cs702

90 ms·

931.
by cs702 3y ago
Touché. I just updated my comment :-)
932.
by cs702 3y ago
Thanks for the thoughtful response. Unfortunately, I too think it will be a bit of an uphill battle for you. If you haven't already, take a look at Mojo and Julia. Both offer many of the benefits of Fortran, but unlike it, they are see
933.
by cs702 3y ago
Fortran! If you don't mind me asking, why Fortran? I know it underpins a lot of time-tested scientific code, often wrapped by libraries like PyTorch and Numpy, but Fortran isn't exactly a popular language nowadays. What's y
934.
by cs702 3y ago
If this is true, shame on the company .
935.
by cs702 3y ago
Can't argue with that! :-) For what it's worth, you can keep both, and make parallel vs sequential execution an option, with a boolean flag. You can also leave the sequential code as a comment explaining what the parallel code doe
936.
by cs702 3y ago
This looks really nice . Thank you for sharing it on HN! In case you didn't know, you can parallelize the slow Python loop in selective_scan that computes all the x's: x = torch.zeros((b, d_in, n)) for i in range(l):
937.
by cs702 3y ago
I have three acquaintances who own Teslas whose family members have gotten into accidents because they were looking at, and tinkering with, that gorgeous giant shiny screen... without first engaging any form of autopilot, not even cruise c
938.
by cs702 3y ago
I have three acquaintances who own Teslas whose family members have gotten into accidents because they were looking at, and tinkering with, that gorgeous giant shiny screen... without first engaging any form of autopilot, not even cruise c
939.
by cs702 3y ago
You're probably thinking of this article: https://www.experimental-history.com/p/the-rise-and-fall-of-...
940.
The rise and fall of peer review(experimental-history.com)
22 points by cs702 3y ago|2 comments
941.
by cs702 3y ago
Surely those seemingly smart anonymous reviewers now feel pretty dumb in hindsight. Peer review does not work for new ideas, because no one ever has the time or bandwidth to spend hours upon hours upon hours trying to understand new thi
942.
by cs702 3y ago
Physical buttons: * work much better than bad touchscreen software (which really sucks!), and * are much cheaper than great touchscreen software (which is very costly to develop and maintain!). Personally, I'd rather have great tou
943.
by cs702 3y ago
TL;DR: A prolonged closure of the Suez route... * would pause a major source of income for Egypt, which is already in the midst of a financial crisis; * would raise the costs of trade as shipping is rerouted around Africa, taking more time,
944.
by cs702 3y ago
Unmentioned by the OP are... cryptocurrencies , which surely are being used to move savings out of the country's financial system.
945.
by cs702 3y ago
Very sad, and evidently very hard to solve within existing societal constraints . A few days ago there was a recent story shared on HN describing a perfect example of why it's become so difficult to solve this horrible problem: https
946.
by cs702 3y ago
That's a huge drop. The only thing that comes to mind that might explain such a huge drop is... ...the university administration's lack of consistent standards when it comes to dealing with prejudice and attacks against differen
947.
by cs702 3y ago
That happened after the paper had been rejected the previous year!
948.
by cs702 3y ago
Thank you. You're right. I edited my comment so it refers to "important/popular" instead.
949.
by cs702 3y ago
I'd be wary of programmatic lists that claim to track the most important/popular recent papers. There's a ridiculous amount of hype/propaganda and citation hacking surrounding new AI research, making it hard to disce
950.
by cs702 3y ago
Build something of personal interest to you. Start by looking for similar open-source projects online. Look at the online posts of the authors. Then look for the papers that you think will be useful for your project. Before you know
951.
by cs702 3y ago
Biggest takeaways for me: * The papers and research that will be considered most important 10-20 years down the road, with the benefit of hindsight, are not necessarily those that receive the most attention today. * Be skeptical of any comm
952.
by cs702 3y ago
Copying and pasting Urban Dictionary's definitions of "to schmidhuber"[a], "schmidhuber"[b] and "schmidhubered"[c]: --- to schmidhuber When you publicly claim that someone else's idea that is remotely
953.
by cs702 3y ago
Hanlon's razor + Sinclair's dictum[a]: Never attribute to malice that which is adequately explained by ignorance, especially when such ignorance is motivated by financial interests. --- [a] https://www.goodreads.com&#x2
954.
by cs702 3y ago
> A condition has been written in the computer code to disable the ability to run a train if it spends at least 10 days in one of these workshops. Wow, that's just... so wrong.
955.
by cs702 3y ago
Yes, I agree, it should be required reading -- as part of an introduction to 20th century European history, of which a lot people today seem to be utterly ignorant. The parallels you allude to are... unsettling, at the very least. Alas, I h
956.
by cs702 3y ago
Nice work, cool result, and -- in hindsight -- not a surprising finding.
957.
by cs702 3y ago
...compared to 1+ year for the Empire State building: https://patrickcollison.com/fast Clearly, whatever we've been doing to solve the shortage of low-income housing in the US is not working . At all. We ought to try
958.
Why It Took 17 Years to Build 49 Housing Units in Los Angeles(wsj.com)
32 points by cs702 3y ago|14 comments
959.
There's a New Bot in Town(twitter.com)
6 points by cs702 3y ago|0 comments
960.
by cs702 3y ago
Developing, training, retraining, and delivering a competitive AI service today requires an army of expensive people and a huge investment in infrastructure. It's not cheap! Moreover, any improvement, whether it's better generatio

More ›