First things first

A startup called Diffbot today launched a knowledge graph that it claims includes a trillion facts about people, places and things. It also claims the graph is powered by artificial intelligence, which sounds like a critical aspect to its scale and accuracy, but is almost secondary from a business perspective.

As I've written before -- mostly (recently, at least) about Apple's plans around AI and its hiring of former Google/Metabase exec John Giannanddrea -- knowledge, however derived, is going to be the major point of distinction for smart-home platforms. Hardware, partnerships and ecosystems could all all look pretty similar in the end, which means brains, for once, will probably trump beauty. Google didn't have the first voice assistant or the first smart-home device, but everyone is playing catch-up on actual utility because of its huge headstart in search and knowledge graphs.

In that world, a company like Diffbot can play an important role by providing those brains via its API. Diffbot already claims a list of web and search-engine customers you've probably heard of (including Microsoft, eBay, Yandex and DuckDuckGo), but it's hard to overlook the excitement around speech recognition, augmented reality and smart devices (even if, like me, you'd like to). Last year, I interviewed the founders of another company, Ozlo, building a knowledge graph for voice assistants. Facebook acquired it a few months later.

I included Apple in the subject line not just because of the Giannandrea angle, but also because news broke today that it has acquired an AR hardware startup called Akonia Holographics. Like most people not involved in that deal, I have no actual idea what Apple plans to do with the AR lens IP it now owns. But I feel pretty confident saying that if smart glasses really do become a mainstream thing beyond Apple Watch on your face, they'll need to connect with some real-world knowledge on the backend.

I suspect we'll start hearing more about Diffbot, and perhaps the handful of other knowledge graph companies out there, as we begin to expect more from our "smart" devices and apps. And although it's not as sexy as AI a few years ago, it wouldn't surprise me to see a knowledge graph acquisition spree happen in the next year or so. It also wouldn't surprise me to see the major cloud providers get into this space to complement their existing suites of AI services and APIs.

One last thing: Please fill out this survey if you haven't yet:

And one actual last thing: This week's podcast with Hilary Mason was super fun and, I think, super insightful. Take a listen: Hilary Mason on data ethics and figuring out what's real in AI.

Read and share this issue online here.

AI and machine learning

Cloud and infrastructure

Data and analytics