antpls

55 ms·

541.
by antpls 8y ago
In addition to other comments : I guess "threat actors" includes non-human autonomous hacking systems ("AI"), and humans (or organizations) who are neither good or bad intended, but whose actions happen to have negative
542.
by antpls 8y ago
Benchmarks about mobile NN frameworks : https://github.com/XiaoMi/mobile-ai-bench , not sure it contains ArmNN, but it's a start and you could open an issue about it maybe
543.
by antpls 8y ago
It's nice to see that Spark is on par with Unix tools on low data volumes. The Spark program can be applied as-is for data volume 10 times or 100 times bigger without being modified. You would have to reinvent Spark if you started from
544.
by antpls 8y ago
Cannot it be reused because of the splash or because of the salt? What about splash landing in a natural/artificial lake?
545.
by antpls 8y ago
I understand, I read my comment again a few hours later and it could be misinterpreted. Or maybe the parent edited the original message (can't remember) I wanted to point the fact that the argument "exploiting people who are bad a
546.
by antpls 8y ago
Well, if you were better at math, you wouldn't give your free time to post content on HN, would you ? As soon as you talk about future costs, different people have different believes and different bets, because they have different risk
547.
by antpls 8y ago
Firefox is already valuable for browsing on mobile phone, where there is not much space on screen to have Dev console anyway. I recommend trying Firefox Mobile
548.
by antpls 8y ago
> Wasting intellectual horsepower by refuting bad arguments, experts have less and less time for more productive endeavors I disagree that it's a waste of effort. It may be considered a waste of effort from the experts perspective,
549.
by antpls 8y ago
> Fundamentally, a schema is a guarantee that your data conforms to the structure you intend. My opinion : The reality is that you never have enough information about how the data will be queried and connected together in the long run. A
550.
by antpls 8y ago
Are neural networks (for example from TensorFlow) considered data structures? They store data, so maybe it would have been nice to mention them in the paper even if it's to put them out of scope for X reasons.
551.
by antpls 8y ago
Can you prove it?
552.
by antpls 8y ago
Those rankings based on countries don't say much IMHO. It depends on : - the score. If the top 5 achieves 85/100 and the the rest of the top 20 achieves 83/100, they are all on par. - the field. If it's academic with ope
553.
by antpls 8y ago
Why not Julia : https://juliacomputing.com/about-us.html 27 men and 1 woman, which is a bad parity score, even for a tech company.
554.
by antpls 8y ago
In all those papers, there at least one member of the team in the author list, or a member of a partner organization (Intel, Lawrence Berkeley National Laboratory, etc). I wouldn't call them fair independent reviews of the language.
555.
by antpls 8y ago
Can't even access the article with javascript turned off.
556.
by antpls 8y ago
This was published on their own website, it is therefore biased toward a good perception of them. It doesn't seem to be a fair and independent review to help build an opinion.
557.
by antpls 8y ago
Again, you don't give any sources or references about what you are stating. If I can chose an indicator, then I pick my personal experience. I live in Europe and I'm happy with the digital experience of social services I have. I&#
558.
by antpls 8y ago
> Especially not from a user perspective, if you’re a social worker you now have to know how to use 5-10 complicated and error prone IT systems, on top of your regular job, and we just keep on adding more. I mean,those 5-10 systems are l
559.
by antpls 8y ago
> we have the comfortable lives we have now "We" = the 1% which reads Hacker News? Speak only for yourself You uses fallacies to spread misinformation about war. No source, no number, no reference. Convenient.
560.
by antpls 8y ago
Not sure why you are downvoted, I agree with you. At that point it becomes ridiculous. Depending on the license, you might fork the whole project and change the name to a genderless name. Keep the project in synch with upstream, only change
561.
by antpls 8y ago
It's beautiful that the reduced model itself is only 175 lines of Python code, thanks to TensorFlow : https://github.com/openai/gpt-2/blob/master/src/model.py
562.
Google and Huawei Published More Papers on ArXiv about Federated Learning(arxiv.org)
1 points by antpls 8y ago|0 comments
563.
by antpls 8y ago
I would drop the concept of "paper", and use a unique, gigantic graph of hypothesis and experiments that keeps growing as research goes on. Everyone would contritube to the same graph without duplicate data, or unaccepted piece of
564.
by antpls 8y ago
Regarding me, I could be happier without working, or working less than I currently do. So, according to the article : "this is a triumph"
565.
by antpls 8y ago
Just curious : why tipping in cash when they are numerous apps to send and receive money between "friends" (ala PayPal) ? (By the way, how long before we can tip people on WhatsApp?)
566.
by antpls 8y ago
One might ask in a first place : who needs $10,000/month to live (including water, food, private place to sleep, and some entertainment) ? If some people can live with $1,500/month, how come other receive so much and for what? I g
567.
by antpls 8y ago
I think the point made by paulgb is that it won't go unnoticed (it will accumulate records) which is a first step to fight an injustice
568.
by antpls 8y ago
If you are a Google employee and you don't care about the executive visions, you are a living proof that Google is made of people who don't even understand what they are working for, beside having a good salary imo. When you wor
569.
by antpls 8y ago
Maybe you should send your code to the PyPy team. I bet they will look into it and see how they could improve their JIT compiler. Last time I checked the project, they had a full benchmark running for each release to spot regressions and to
570.
by antpls 8y ago
> This is wonderful. This could become the best way to move Python projects to Rust: initially just run on the RustPython interpreter, but then optimize low level routines in Rust. In 15 years I wouldn't be surprised if this or some

More ›