's conceivable that optimal shapes be very different from what
98.
by MITSardine 1y ago
This irks me to no end, why not just call it applied mathematics algorithms to not use specific terms, rather than AI. Is grep AI? Is your web browser AI?
99.
by MITSardine 1y ago
This doesn't even seem to be ML, though.
100.
by MITSardine 1y ago
I would distinguish between: - methods that were devised with domain knowledge (= numerical methods) - generic methods that rely on numerical brute forcing to interpolate general behaviour (= AI) The qualitative leap is that numerical brute
101.
by MITSardine 1y ago
I think this disregards the costs associated with using AI. It used to be you could learn to program with a cheap old computer a majority of families can afford. It might have run slower, but you still had all the same tooling that's f
102.
by MITSardine 1y ago
If I may play the devil's advocate, nothing is deterministic. A neutrino could cause a bit flip in your calculator. More commonly, the lower abstractions we build upon without knowing their innards can have bugs. Even the most popular
103.
by MITSardine 1y ago
I would argue that the ratio of work to breakthroughs is not a form of inefficiency, but something inevitable about the nature of breakthroughs. In my opinion, a breakthrough is not the production of new knowledge, it is rather its adoption
104.
by MITSardine 1y ago
Beg to disagree, plenty of people love puzzle games. This doesn't need to appeal to everyone. The article cites two examples of such games that lead to scientific progress.
105.
by MITSardine 1y ago
Reverse ant algorithms?
106.
by MITSardine 1y ago
Super computers are being used daily for much toy-ier codes in research, be glad this at least interests the public and constitutes a foray of academia into new areas.
107.
by MITSardine 1y ago
University and research clusters are built to run research code. I can guarantee this project is 10x as impactful and interesting as what usually runs on these machines. This coming from someone in the area that usually hogs these machines
108.
by MITSardine 1y ago
I think EPFL and ETH are generally well known internationally, but Switzerland being rather small (9M pop), it's only natural you don't hear much about it compared to other larger countries!
109.
by MITSardine 1y ago
What do you call the origin of a vector? I don't think that's a notion in vector spaces.
110.
by MITSardine 1y ago
Sorry, I couldn't find the page in English, but what you're talking about is a Hilbert basis: https://fr.wikipedia.org/wiki/Base_de_Hilbert . There is a paragraph on this in the orthonormal basis English page
111.
by MITSardine 1y ago
This idea is also at the core of the Finite Element Method (and other Galerkin methods, such as DG) which are probably the most widespread methods to approximate the solutions of partial differential equations (PDEs) used everywhere in phys
112.
by MITSardine 1y ago
That's not really the assumption of the article, that's the assumption the article makes of the reader's knowledge of vectors as they begin reading. The article is trying to make the point that functions are also vectors, str
113.
by MITSardine 1y ago
I'm not sure I'm sold on the necessity to detect lines and planes specifically. My issue is, suppose you could do that perfectly, then what of the rest of the geometry? If you're aiming for a CAD model in the end (BREP), you&
114.
by MITSardine 1y ago
Have you tried "traditional" approaches like a Delaunay triangulation on the point cloud, and how does your method compare to that? Or did you encounter difficulties with that? Regarding what you say of planes and compression, you
115.
by MITSardine 2y ago
What does this context window mean, is it the size of the prompt it can be made aware of? In practice, can you use any of these models with existing code bases of, say, 50k LoC?
116.
by MITSardine 2y ago
If my C++ project is a simple utility supposed to take some files, crunch numbers, and spit out results, is there still the possibility it can be used for nefarious purposes?
117.
by MITSardine 2y ago
I insist that "monopoly" is not the correct term. There is a much more precise term (with fewer associations): exclusivity. Saying "the heart has a monopoly on pumping blood" is understandable, but a better sentence woul
118.
by MITSardine 2y ago
Are you talking about the Turkish grad student at Tufts who got arrested and deported to Louisiana recently?
119.
by MITSardine 2y ago
Though people may use CPUs several years old, they generally weren't at the moment they were bought, and the decision came from comparing with the competition. This argument of "my software will be faster when computers are faster
120.
by MITSardine 2y ago
The hardware is a real selling point, I researched a little before buying my current MBP, but none of the non-mac laptops convinced me. The Mac ARM processors are just that good, it seemed I could only get similar performance with twice the

More ›