rd11235

8 ms·

1.
by rd11235 1y ago
Yeah - my guess is this was just a very roundabout solution for setting axis limits. (For some reason, plt.bar was used instead of plt.plot, so the y axis would start at 0 by default, making all results look the same. But when the log scale
2.
by rd11235 1y ago
I don’t understand the trend of individual studies making it to the top of HN, while metaanalyses are ignored, despite examining 10s or 100s of such individual studies in aggregate to reduce noise. There are we-are-human reasons, but are th
3.
Student kicked out of Columbia for LeetCode interview cheat tool [video](youtube.com)
2 points by rd11235 1y ago|0 comments
4.
by rd11235 2y ago
Good motivation for a PSA: This happens more and more often, and there is a fairly easy + popular workaround (which also comes with 99% ad blocking as a bonus). Just either set up pi-hole locally OR use a hosted DNS service that does essent
5.
by rd11235 2y ago
You didn’t mention an important point: speed. Suppose conda had projects. Still, it is somewhat incredible to see uv resolve + install in 2 seconds what takes conda 10 minutes. It immediately made me want to replace conda with uv whenever p
6.
by rd11235 2y ago
> The meticulous PT format and exercise selection allows them to achieve more muscle gain in 20 minutes per week than median trainees achieve in 2-3 hours. This is a strong statement that is presented without evidence, and which conflict
7.
by rd11235 2y ago
I agree but the opposite can be true too. Sometimes the narrator seems to target some general audience that doesn’t fit me at all, in a way that makes me cringe when I listen, until I stop listening altogether. In these cases I’d rather lis
8.
by rd11235 2y ago
An obvious question that isn’t answered (in this article - not sure about the paper itself) is whether feeding fructose results in MORE tumor growth than feeding glucose (or other sources of calories) Without knowing this, it doesn’t make a
9.
by rd11235 2y ago
> it was just the wrong abstraction: too easy to start with, too difficult to create custom things Couldn’t agree with this more. I was working on custom RNN variants at the time, and for that, Keras was handcuffs. Even raw TensorFlow wa
10.
by rd11235 3y ago
What causes comments like this one to get faded out? I guess many people flagged it? Isn’t it just a logical statement?
11.
by rd11235 3y ago
But the chain rule for ordered derivatives is exactly the backprop rule. It's just the mathematical representation of 'the simple implementation' I mentioned. I think what you're saying is that you find the process intui
12.
by rd11235 3y ago
I don't have a great answer. Most modern descriptions are shallow and/or unclear. My favorite discussions were actually in Werbos's original papers. A nice overview was Backpropagation through time: what it does and how to d
13.
by rd11235 3y ago
Actually by 'ordinary chain rule' I am referring to what you're referring to as 'the chain rule for partial derivatives'. It seems like backprop follows very quickly even from that, but it does not.
14.
by rd11235 3y ago
> chain rule is defined for partial derivatives I agree. That's what I'm referring to as 'the ordinary chain rule'. > so it's still technically just chain rule No. Go try to derive backprop for general DAGs us
15.
by rd11235 3y ago
Anyone who believes that this completes their understanding of automatic differentiation is tricking themselves. When your graph is a TREE, then everything is very simple, as in this post. When your graph is instead a more general directed
16.
by rd11235 3y ago
I’m surprised no one has mention AdGuard. In settings, blocking “Annoyances” is off by default. By toggling it to true, the Google login banner will be blocked, along with the “Chrome is the best!!11” banner and others.
17.
by rd11235 3y ago
It seems that many agree with this. At the risk of getting downvoted I want to share an opposing opinion: This way of thinking is not just unhelpful but even harmful. If one would often benefit from these checks while coding, then they shou
18.
by rd11235 3y ago
Flathabits looks great. I used to use a paid iOS app until they pushed enough UI changes and buggy updates to motivate me to create my own bare bones solution. Wish I had seen your app before that. (Mine is just a personalized spreadsheet,
19.
by rd11235 4y ago
My comment isn’t about the paper. It’s only about the surprising response it’s getting from HN. This small level of evidence can be found in some paper for almost any fitness claim one can imagine. (And that is not at all a knock against th
20.
by rd11235 4y ago
From the actual paper: > Methods Non-resistance-trained young adults were assigned to one of the four groups: CON-ECC (n= 14), CON (n=14) and ECC (n= 14) training groups, and a control group (n=11) that had measurements only. Not sure wh
21.
by rd11235 5y ago
Agree. To put it in the context of the article, if the author looked at getting desired outcomes as a video game, then this article is just a rant, and time would have been better spent analyzing the situations for how to improve. The recu
22.
by rd11235 5y ago
Agree. Current title is inaccurate and click baity. Also, the confirmation requires authentication (at least by default, unsure if this can be changed).
23.
by rd11235 6y ago
I've personally had (many) more signal issues than WhatsApp issues, and not one of the signal issues was accompanied by a banner. I am not vouching for WhatsApp. I just don't think we should pretend that Signal is more reliable th
24.
by rd11235 6y ago
The two that stand out are telegram and signal, and which is better for you depends on your personal balance of usability vs. privacy. I personally can't use signal. It's great that it's secure, but I can't tolerate freq
25.
by rd11235 6y ago
This article says this result may transform communication, but seems to make no mention of why . What improvements does this have over classical communication?
26.
by rd11235 6y ago
No. It's a random Chase customer.
27.
Ask HN: How do I convince Chase that I shouldn't be receiving other's info?
30 points by rd11235 6y ago|16 comments
28.
A Friendly Introduction to Cross-Entropy Loss(rdipietro.github.io)
5 points by rd11235 10y ago|0 comments