argonaut

69 ms·

661.
by argonaut 10y ago
I was referring to ImageNet as to what I could do (because my experience is in computer vision). But more broadly, ensembling in general is really really obvious . I mean, for heavens sake, it's the trick everyone uses on Kaggle. You
662.
by argonaut 10y ago
Not to be too harsh, but if all you did was ensemble several different embeddings using methods that were mostly from other papers, it would be pretty obvious you'd get state of the art performance. But that is not very interesting fro
663.
by argonaut 10y ago
in the first place . They will eventually sell in the US.
664.
by argonaut 10y ago
Not true. If Xiaomi sells anything in the US, they can definitely be sued.
665.
by argonaut 10y ago
I'm very happy with my iPhone. I'm certain most iPhone users are happy with theirs.
666.
by argonaut 10y ago
The original poster was talking about what would be better for the HN community. The HN community is a discussion forum, not a system of people privately emailing.
667.
by argonaut 10y ago
The downturn is in startups. I see these bootcamps graduates being snapped up by large corporations all the time. Including lots of companies that are not as sexy as the big Silicon Valley firms, yet still desperate for tech talent.
668.
by argonaut 10y ago
I'm with you until you started talking electric vs. gas vehicles. That has nothing to do with autonomous vehicles.
669.
The Skynet Salesman(multithreaded.stitchfix.com)
1 points by argonaut 10y ago|0 comments
670.
by argonaut 10y ago
Since you're the one making the insinuations/argument, you should be answering those questions for us.
671.
by argonaut 10y ago
Wrong issue. That's for patent lawsuits.
672.
by argonaut 10y ago
1) HN has 300k daily uniques ( https://news.ycombinator.com/item?id=9219581 ). Of those readers, nearly all do not comment (we don't have 300k comments, do we?). Of those commentors, only a fraction comment on a given
673.
by argonaut 10y ago
The judge, a human who possesses common sense , sees you've registered lots of domains and spammed C&D letters to Google, calls you out on your BS, and orders you to pay Google's legal fees?
674.
by argonaut 10y ago
For any given arxiv article, I'd wager 98% of people on HN are unqualified to critically review the article. They're not aware of prior work that might contradict the article. They're not aware of the limitations of prior wor
675.
by argonaut 10y ago
Yes, but I find a $400M retention package to be highly implausible. We're talking the kind of compensation that Google/Facebook would spend to retain 40 top executives.
676.
by argonaut 10y ago
Do you have inside knowledge we're unaware of? Otherwise it's hard to say "in reality...." $1B is plausible, but much more questionable given this article.
677.
by argonaut 10y ago
I don't think it's plausible. Sure, it is common for the majority of payouts in acquihires that are in the $1M-$50M range to be through handcuffs. When a company is acquired for $600M, that is not an acquihire, and I would be very
678.
GM paid less for Cruise Automation than was reported(businessinsider.com)
47 points by argonaut 10y ago|21 comments
679.
by argonaut 10y ago
Neither. The values he gives are reasonable (not bad choices) but also quite broad (covering several orders of magnitude usually - not valuable per-se). Of course, things become much clearer when you actually know what the hyperparameters r
680.
by argonaut 10y ago
Because your logic is obviously fallacious (e.g. moving the goalposts / excluding contrary evidence).
681.
by argonaut 10y ago
Sure Google and Intel don't have great executives that leave for other companies, if you exclude the really successful ones. /s It's also incredibly rare for executives to leave these companies in the first place. They are co
682.
by argonaut 10y ago
The elephant in the room: Sheryl Sandberg (VP of sales and ops at Google before FB). John Doerr was also a sales executive at Intel in the late 70s. Also, most executives don't become famous. There are dozens of ex-Google and ex-Inte
683.
by argonaut 10y ago
That patent war was started by Microsoft under Steve Ballmer and both companies dropped it after Satya Nadella took over. I genuinely don't believe Microsoft under Nadella would go to patent war except defensively. Everything I've
684.
by argonaut 10y ago
I'm pretty sure Andrew Ng was hired by Baidu as a recruiting tool / evangelist. He hasn't done much research recently.
685.
by argonaut 10y ago
You're mistaken. Convolutional neural networks require annotated data. It's a TED talk; they're obviously not going to mention the very boring details of their annotation protocols.
686.
GM says it bought Cruise Automation for about $600M(twitter.com)
2 points by argonaut 10y ago|0 comments
687.
by argonaut 10y ago
The solution was specific to that subproblem. The left/right-offset cameras were for the sole purpose of providing examples of what it would look like if the car was deviating off path. The same trick would not work for any other probl
688.
by argonaut 10y ago
Sure, some people are exploring the approach. This does not mean it works.
689.
by argonaut 10y ago
Always fascinating how it turns out many ideas in research were already tested in the past (of course DeepMind is most likely aware of and has improved on prior work). From NIPS in 2008: "Managing Power Consumption and Performance of C
690.
by argonaut 10y ago
Technically, it's 0, because Tesla does not use LIDAR at all.

More ›