Richard MacManus.
Microsoft chief executive Satya Nadella recently sent out a tweet announcing a partnership with Publicis Groupe, a multinational advertising and PR company.
Nadella said they would build an artificial intelligence product called Marcel, as some type of collaboration platform to ‘‘empower’’ the 80,000 employees of Publicis Groupe.
Nadella’s tweet didn’t offer much detail about the new AI platform, other than saying it would provide ‘‘intelligent insights’’. But Nadella did link to aYouTube video of him and Publicis Groupe chief executive Arthur Sadoun talking about the partnership.
Good, I thought, they’ll explain to me exactly how Marcel will work.
Of course that didn’t happen. The three-minute video was a series of platitudes between the two men, who each profusely thanked the other for the partnership and said how ‘‘excited’’ and ‘‘thrilled’’ they are. The cultural fit is tremendous, apparently.
The closest we got to specific product detail was Sadoun’s comment that Marcel will help Publicis employees to ‘‘learn more’’, to ‘‘share more’’, and to ‘‘create more’’.
So…is this an AI version of SharePoint? In case you don’t know it, SharePoint is Microsoft’s web-based collaboration software that integrates with Microsoft Office. I’d always found SharePoint to be overly complicated, so perhaps it does need an AI to take control.
Next I googled the official press release for Marcel. It was nearly as bad as the video in its lack of specifics. The only semi-useful information it contained was that Marcel will ‘‘leverage’’ Microsoft Azure AI and Office 365.
Unfortunately, this experience – trying to work out what exactly Microsoft is promoting – is becoming increasingly common in the software industry.
Too many companies and startups these days are vague about the technology in their products. Telling me it will help people learn more or share more is not telling me anything.
How specifically will Marcel help someone learn or share? What will the architecture look like?
You don’t have to give away any intellectual property. But some level of detail is required to convince me that your product is more than just talk.
I’ve decided to call these products ‘‘vagueware’’. I define vagueware as software that is only vaguely described and has no discernible features.
Think of it as similar to vapourware, which is when a software company makes big claims about an upcoming product then ends up never releasing it.
Startups are notorious for announcing vagueware.
Take Magic Leap, a well-funded American startup (US$1.9 billion) which has been promising a revolutionary ‘‘mixed reality’’ headset for the past several years.
In early 2016, Magic Leap released a video showing a virtual whale leaping out of a gymnasium floor, to the astonished reaction of a group of children.
Many people came across this video on their daily web surf and assumed it was an actual product demo. But it later emerged it was just a scripted marketing video.
Journalists from Wired and Rolling Stone have since seen Magic Leap’s mysterious software in action, so apparently there’s a real product behind the hype.
So I don’t think it’s vapourware. However it is certainly vagueware, because we don’t yet know how it will work in the real world.
Needless to say, there’s no shortage of vagueware in the cryptocurrency and blockchain market.
Last week I wrote about a local ICO (initial coin offering) which claimed it sold US$80 million of virtual currency. But when I looked closer, there was very little evidence that this company’s products use blockchain technology – which was the main selling point in the ICO.
The company’s nearly 50-page white paper is vague and avoids