3 things you should read today: MongoDB's IPO; Gartner on AI hype; why robots don't want our jobs

TechCrunch reported late Tuesday that MongoDB has confidentially filed for its IPO, meaning we should
ARCHITECHT
3 things you should read today: MongoDB's IPO; Gartner on AI hype; why robots don't want our jobs
By ARCHITECHT • Issue #123
TechCrunch reported late Tuesday that MongoDB has confidentially filed for its IPO, meaning we should see the company’s financial details relatively soon. I’m anxious to see what they reveal, in large part because MongoDB is a different type of company than companies like Cloudera and Hortonworks. Rather than pushing heavy infrastructure products and services built around a new technology, MongoDB built a database that’s relatively easy to adopt and, indeed, has lots of users as a result.
The last estimate I saw was that the company is doing more than $100 million in annual revenue, although how many paying customers it has or how much cash it’s losing I do not know. I’m especially curious to see the ratio of users and paying customers—something about which MongoDB has been dogged for years. Assuming it’s pretty high, that could viewed as a weakness of its business model (and open source generally) or as a potential goldmine of enterprise sales to come.
On the latter point, the company has expanded its lineup of revenue streams pretty heavily over the past year with new products. That’s also a fairly stark departure from other public open source companies, which have tended to focus on infrastructure. 
At any rate, we’ll all find out what’s doing when MongoDB’s S-1 is public. Or, if someone swoops in and buys it beforehand, then not. In the meantime, you can check out my interview with co-founder and CTO Eliot Horowitz from 2016: MongoDB co-creator explains why ‘NoSQL’ came to be, and why open source mastery is an elusive goal. And also my podcast interview with Honeycomb co-founder and CEO Charity Majors about the unfair rap MongoDB sometimes gets.
Two other things I want to highlight today are:
1. Gartner released its new Hype Cycle for Emerging Technologies, which actually includes quite a few artificial intelligence techniques and technologies—more than I would have guessed. I’ll admit that I have a tough time distinguishing between “deep learning,” “machine learning” and “cognitive computing” on the Peak of Inflated Expectations (as well as some of the more specific technologies elsewhere), but if we’re going to break them up I’d actually argue that machine learning is probably mature beyond the rest and well up the Slope of Enlightenment. (God, I love those labels.)
2. WIRED has a good story about why the fears of job loss due to automation might be overblown. You should just read it, but the gist is that there’s really no evidence of large-scale replacement of workers by robots and, in fact, some signs point to the opposite conclusion. If you’re searching for a reason to be optimistic on this issue, this will help—although, I’d note, talking about robots is not the same as talking about automation via software, which is kind of its own discussion.
And here are couple other, complementary pieces on this issue from today:

Sponsor: DigitalOcean
Sponsor: DigitalOcean
Artificial intelligence
Just what it sounds like: more details on how OpenAI built its bot to tackle yet another complex game. Some of the original criticism was a fair response to inflated hype about the accomplishment, but it’s still an impressive accomplishment.
The headline pretty much says it all. If Ford is successful, that billion dollars will seem like chump change (kind of makes Snap’s valuation seem even crazier). But this is a very competitive field, and I’m not sure anybody’s certain what will ultimately be the difference between table stakes and huge advantage.
The details are interesting, but the broader context is all at the end. Basically, the cutting-edge AI video-game research coming out of places like DeepMind and OpenAI is about dealing with unknown, complex situations—something cars, drones and anything in the real world will also have to do.
I think this is pretty well known when it comes to text, because the use of slang, dialects, mispellings, etc, means you need to cast a broad net to catch everyone commenting on an issue. But I hadn’t thought about this with voice products, which will remind users they’re different really quickly if devices can’t understand what’s being said.
Well, not quite everything … but this is still a good intro into why they can be troublesome (they also can be beneficial). It’s definitely an area to think about re: security and safety (e.g., with driverless cars), but probably not an immediate (or easy to pull off) threat in most instances.
This is only research at the moment, the authors are from Nvidia, MIT, UC-Berkeley and Stanford—which is a who’s who of AI institutions. And considering it’s about running models more efficiently on hardware, Nvidia’s presence is notable.
arxiv.org  •  Share
Sponsor: Bonsai
Sponsor: Bonsai
Cloud and infrastructure
I think the most interesting thing here—which is something that applies not just to Kubernetes, but to the entire container ecosystem—is that the business decisions are often more about flexibility, automation and standardization than about infrastructural efficiency or scale.  SIDE NOTE: Jesse Newland, the author of this post, will be on the podcast next week talking about it and site reliability engineering, in general.
As with most things cloud, AWS got the drop on the competition by being first to market with Lambda and making a big splash with it. But Microsoft seems determined to differentiate its serverless play via UX.
Linux, Windows and … IBM System z (yes, the mainframe OS). I could be wrong about this, but the real benefit here seems to be around consolidating workloads. That said, standardizing deployment across new and legacy apps has its benefits, as well.
Consider the source—this is written by someone who sells embedded FPGAs—but there’s every reason to believe we’re yet done seeing the shift to alternative hardware platforms. The big X factor, of course, is what the cloud-scale buyers will buy into.
If you enjoy the newsletter, please help spread the word via Twitter, or however else you see fit.
If you’re interested in sponsoring the newsletter and/or the ARCHITECHT Show podcast, please drop me a line.
Use Feedly? Get ARCHITECHT via RSS here:
All things data
I wouldn’t normally consider this a data story, but Pattern seems to have focused its business collecting user data to build relationship graphs. Last year, Workday acquired Platfora, which made a splash several years ago by selling analytics software built around a foundation of Hadoop and then Spark.
This is really a look at what to expect from AWS’s forthcoming Postgres-on-Aurora general availability, but it also provides some good general insights into the intricacies of selling cloud database services and convincing customers to switch database systems and/or vendors.
This is a good primer on how to think about data as part of the online personalization experience. However, I think the bigger picture is how to determine what actually needs to be personalized and how much benefit it will actually bring to users or the company.
New ARCHITECHT Show every Thursday; new AI & Robot Show every Friday!
New ARCHITECHT Show every Thursday; new AI & Robot Show every Friday!
Did you enjoy this issue?
ARCHITECHT
ARCHITECHT delivers the most interesting news and information about the business impacts of cloud computing, artificial intelligence, and other trends reshaping enterprise IT. 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.