Go to Top

Mobile Eats the Cloud

Andreessen Horowitz VC Chris Dixon recently wrote about the Decline of the Mobile Web and that this “favors the status quo over new innovations”. This view is supported by Union Square Ventures VC Fred Wilson in The Mobile Downturn where he also states: “The downturn is always followed by something radical and new that starts on the fringe and becomes mainstream.”, let me get back to just that.

I strongly disagree with Chris’ claim that this favors the status quo over new innovations.

The return of the Fat Client, but this time Mobile

The Web gave birth to the first widespread thin clientthe Browser (I’m sorry VT100), and Cloud services has been a natural evolution of the Web, but both used from a browser and in mobile apps. Many of the things we used the first widespread thick clientthe PC – to do in the past have been replaced by Cloud Services (e.g. this is written using WordPress cloud service).

On mobiles, tablets and wearables it has been similar application of cloud services (REST APIs) as on the desktop but from apps, this has dramatically reduced time spent on the mobile web as Chris points out (from 20% of time spent to 14%, i.e. a 30% reduction, in just one year). HTML5 has been mentioned as the technology that will bring back the mobile web, but trend towards “natification” of mobile apps doesn’t seem to support that.

The Rise of Bring Your Own Supercomputer Mobile Device (BYOSMD)

As Marc Andreessen – the inventor of the web browser – points out in this tweet is that his mobile phone is faster than supercomputers of the past, and this is just the beginning of mobile supercomputing.

1. Mobile CPUs and GPUs

Marc’s phone has a very fast 64 bit Apple A7 Cyclone CPU (see also this), and the corresponding tablet – iPad Air – has a set of GPUs that Bartosz Ciechanowski (@BCiechanowski) recently demonstrated in Exploring GPGPU on iOS can be up to 64 times faster than A7 using OpenCL, or as Bartosz says: “That’s right, the GPU was over 64 times faster. Sixty four. Times. Faster.” (link to his code on github). And of course the competition doesn’t stand still, e.g.
Qualcomm recently announced mobile 64 bit CPUs Snapdragon 808 and 810, Intel recently announced the 22 nanometer 64 bit Atom CPU, Samsung is working on 64-bit CPU and Nvidia announced the Nvidia Tegra K1 with 192(!) cores. I recently got to see a sneak peak of what an underdog in this mobile super-CPUs has created: Adapteva‘s Parallella with the Epiphany Multicore Accelerator (with 16 or 64 cores).

Dear Chris Dixon: The effect of Mobile super-CPUs on Innovation is largely unexploited territory (perhaps with the exception of Infinity Blade 3..)

2. Mobile On-Device Storage

Mobile on-device storage is currently strongly neglected compared to cloud storage (ref: forthcoming Box IPO and large new investment rounds for Dropbox), but this is perhaps the most exciting and enabling field for innovation, e.g. Sandisk’s MicroSD has 1000x (thousand times!) more storage capacity in 2014 than just 9 years ago (2005), Lexar’s latest SD card can store 0.5 TeraByte, Kingston’s USB stick can store 1 TeraByte, and fortcoming RRAM from Crossbar can store 1 TeraByte on a 2 square centimeter chip. Note: Latency for accessing local storage on a mobile device can be roughly 1000 (thousand times!) less than accessing 4G Network. To put this into perspective: 1 TeraByte is more storage capacity than Google and Altavista Web Search Engines had combined in the late nineties.

And if one really would like to extrapolate one can also store (at least) 700 TeraBytes in 1 gram of DNA.

Dear Chris Dixon: The effect of Mobile Storage on Innovation is largely unexploited territory (but I must admit: Memkite is working on it)


What will this mean for the Cloud?

Given the tremendous increase in processing power and low-latency storage capacity of mobile devices, a lot of processing is likely to happen on device instead of in the cloud (since it can improve latency, privacy and by always being available), but the amount of data prepared for and moved from the cloud to mobile devices is likely to increase by a large margin hence increasing the need for the cloud, in particular cloud infrastructure. But cloud services that requires a REST call even for the tiniest interaction with the user I am not so sure.

What will this mean for Mobile?

Who knows, but it will definitely be a catalysator for innovation. Having a mobile supercomputer gives some exciting opportunities in search and machine learning, in particular if one assumes that it can be connected to the brain in 10-20 years time with EEG-based technologies or as an implant.

It feels right to round off with Pete Warden’s (CTO of Jetpac) really cool demonstration of How to add a brain to your smart phone by using Deep Learning on a Mobile Device to make the Phone into a Sonar-like device to detect .. drumroll .. a cat.

Dear Chris Dixon: Mobile Eats the Cloud

Feel free to check out our previous blog post – Technical Feasibility of Building Hitchiker’s Guide to the Galaxy, i.e. Offline Web Search – Part I

Best regards,
Amund Tveit (@atveit), co-Founder of Memkite

About Amund Tveit (@atveit - amund@memkite.com)

Amund Tveit works in Memkite on developing large-scale Deep Learning and Search (Convolutional Neural Network) with Swift and Metal for iOS (see deeplearning.education for a Memkite app video demo). He also maintains the deeplearning.university bibliography (github.com/memkite/DeepLearningBibliography)

Amund previously co-founded Atbrox , a cloud computing/big data service company (partner with Amazon Web Services), also doing some “sweat equity” startup investments in US and Nordic startups. His presentations about Hadoop/Mapreduce Algorithms and Search were among top 3% of all SlideShare presentations in 2013 and his blog posts has been frequently quoted by Big Data Industry Leaders and featured on front pages of YCombinator News and Reddit Programming

He previously worked for Google, where he was tech.lead for Google News for iPhone (mentioned as “Google News Now Looks Beautiful On Your iPhone” on Mashable.com), lead a team measuring and improving Google Services in the Scandinavian Countries (Maps and Search) and worked as a software engineer on infrastructure projects. Other work experience include telecom (IBM Canada) and insurance/finance (Storebrand).

Amund has a PhD in Computer Science. His publications has been cited more than 500 times. He also holds 4 US patents in the areas of search and advertisement technology, and a pending US patent in the area of brain-controlled search with consumer-level EEG devices.

Amund enjoys coding, in particular Python, C++ and Swift (iOS)

5 Responses to "Mobile Eats the Cloud"

  • Jinghai Rao
    April 23, 2014 - 6:38 pm Reply

    Hi Amund,

    I totally agree with you. Instead of big internet companies (like google) who promote to put all computation and storage to the cloud side, I believe the future will be a hybrid model combining mobile devices, personal storage devices (NAS, like Western Digital MyBook Live) and cloud computing. How to fragment the data and computation to make best user experience while preserving privacy would be a very interesting question. I don’t have any mature idea yet, but we can discuss if you are interested too.



  • Peter Newman
    April 24, 2014 - 3:15 pm Reply

    Mobile Eats the Cloud is spot on. Cloud is much too tenuous to keep up with every camera and microphone on all the time. We have to push the data reduction closer to the data source – fatter clients indeed.

    We hope you are also spot on about mobile machine learning. Our company, Artificial Learning, is designing the next generation of hardware that can free mobile machine learning from the cloud and its data warehouses.

    Best regards,
    Peter Newman – Director – Artificial Learning Ltd

    • memkite
      April 24, 2014 - 6:24 pm Reply

      Peter, what kind of machine learning algorithms do you support? Do you have any hardware available for testing?

      Best regards,

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>