ArchiTECHt Daily: Why we play games with AI

Computers that can play games better than humans are nothing new, and yet they still get a lot of att
ARCHITECHT
ArchiTECHt Daily: Why we play games with AI
By ARCHITECHT • Issue #23
Computers that can play games better than humans are nothing new, and yet they still get a lot of attention. After all, they make for fun stories, challenge researchers and elicit some existential questions about just how smart we really are. However, for the purposes of applied artificial intelligence, I think the discussion about these systems often doesn’t go far enough.
I—and I assume a lot of CIO types—would like to know how reinforcement learning, for example, applies beyond defeating human champions at the game of Go. Particularly, it would be interesting to know how these types of algorithms might apply to some tough business problems. If you think hard enough you’ll conjure up some ideas but, frankly, most people don’t have the time to dwell on the subject, much less the expertise to start building anything.
The risk here is that despite some apparently revolutionary technologies, their actual applications beyond playing games or improving things at companies like Google and Facebook could be quite limited. We could find ourselves stuck trying to solve many of the same enterprise problems as we were 10 years ago—fraud detection, sales optimization, recommendations, etc.—simply due to lack of imagination. Don’t get me wrong: Those are big and important problems and we should use AI to solve them if we can, but there have to be some optimal use cases for AI should that are different than the optimal use cases for the first iteration of big data.
It’s a tricky proposition because the toughest business situations are rarely as black and white as games, with their clearly defined rules and often obvious reward/penalty outcomes. And they probably can’t be solved with computer vision or voice recognition alone. But if smart entrepreneurs are willing to engage with smart business execs to identify and solve some of their toughest—and perhaps untapped—challenges, I’m confident we can see some real progress.
End of rant. Here’s what got me thinking about this:
P.S. Sorry the newsletter was so late today.

What's new on ArchiTECHt
Peter Yared. Source: Sapho
Peter Yared. Source: Sapho
The latest episode of the ArchiTECHt Show podcast features Peter Yared—co-founder and CTO of Sapho, and former CTO/CIO of CBS Interactive—discussing, among other things, how containers and Kubernetes could help ISVs get back to selling software rather than SaaS.  Continue reading on ArchiTECHt »
Around the web: Artificial intelligence
Who doesn’t like a listicle!? This one is pretty good, too, focusing on methods for standardizing deep learning and integrating it with existing technologies. The latter will be critical for general-purpose software.
I think I’ve pointed out before how well suited Pinterest is to adopt and, eventually, push the envelope on computer vision. This is a good profile of what they’re up to.
This is one of those potentially very useful and chronically unsexy use cases for AI. Someone should figure out how to monetize it soon.
This is related to the link above about MDM. Not only is there is a lot of grunt work that AI could automate, but there’s also a lot of grunt work associated with AI that could stand to be automated.
This is what I was talking about up top. There are so many other areas to focus on beyond assistants and search.
Reportedly, it was only correctly responding to 30 percent of requests without relying on humans. That being said, part of the problem seems to stem from a bigger issue than M: helping users understand from the start what AI can and cannot do.
A decent read on Baidu’s efforts around scaling the hardware it uses to train AI models, including GPUs and supercomputers. It recently open sourced this GPU-training library, by the way.
What’s good for real universities is good for MOOCs. Hopefully, institutions will be able to help folks who really want to finish, well, finish.
arxiv.org  •  Share
Source: Dropbox
Source: Dropbox
Around the web: Cloud and infrastructure
Probably a good move for Rackspace given its recent layoffs and desire to focus on making money. Dev tools aren’t known for being too lucrative, and playing up the stack isn’t in every company’s DNA.
The latest release of OpenStack is out, and containers projects are the fastest area of growth. That’s an understandable, but possibly short-lived situation as users try to cut out the middleman and go straight container platform.
Would you pay for a specialized appliance for running containerized workloads? As with all appliances, there’s an easily identifiable logic, but they have been a tough sell in the cloud era.
And for sake of database users everywhere, we should all hope CockroachDB and Cockroach Labs are wondrously successful. Open source will be a major check on cloud lock-in.
This could be a trickier problem than people give it credit for, especially for new apps that should take advantage of new architectures. Nice analysis of where we’re at today.
And if it’s a big deal to developers’ careers, it’s going to become an even bigger deal for companies. Fewer people want to work in a black box environment.
medium.com  •  Share
ChatOps are coming for you … to protect you.
Want to sponsor this newsletter, or the ArchiTECHt Show podcast? Email advertise@architecht.io for more information.
Around the web: All things data
Another startup doing next-gen BI, targeting data stored in Hadoop, MongoDB, etc. I would have thought we’ve reached capacity in this space, but perhaps it’s not solved yet.
So if you’re a scientist wanting to see what Google’s cloud can do, now’s the time to apply for an NSF grant.
blog.google  •  Share
The more data, the better when it comes to mapping. In rural areas, especially, good info can be hard to come by (if the map knows what road you’re to begin with).
If you needed more evidence that Palantir really is involved in mass surveillance, here you are. I would be more concerned about what’s coming next.
Chris was an early employee at Aster Data back in the day, and co-founded BI startup DateHero a few years ago. Founders couldn’t ask for a nicer mentor.
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.