Innovation

Why there will be no more AI startups in 2 years

Every enterprise wants to use AI to market their otherwise unexceptional products, but soon we'll hear a lot less about AI because it will table stakes for nearly every organization.

We're maybe at 15% of organizations getting into production with machine learning and AI, suggesting many years before these technologies transform the world. Even so, Andreessen Horowitz partner Frank Chen is probably correct to posit that we're just two years away from VCs no longer investing in AI startups.

Or, more accurately, we're two years from an end to the silly AI washing that currently inundates our industry. It can't come soon enough.

Washing with AI

The AI washing silly season is completely predictable. It's what we do with every significant new trend. For years we had to endure every vendor dressing itself up in open source clothes, most of them without bothering to release much source code. Then came cloudwashing. Now we're AI washing. It was just a matter of time.

What does AI washing look like? It's when every vendor magically transforms their tired, staid products into unicorns by waving a bit of AI marketing pixie dust over them. Basic analytics or personalization becomes much, much more when the AI wand is waved.

SEE: Microsoft used AI to combat global tech support scams (TechRepublic)

Suddenly, AI is everywhere, including in a pitch for a bedroom toy I received via email. Color me skeptical, but I somehow doubt that advanced AI is involved in that product, but that's par for the AI washing course.

The antidote to AI washing, just as happened with open source washing and cloud washing, is time.

Getting real with AI

Over time, things like open source become expected ingredients of successful software strategies, and not overt strategies themselves. This is the point that Chen is making:

[In two years] investors will assume the startup is using the best available AI techniques to solve the problem they are solving. Not having state-of-the-art AI techniques powering their software would be like not having a relational database in their tech stack in 1980 or not having a rich Windows client in 1987 or not having a Web-based front end in 1995 or not being cloud native in 2004 or not having a mobile app in 2009. In other words, in a small handful of years, software without AI will be unthinkable.

SEE: AI could lead to 4-hour workdays and World War III, says Alibaba's Jack Ma (TechRepublic)

As mentioned above, we're nowhere near this saturation point today, and so marketing departments have gone wild pretending at AI prowess that they don't yet have. One way to track this is by simply counting the number of mentions of AI in earnings calls (hint: that number has gone through the roof in the last two years, even as successful AI deployments have...not).

Which, by the way, is perhaps the best way to gauge the end of the AI washing phenomenon. We'll know when AI has truly gone mainstream when we stop hearing about it. The more a company feels the need to call it out, the less likely it is that they're actually doing something meaningful with it.

Also see

aitrends.jpg
Image: iStockphoto/Jirsak

About Matt Asay

Matt Asay is a veteran technology columnist who has written for CNET, ReadWrite, and other tech media. Asay has also held a variety of executive roles with leading mobile and big data software companies.

Editor's Picks

Free Newsletters, In your Inbox