ARCHITECHT Daily: AWS is not done growing, by a long shot

Yesterday, GeekWire highlighted a chart produced by CB Insights about all the hiring Amazon is doing
ARCHITECHT Daily: AWS is not done growing, by a long shot
By ARCHITECHT • Issue #62
Yesterday, GeekWire highlighted a chart produced by CB Insights about all the hiring Amazon is doing right now. Remarkably, nearly a third of all open listings—more than 5,600—were for Amazon Web Services. (When I looked on the Amazon site this morning, the number was a mere 5,353.) That’s a far larger number than for any other Amazon division.
A couple of things:
  1. That’s a big percentage for a unit that contributed less than 10 percent to the company’s total revenue in 2016. Google, by comparison, appears to have less than 200 cloud jobs listed on its careers page. Microsoft only lists the top 500 results on its careers page.
  2. This makes the AI-jobs report I linked to last week seem like it’s probably accurate in terms of what Amazon is hiring. In fact, a quick search for “artificial intelligence” on the Amazon site now turns up 173 positions.
Assuming actually intends on filling all those 5,000-plus AWS positions, the writing on the wall is that it expects cloud to be a huge-growth business (which is what everyone else expects, too) and that wants to give rivals Microsoft and Google as little breathing room as possible. It won’t be able to snuff them out completely—as I explained in March, there are good reasons to like the chances of both Microsoft Azure and Google Cloud going forward—but blanketing the world with sales, support and new services could be an effective mitigation strategy.
Even if it didn’t have competition, it’s also worth remembering that AWS was responsible for the majority of Amazon’s total operating income in 2016. You can understand why Amazon might want to throw fuel on the fire to keep it growing as the company continues fighting lower-margin battles elsewhere to expand its footprint.

Sponsor: Cloudera
Sponsor: Cloudera
Artificial intelligence
I’ve been saying this for a while now, just less eloquently. I hope Nvidia is paying attention and thinking about the future.
As I wast just saying, everyone is getting in on these custom AI architectures. Fujitsu joins Google, IBM, Intel, at least a few startups, and probably Amazon and Microsoft, as well.
The Alibaba boss thinks fewer jobs and longer lifespans cloud be a recipe for disaster if not managed correctly.
Even in business, creativity matters. Humans figure out all sorts of ways to solve problems, often not involving software at all.  •  Share
Let’s not forget the recent U.S. election as we make predictions about the effects of robots and AI. If people don’t trust them, there will be pushback.
It’s table stakes in the cloud now. 
These researchers from Singapore say existing distributed schedulers can’t handle the requirements of TensorFlow, MXNet, etc, running at the same time. They suggest a new system called Dorm.
Cloud and infrastructure
But the majority of work still happens in its own facilities. This is a good breakdown of Apple’s energy and data center footprint.
At least $33 million all-in if the proposed pay package is approved. It seems like a lot (too much?) given people’s, shall we say, mixed reactions to IBM’s recent performance. Also, I would like to be IBM CEO for a week.
Kudos to Google for acknowledging that customers might want to use, or already be using, AWS. This example connects S3 and Lambda to to Google Cloud Endpoints.
Less than $3 million for a team of engineers could be a good investment. On the other hand, I’m not certain how it works to acqui-hire people from a company that still exist.
This is a fine writeup of how software development has evolved in the past decade-plus. From the CEO of Wercker, which is soon to be part of Oracle.
All things data
A group consisting of members from Google, Amazon and more is pushing for more access to public datasets. They’re concerned with AI primarily, but open data is important for many more reasons.
If rumors of a major new funding round are accurate. This seems like an awful lot of money in a space that must be pretty saturated. But I’m also woefully ignorant of this space.
Security for real-time data processing is an important topic, but one has to wonder how broadly Metron or any Hadoop projects will be adopted beyond their main sponsors. In this case, Hortonworks.
A CS student implores her peers to take a broader worldview. This isn’t just a data problem, but in a way it’s mainly a data problem. At least if you think about data vs. information, or quantitative vs. qualitative.
Did you enjoy this issue?
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
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.