ARCHITECHT Daily: Applied artificial intelligence FTW

We all love groundbreaking (or at least really interesting) AI research, but it can be tough to draw
ARCHITECHT
ARCHITECHT Daily: Applied artificial intelligence FTW
By ARCHITECHT • Issue #96
We all love groundbreaking (or at least really interesting) AI research, but it can be tough to draw a clear line between a computer model that’s great at playing games and a computer model that can transform your business—as so many experts are predicting will happen. There’s even a growing sentiment, especially among certain investors, that the general-purpose AI-as-a-service options being pushed by cloud providers won’t end up moving the needle much. Basically, the argument goes, enterprise AI adoption is going to happen via bespoke systems and industry-specific applications.
If that is indeed the case, then enterprises got some good news on Wednesday, with the announcement by Element AI that it has raised about $102 million in venture capital. It’s a lot of money for such a young company (Element was only founded in October), but then again, not every company can deep learning master Yoshua Bengio as a co-founder. The last AI company founded by one of the deep learning grandmasters was Geoff Hinton’s DNN Research, which Google acquired in 2013.
In today’s IT climate, it’s hard to find a better description of what your company does than this one from Element: “We’re turning the world’s leading AI research into transformative business applications.” That’s a powerful promise, but the company appears to have a sound strategy—not to mention to brainpower—to pull it off. Doing custom work for every customer is expensive, but having $102 million in the bank will help with that.
Here are a couple more interesting applied AI items that I came across today:
For more insights into enterprise AI adoption, I would also suggest checking the following episodes of my ARCHITECHT Show podcast:

Sponsor: Cloudera
Sponsor: Cloudera
Artificial intelligence
Seems a bit optimistic but, then again, I don’t sell CRM software. Seriously, though, I’m sure it will be helpful, but I also don’t think CRM is the ideal app for these technologies.
This is one of the best pieces I’ve read in a while about the progress being made on quantum computing, and why people are so excited about them. In other quantum news, here’s a partnership between Accenture, 1QBit and Biogen on drug discovery.
I linked to this paper earlier. Here’s an explanation of how it works and why it matters.
Reinforcement learning strikes again! Maluuba is apparently Microsoft’s answer to Google DeepMind. And Microsoft is thinking about commercial application, including in sales.
Facebook has been doing some cool research on conversational AI, and this is no exception. I would love to see this somehow taken beyond chat, though, and perhaps into dynamic pricing or other areas where negotiation comes into play.
When people suggest that Apple is behind on AI, this is part of the reason why. Google is putting in the work to put AI on mobile devices and to get its stuff broadly adopted.
This type of work is obviously important for SnapChat-type apps, but don’t mistake this for a competition with Google, which has bigger fish to fry by putting deep learning on mobile devices (i.e., Android world domination).
arxiv.org  •  Share
Sponsor: DigitalOcean
Sponsor: DigitalOcean
Cloud and infrastructure
Don’t expect Alibaba to challenge AWS in the United States or, probably, Europe (although data protection laws there could have unforeseen effects), but there’s no reason I can see why it can’t rule across Asia. Also, Alibaba is using liquid immersion cooling for servers, and plans to share its technique with the Open Compute Project.
MemSQL is by all accounts good technology, although there’s a lot of competition in the cloud data warehouse market, including from another startup, Snowflake. Still, if customers want cloud options and will pay, it’s probably best to deliver.
I get the carbon-neutral thing, but it’s pretty surprising there isn’t yet a government region near … the government.
gcn.com  •  Share
Exactly what it sounds like. I’m sure it’s very useful for a lot of situations, including, as AWS suggests, serverless apps, but it’s important to keep track of what’s really happening with your managed databases.
Like site reliability engineering, which Google made popular a couple years ago. The good news is that this probably means your database future will be brighter than your database past.
I just thought this was interesting, even though it’s not particularly newsworthy. There are some cloud, storage and AI companies among the investments.
This team includes researchers from Expedia and IBM, who have devised a method for letting mobile users—and services—hop from edge node to edge node as devices move.
arxiv.org  •  Share
All things data
I prefer the old eye test, or perhaps the sniff test, but actually capturing environmental data could work, too.
I’ve noted before that Slack can possibly succeed with this idea better than most previous attempts, because it has high-quality data in a fairly confined universe. It also helps that everyone’s already actively using it.
I have tried a lot of GUI-based data analysis/visualization tools over the years, and none really delivered good interactive charts that a non-coding journalist could use. Maybe Microsoft is onto something with PowerBI.
This is some good perspective on the spate of big data cloud services coming online lately. He’s focused on Cloudera Altus, but it might well apply to what Hortonworks, Databricks and Qubole are doing, as well. Also, I don’t think we’ve seen the end of where these companies hope to take their cloud offerings.
Listen the the ARCHITECHT Show podcast. New episodes every Thursday!
Listen the the ARCHITECHT Show podcast. New episodes every Thursday!
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.