ubj

14 ms·

121.
by ubj 3y ago
This type of problem has been going on for weeks on multiple Linux distros. See for example the Github issue [0]. [0]: https://github.com/bitwarden/clients/issues/6560
122.
by ubj 3y ago
If this has an offline mode, I'm sold. If not, I'm unconvinced it's much better than Notion yet.
123.
by ubj 3y ago
Fair points. But given the potential impact this would have in practice, I'm still surprised there isn't more interest in exploring whether it's possible to scale the approach. Transforming neural network training into convex
124.
by ubj 3y ago
If an appropriate dual problem is found this could be possible. I'm baffled why Mert Pilanci's work in this area hasn't received more attention. His proofs of a zero duality gap for neural networks are impressive.
125.
by ubj 3y ago
The paper only gives a representation result, so no method is known yet for constructing the component functions `g` for the discontinuous case. See the last paragraph of the Conclusion. It may turn out that the correct representation can b
126.
Kolmogorov Neural Networks can represent discontinuous functions(arxiv.org)
133 points by ubj 3y ago|31 comments
127.
by ubj 3y ago
Very cool! Now what would be really interesting is combining this with steganography techniques to hide the horcrux fragments inside totally unrelated-looking files.
128.
by ubj 3y ago
PackageCompiler.jl is more fully featured, and I'm definitely looking forward to further developments :) But it still has some pain points: * Massive executables (which you mentioned). This makes it very difficult to use with embedded
129.
by ubj 3y ago
Yep, I was aware of StaticCompiler.jl. I wish it was more mature. Static compilation is indeed possible with Julia. But it's very limited in its capabilities and certainly not as effortless as a simple `mojo build myfile.mojo`.
130.
by ubj 3y ago
In my opinion, the issue that will make more of a difference in the long run is Mojo's first-class support for AoT compiled binaries (as well as JIT compilation). Julia's poor AoT support (with small binaries) is a major Achilles
131.
by ubj 3y ago
Second this point, especially when flying with young children. But I imagine that slight modifications could be made to get a hybrid approach.
132.
by ubj 3y ago
Thanks for sharing these articles. I've been interested in MLIR ever since learning that the Mojo language would integrate with it [1], but it's been hard to find reference material geared towards beginners. [1]: https:/&#x2
133.
by ubj 3y ago
Apparently commercially produced "foot mouse" devices have been marketed for quite some time, as early as 1985 [1]. TIL. [1]: https://en.m.wikipedia.org/wiki/Footmouse
134.
by ubj 3y ago
I don't see a description on the GitHub page of any experiments performed showing the model can pass an exam. Where are these details, and how was the testing performed? EDIT: The git commit was 4c6d52a when I originally posted this co
135.
by ubj 3y ago
I wish I could upvote this twice. This is the best basic introduction to concepts in functional analysis that I've seen. Another great overview that goes deeper into the math is [1]. Another fantastic application that the website doesn
136.
by ubj 3y ago
Well, a couple key differences from Mojo are 1) LPython is open source (BSD 3-clause), and 2) LPython is available to download and use locally today. Mojo is still only available on private Modular notebooks. Granted, I'm optimistic fo
137.
by ubj 3y ago
Fantastic, looking forward to trying this out! And good point about JIT--it does have its place. What I should have said is that I wish there were better options with high quality support for _both_ AoT and JIT. Most of the options I'm
138.
by ubj 3y ago
Looks interesting! Thank you for focusing on AoT compilation and not just JIT compilation. To be honest, I'm sick of JIT compilation. In theory it seems like the best of both worlds, but in practice it turns out to be the worst of both
139.
Hyundai Owner Rebadges Their Car as a Ford to Deter Thieves(thedrive.com)
3 points by ubj 3y ago|1 comments
140.
by ubj 3y ago
This may be relevant: https://www.sematic.dev/blog/tuning-and-testing-llama-2-flan... It's the most straightforward explanation I've found so far. I'd love to hear if anyone's found something better
141.
by ubj 3y ago
Disagree with this comment. See the Wikipedia page on Explainable AI [1]. Many models today fall under the "black box" category described on that page. Also note the last paragraph in the Criticism section: > The goals of XAI a
142.
Even the scientists who build AI can’t tell you how it works(vox.com)
7 points by ubj 3y ago|4 comments
143.
Microsoft's Outlook.com is down again on mobile, web(bleepingcomputer.com)
4 points by ubj 3y ago|0 comments
144.
by ubj 3y ago
Another interesting robot that was created a year or two after this article: https://techfinder.stanford.edu/technology/new-soft-robot-sh... Veritasium did a video segment on it: https://youtu.be/058hRt
145.
by ubj 3y ago
I'll second this. I've been a part of several fantastic church communities in the various places I've lived, both before and after I had kids. And yes, a great side perk of having children is connecting with other parents :)
146.
by ubj 3y ago
> No, it did not “double-check”—that’s not something it can do! And stating that the cases “can be found on legal research databases” is a flat out lie. What’s harder is explaining why ChatGPT would lie in this way. What possible reason
147.
Two-thirds of North America at risk of energy shortages during extreme demand [pdf](nerc.com)
117 points by ubj 3y ago|176 comments
148.
by ubj 3y ago
Second this. Not only is it faster, but the error messages in Mamba are much more helpful and sane.
149.
by ubj 3y ago
I realize there are different opinions about this article, but I personally agree with the overall idea. Learning is about recalling information. The more you practice recalling a fact, the more the memory "sticks".[1] Writing sum
150.
LAGrad: Statically Optimized Differentiable Programming in MLIR [pdf](pengmai.github.io)
2 points by ubj 3y ago|0 comments

More ›