ARCHITECHT Daily: Google is the epicenter of AI, but ...

Make no mistake: Google is currently the sun around which the artificial intelligence community orbit
ARCHITECHT
ARCHITECHT Daily: Google is the epicenter of AI, but ...
By ARCHITECHT • Issue #60
Make no mistake: Google is currently the sun around which the artificial intelligence community orbits. 
Between the teams at Google Brain and DeepMind, the company is behind many of the big advances in AI and many of the most popular research papers. Google is the company behind the popular TensorFlow framework, and ceaselessly announces new AI-powered products and features (some eye-opening, others incremental). If you keep up with AI news, the chances are you’re reading several stories about Google per day.
Want proof? Here’s yesterday’s Google AI news (although probably only a portion of what’s actually out there):
You will not find, on any given day, stories on 4 separate topics relating to any other company’s AI technology or personnel.
But … despite all of Google’s investment and innovation in AI, it’s possible Google won’t be the company to capitalize on AI most successfully. Amazon is quietly (and covertly?) making its own investments in AI. 
I cannot vouch for the accuracy of this report at all, but job site Paysa estimates that Amazon currently has AI job openings totaling more $227 million in salary, compared with just $130 million for Google. Amazon, according to this study, has more than twice the number of job openings than Google, but with a lower average salary.
These numbers could be completely off. Or they could be accurate, but only because Amazon is so far behind Google in terms of AI talent that it has to hire lots of people to catch up. In terms of research presence and public perception, Amazon certainly does have a long way to go. 
However, it also could be that Amazon is applying the same principles to AI as it has to other technological areas and its business, in general. It could be using AI to drive efficiencies internally and power some customer-facing products, without caring too much whether it’s viewed as research hub or a particularly open book.
Amazon CEO Jeff Bezos wrote quite a bit about AI in his recent shareholder letter, including commenting on efforts to democratize access to AI via everything from Alexa to AWS. (I previously wrote about the Amazon-Google-Microsoft battle for cloud computing and IoT dominance here, and about the importance (or not) of openness here.)
But this paragraph from the Bezos letter stuck out to me (emphasis mine):

“But much of what we do with machine learning happens beneath the surface. Machine learning drives our algorithms for demand forecasting, product search ranking, product and deals recommendations, merchandising placements, fraud detection, translations, and much more. Though less visible, much of the impact of machine learning will be of this type – quietly but meaningfully improving core operations.”
For large corporations like Amazon and Google, this is where the rubber meets the road. Google’s massive AI investment will likely pay off immensely in the decade to come. But something tells me Amazon will be happy to pad its bottom line without matching Google in terms of AI moonshots, GitHub stars and articles in Nature

Sponsor: Cloudera
Sponsor: Cloudera
The ARCHITECHT Show podcast
Segment CTO on trimming $1M off the AWS bill, and why he’s all in on cloud
Segment CTO Calvin French-Owen explains how his company saved $1 million a year by troubleshooting its AWS environment, and much more.
Backblaze CEO on the economics of cloud storage and the business value of blogging
Backblaze CEO Gleb Budman tackles a wide range of topics, ranging from the economics of cloud storage to the benefits of good blogging.
Artificial intelligence
If you haven’t already read it, carve out about 2 hours to do so.
How do you look for changes due to mining across 37 million acres worth of imagery? Machine learning and a Kaggle competition, of course.
An O'Reilly podcast featuring Reza Zadeh—adjunct Stanford professor, former Twitter, former Twitter data scientist, and current CV startup CEO.
Correct me if I’m wrong, but isn’t this essentially the idea behind unsupervised learning, and what companies like Ayasdi are up to?
I guess they never heard of Watson. Seriously, though, cool project and a good use case for Microsoft’s myriad products.
That’s question-answering, not the other, less fun, definition. What’s interesting is the knowledge-based angle AI2 is taking on its research compared with almost everyone else.
arxiv.org  •  Share
Cloud and infrastructure
Perhaps it’s fitting that Yahoo is one of the largest OpenStack users willing to speak publicly about it.
When the CEO is commenting at all, something is going to happen. Cloudyn wouldn’t give Microsoft a whole lot, but perhaps some IP for modeling customer spending—and insights into spending/usage on AWS.
fortune.com  •  Share
But in the end, the cost of all basic cloud tasks will approach zero.
I know that technical folks hate buzzwords and overly broad definitions, but we need a way to talk about containers, etc., that people can understand and use as shorthand. It’s more important than you might think.
An MIT professor ran an HPC job on Google’s cloud using 220,000 cores on Preemptible VMs (Google’s answer to AWS Spot Instances).
This is probably a good thing for Spanning, which focuses on backing up SaaS data.
fortune.com  •  Share
Sponsor: Marshal.io
Sponsor: Marshal.io
All things data
I always want to be excited about what Wolfram is doing, but I can’t get a sense how widely it’s actually used. I’d love some insights into this.
You have to think Kafka has the edge for the next few years, at least. We’ll see how cloud-based options—and users’ preferences on topics like open source—evolve over time.
Survey says: Data scientists love their jobs, but do not love data munging.
A boring, but very important, aspect of cloud computing: knowing for sure whether your cloud provider is deleting the data you promised users you’re deleting.
Did you enjoy this issue?
ARCHITECHT
The most interesting news, analysis, blog posts and research in cloud computing, artificial intelligence and software engineering. Delivered daily to your inbox. Curated by Derrick Harris. Check out the Architecht site at https://architecht.io
Carefully curated by ARCHITECHT with Revue. If you were forwarded this newsletter and you like it, you can subscribe here. If you don't want these updates anymore, please unsubscribe here.