I don't think that /s is necessary. I think the years have proved that china's trade barrier has worked wonders for home grown businesses. Look at didi chuxing. Blindly aped uber and worth more than them now. On the other hand look at India's own uber competitor: ola. They're struggling almost as much as flipkart while uber thunders ahead.
But as an Indian consumer, I have happy that I have a choice and I believe in the long run the more open Indian market would prove more beneficial to local businesses.
Their market caps notwithstanding - Alibaba, Didi Kuaidi and Baidu are not global companies. India already has one of the largest global ad networks outside of Google i.e. InMobi and I believe over the next 20 years, you'll see much more truly global Internet companies started out of India vs. China.
Unfortunately most Chinese are not tech workers reaping in cash, they're consumers who have to make do with poorer quality local offerings like Baidu and Weibo.
Of course trade protectionism is good for the protected industries. That should be self-evident.
The more important question is whether the benefit to domestic industry offsets the harm to consumers. Traditional economics would say no and I'd be inclined to agree.
Amazon is years ahead in customer support and services. Yes. but in terms of products sold, I'd argue that they are mostly the same: made in China or southeast Asia.
Works for whom? Definitely not the consumer. I am happy my friends and family are enjoying the reliability and customer support of Amazon. I wished flipkart upped their game, but apparently they did not. Amazon has a well of experience, even though the market is new, they know what parameters to look for and best of all, can wait for a decade before turning profit.
The Chinese government backs several companies in the same industry, there's no reason to believe they don't have both company's backs
>2. Their decision to make their OS to show crazy amount of ads is really short sighted.
This must be a Chinese mainland thing, I've never seen this on any of their devices.
> 3. They are falling dangerously behind Huawei in both sales and quality of product.
I don't think so, I find the Xiaomi products to be very inticing and well built for the price. In the new era of smartphones where people will no longer be willing to shell out big bucks and subsidies of smart phones from providers are faltering (or just ending.) Companies like Xiaomi stand to gain.
My experience doing business with Huawei and being on site for a while was that they even pit different divisions against each other. The fear of competition or being beaten by another Huawei division was much more evident than any external company.
They hide ideas for other divisions and act very competitive with each other.
> I don't think so, I find the Xiaomi products to be very inticing and well built for the price.
Have you ever used actually used a Xiaomi product for some time? Sure they are cheap and look ok but they are so unreliable it gets unbearable after a few months.
I have: Xiaomi speakers, plugs, air purifier, water quality tester, water purifier, router, backpack, portable charger, fitness tracker, lightbulb, etc.
They all work very well, and are exceptional quality, considering their price.
That quantizes the gradient for compression of the communication between nodes, which is cool, but each node must still calculate a 32 bit floating point gradient locally. What GP is asking for is a way to avoid having any floating point math at all.
If you could implement training with only single-bit operations rather than floating point math, a hardware implementation could be several orders of magnitude faster and more efficient than current CPUs/GPUs. That would certainly usher in a revolution in computer architecture.
jsno2, jsno3, jsno4 are accounts created within an hour only to comment positively for Qihoo. Are there any rules on HN against this kind of behavior or can someone ban these fake accounts?
It's a hard problem to solve, and it might be a good crowd-sourcing opportunity. They need to expose some API for managing traffic, so that the community can experiment with new routing and scheduling strategies.
