In a few years, no investors are going to be looking for AI startups

This post was written by Frank Chen, Partner at Andreessen Horowitz

Despite the click-baity title, I mean it. I believe that in 2 years, no investor is going to be explicitly looking to fund AI-powered startups.

Given that nearly all the startups we see these days are dressing themselves in AI clothes (powered by machine learning! uses the latest RNNs and GANs! post-deep learning architecture!), this might seem counter-intuitive, to say the least.

But the reason I believe that no investor will be funding startups calling themselves AI-powered startups (and no startup CEO will differentiate themselves as an AI-first company like Google) is because 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.

So ambitious founders will need to invest some other way to differentiate themselves from the crowd — and investors will be looking for other ways to decide whether to fund a startup. And investors will stop looking for AI-powered startups in exactly the same way they don’t look for database-inside or cloud-native or mobile-first startups anymore. All those things are just assumed.

In this way, AI is no different than any other foundational computer science technique that gets widely adopted. Every now and again, the industry invents a new set of “must adopt” computer science techniques that find their way into all important software we use.

It begins with cool kids cloning code from Github and talking about what they found in a sub-reddit, then the research papers start dominating Arxiv, then the meetups go from dozens of enthusiasts to sold out extravaganzas at venues like Moscone Center, then the coding bootcamps and MOOCs offer the nanodegrees, and then eventually the techniques are embodied in the mainstream IDEs like XCode and Visual Studio, and then you discover that you can’t earn a computer science degree without a healthy dose of the stuff in the Standard Curriculum.

Until we get to that point, we’ll have some froth in salaries and valuations and pitch decks — just like we’ve always had when the platform shifts. Remember when PowerBuilder or Win32 or Objective C or Hadoop coders or [insert favorite platform tech here] engineers were over-priced?

I’m looking forward to the day when AI is powering all the software we write by default. And also looking forward to hearing what the shiny new thing will be then.

Join 25,000+ people who read the weekly ??Machine Learnings?? newsletter to understand how AI will impact the way they work.

This post originally appeared on the Machine Learnings blog