Furthering its push into conversational AI (artificial intelligence), Microsoft announced this morning that it has purchased Semantic Machines, an AI startup, for an undisclosed sum. But this purchase may have a chilling effect on Cortana, Microsoft’s little-used personal digital assistant.
“Semantic Machines has developed a revolutionary new approach to building conversational AI,” Microsoft corporate vice president David Ku explains in the announcement. “Their work uses the power of machine learning to enable users to discover, access and interact with information and services in a much more natural way, and with significantly less effort.”
Curiously, a number of publications are claiming that Microsoft will use this acquisition to improve Cortana, its digital personal assistant. But the Microsoft announcement doesn’t say that: Instead, it suggests that the Semantic Machines technology is superior to its own conversational AI work with Cortana. And it says that the firm will now combine its in-house efforts with Semantic Machines’ technology to “deliver powerful, natural and more productive user experiences that will take conversational computing to a new level.”
Sign up for our new free newsletter to get three time-saving tips each Friday — and get free copies of Paul Thurrott's Windows 11 and Windows 10 Field Guides (normally $9.99) as a special welcome gift!
"*" indicates required fields
This is notable because Microsoft has recently demoted Cortana from being a standalone assistant—which might be found in phones, PCs, smart speakers, cars, and elsewhere—to being an “assistance” technology that is built-in to other products and can integrate with popular assistants.
Cortana is not going away, according to Cortana head J Soltero, who was recently interviewed by Mary Jo Foley. But it will likely merge more into Microsoft’s other AI/machine learning efforts and be promoted as a set of capabilities and skills, and be less forward-facing as a brand.
To that end, Microsoft says that it will establish a conversational AI center of excellence in Berkeley, California to “push forward the boundaries of what is possible in language interfaces.”
Notably, that’s where Semantic Machines is based.
Bats
<p>Why am I not surprised? By following this thurrott.com along with MaryJo and small number of Microsoft beat reporters, I pretty much know and feel the pulse of Microsoft.</p><p>This news seems to me, as a direct reaction to Google's Duplex offering. Google Duplex is awesome and takes AI technology to the next level. It's so mind blowing, that I am not surprised of there being an extremely high level of suspicion with it.</p><p>Perfected, this AI technology can possibly eliminate or reduce (partially) the personal assistant job market and practically give every single person in this world their own "Alfred", "Jarvis", "Friday"…."Majel Barret."</p><p>Microsoft, of course, is late to this. I always joke that Microsoft is so big that Redmond may be the home hq of the company, but their hq for R&D is in Mountain View, California.</p><p>Here is a few things, I think we can all expect. One, in the next Microsoft event, they will talk about conversational AI (duplex whatever) and make it sound like they have been in on this for sometime. Nadella will lecture on the it's importance and make connection between this amazing technology and Microsoft's and it's part in the company's overall AI strategy. While Microsoft is making it seem like they have been in on this, let's not forget the date of May 21, 2018. The date where Paul Thurrot reports Microsoft's intent to buy the an AI Startup. Second, after Google releases their Google Duplex product, Microsoft will release theirs a year or two later.</p><p>Ya know, this is exactly how it's going to happen.</p>
RR
<blockquote><a href="#277806"><em>In reply to Chris_Kez:</em></a></blockquote><p>I have half a mind that it could not have happened in 2 weeks, but. I am not so sure, strategic M&A can move very fast when the stars align. Also, these are industry insiders. Things that happen in public they probably have some rough sense it could happen sometime before that, and you measure the timeline from that not when the public sees it. </p><p>Another way you can look at this is not from Microsoft's perspective, but that of the aquiree, Semantic. It could be that the Google I/O demo basically crapped their value. Like, this is the same thing they wanted to sell to these companies, and look, the companies are already doing it themselves, and they are therefore making a sensible countermove for their longer term survival by seeking an ark rather than waiting until they are downing before acting.</p><p>Related is Microsoft has shown a propensity for bottom fishing over the last couple years for Bay Area companies. They did it with LinkedIn, with that 3D/social media startup that actually closed and they opened back up, and probably other examples that we would never know why the target sold. So, if Google I/O basically sent these guys reeling, well, "Hello, Microsoft". It could just as well be that rather than Microsoft desperately needing this. I think Microsoft knows they must expand their Silicon Valley presence & linkages. Apple & Google have used local champion status to play against Microsoft to devastating effect in the developer & public opinon wars, so with that larger context there may be a lot more sense to it that just playing catchup. </p>
RR
<p>Microsoft London demo reported yesterday that shows Xiaoce has been doing this duplex conversation thing that Google demoed, but with real customers for the past 6 months shows that they are ahead, not behind Google in this tech. Therefore the reflex assumption, expressed in some comments, that this acquisition was a desperate catch up move on their part is wrong, I did not know this info when I commented below though.</p><p>But it is instructive. There are two sides to any deal, it is not just one side that can have motivations. For example, it is very likely the Semantics Machine folks knew what Microsoft had achieved in China, even though Microsoft didn't make a media splash, it was still a public trial, in their direct line of work. The deep tech insiders may be signaling something here by which team they decide to join, when it came time to sell. Doesn't mean Microsoft won't squander the tech advantage though, they do that regularly. </p><p><br></p>