ARCHITECHT Daily: Canada's quest to become an AI epicenter is the right one, but a daunting one

There was an interesting column published in the Montreal Gazette on Tuesday, calling for the provinc
ARCHITECHT Daily: Canada's quest to become an AI epicenter is the right one, but a daunting one
By ARCHITECHT • Issue #83
There was an interesting column published in the Montreal Gazette on Tuesday, calling for the province of Quebec to invest billions of dollars in an attempt to ensure Montreal and the surrounding cities become the AI hub they have the potential to be. Essentially, the argument goes, artificial intelligence has the potential to be a huge economic boon, and the time to strike is now before Silicon Valley vacuums up all the talent and the United States get a serious government in place again.
If you’ve been following the AI world, of course, the idea of pitching Canada as the epicenter (or at least an epicenter) of AI is nothing new. The Canadian government and private institutions have already committed to invest hundreds of millions of dollars supporting AI in the Toronto-Montreal corridor. The political situation in the U.S. certainly helps with timing, as explained in a recent New York Times story, but so does the AI infrastructure already in place.
The University of Toronto and University of Montreal, in particular, produce large numbers of AI experts, often under the tutelage of deep learning masters like Geoff Hinton and Yoshua Bengio. Google, Microsoft, Uber and others have offices and labs in the region. The newly announced Vector Institute should help foment more innovation, and accelerators like Creative Destruction Lab are providing high-quality entrepreneurial guidance, as well as investment capital.
I think it’s great that Canada recognizes this opportunity and is trying to capitalize on it, but officials would be wise to set realistic expectations and invest accordingly. Lots of cities around the U.S.—including Las Vegas, where I live—have tried to become “the next Silicon Valley,” but few have actually been very successful at it. Some have a healthy startup scene, satellite offices of major web companies and have even had some local success stories. But epicenters of technology on par with Silicon Valley? No.
Eastern Canada is well positioned in the sense that it has the university infrastructure in place to attract and churn out a steady stream of talent. Where it’s arguably lacking (and people who study this closer than me might have a much different opinion) is in venture capital infrastructure and major anchor businesses (e.g., Google, Facebook, Microsoft, Amazon, Cisco, VMware, etc.). Those types of networks are immensely valuable in terms of money, meetings and experience.
Silicon Valley and Seattle have all three, which is why there’s a seemingly endless of stream of AI, cloud and other-cutting edge tech activity happening around them. It’s certainly possible that Toronto and Montreal could be to the AI era what Silicon Valley and, increasingly, Seattle have been to silicon/web/cloud eras, but it’s going to take a concerted effort to make that happen. Money is a big part of it, for sure, but it needs to be money well spent.
Building the foundation to keep the region’s AI scene humming once government subsidies dry up a decade from now is probably more important than spawning a few dozen startups tomorrow. Otherwise, Canada risks spending a lot of money to become a feeder to Silicon Valley. That’s not the worst thing in the world, but it’s not a national game-changer, either.

Sponsor: Cloudera
Sponsor: Cloudera
Artificial intelligence
Someone thinks Nvidia’s machine learning business is here to stay—and then some!
Speaking of Nvidia, it looks like AMD wants a piece of that machine learning action, as well. Good luck! If I had to guess, I’d say this goes a lot like its battles with Intel 
We’ve covered this before, but a little reminder never hurts. We are literally years from perhaps having commercially available quantum computing.
This is important to show that the world’ most famous (right?) AI system isn’t just a one-trick pony. It’s worth noting, though, that real-world applications can have obstacles that tech alone won’t solve
This could end up being a pretty big deal, in the sense that reproducibility is often a point of contention in AI (and other spaces). OpenAI is doing the lord’s work.
This is one of those ideas that might make shareholders smile, but drives consumers up the wall. Just because we can optimize much more using AI, that doesn’t always mean we should optimize.
By analyzing users’ microbiomes for markers of future diseases. Seems like snake oil, or at least wishful thinking.  •  Share
This is pretty cool and potentially very powerful research. We could stitch together computer vision and computer listening technologies, or we could build systems that do both simultaneously.  •  Share
Sponsor: DigitalOcean
Sponsor: DigitalOcean
Cloud and infrastructure
So says an Israeli news outlet. This would be Microsoft’s second Israeli acquisition this year, following its reported purchase of Cloudant in April.
If you haven’t been following Amazon ECS, you’re probably not alone. Here’s a helpful guide the the last year worth of developments for the container service that doesn’t involve Kubernetes.
It’s former EarthLink startup Joe Eazor. CEO search settled, I’m wondering whether there’s anything interesting left with Rackspace, or if it’s settling into its role as MSP and cloud support expert.
This is some classic enterprise IT hair-splitting and unnecessary boasting. Saying you have a markedly different arrangement that IBM or AWS doesn’t make it so. It also doesn’t mean customers care.  •  Share
Facebook’s automation tools and mobile device lab are really something else. Here’s a good explainer on how developers interact with them.
This tool—from a university, not a company—could be useful if you’re considering the 4 databases it currently supports. Then again, testing them in your environments on your data is probably more accurate.  •  Share
I noted yesterday, in response to its claims that it ships 1 million IoT chips per day, that Qualcomm might be under-appreciated in the AI-chip space. If it’s able to develop a distributed storage architecture optimized for IoT and edge computing, it might have a strong presence there, too.  •  Share
There were also two new Kubernetes projects announced on Wednesday: ksonnet from Heptio; and Istio from Google, IBM and Lyft. Ksonnet focus on making it easier to configure and deploy Kubnernetes applications, while Istio provides a service mesh for managing various applications on a shared cluster. Given the companies behind them, I’ll assume both are important advances in the container space, but everyone could do a better job explaining how new projects relate to other projects, as well as to non-cloud-native counterparts. It’s becoming really tough to keep up with which projects do what—and that is not a good thing for folks in charge of marketing and selling this stuff.
Media partner: GeekWire
Media partner: GeekWire
All things data
Cloudera Altus is the company’s latest foray into the cloud, only this one is built from the group up as a cloud service. Its first capabilities are for data engineering, a la Elastic MapReduce or Qubole.
Many cities still need help making data machine readable and accessible, not to mention thinking of good ways to utilize it and share it. It’s gold mine for the companies that can pull it off.
Metabiota was an early company trying to do important things with big data, in its case monitoring the spread of disease around the world. Its new focus on selling disease risk models to insurers is probably more lucrative.
Analyzing environmental factors, hospital experience, doctors’ notes and other info not included in a medical record could be a big help not just veterans, but all sorts of people. I’m not sure this is the holy grail of that, but it’s a start.  •  Share
And I will throw in $5 for the model that inflates the price of my house by the most ;-)
If, like me, you’re noticing quite a few blog posts lately about how to manage Kafka or build Kafka-Spark pipelines, etc., it’s because Kafka has shed its original big-data suit and is a key piece of many IoT and microservices environments. 
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?
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.