AI’s sudden great leap into usefulness

Despite its huge potential, the artificial intelligence space has been something of a backlog in the investment world. There are companies that have ridden the AI ​​wave in important ways: Google claims to have refined many of its services using AI, machine learning has boosted sales of Nvidia’s GPUs, and TikTok’s algorithm is reportedly a big part of that , which keeps users going back to his short videos.

But it’s hard to find a pure-play AI company that’s risen on the technology’s back, or identify a major new market that’s been created. This picture could soon change, and in a big way.

Something significant has happened in AI over the past few weeks, according to Pat Grady, a partner at Sequoia Capital. Generative systems – those that automatically generate text and images from simple text prompts – have advanced to a level where they could have far-reaching business applications. A partner at another leading Silicon Valley venture capital firm, who describes AI’s recent history as a graveyard for start-up investors, also reports that the race for breakthrough applications for this new technology is on.

Since the introduction of OpenAI’s GPT3 text writing system two years ago, generative models like this have been all the rage in AI. The ethical questions they raise are profound, ranging from biases they might ingest from the data they’ve been trained on to the risk that they might be used to spew misinformation. But that hasn’t stopped the hunt for practical applications.

Three things have changed to turn these systems from clever party tricks into potentially useful tools.

One is that the AI ​​systems have gone beyond text. Last week, Meta introduced the first system capable of producing a video from a text or image prompt. It was thought that breakthrough would be two years or more away. Not to be outdone, Google responded with not one, but two of its own AI video systems.

The biggest AI breakthrough of this year has been in image generation, thanks to systems like OpenAI’s Dall-E 2, Google’s Imagen and start-up Midjourney. Emad Mostaque, the London-based hedge fund manager behind Stable Diffusion, the latest image generation system taking the AI ​​world by storm, claims that images will be the “killer app” for this new form of AI. For the generation that grew up on TikTok and Snapchat, this new creative tool could be profound, he says. It also poses an obvious threat to anyone whose livelihood relies on creating images in any other way.

The second big change comes from the rapidly decreasing cost of training huge AI models. Microsoft’s $1 billion support of OpenAI three years ago highlighted the prohibitive cost of ever larger models. New techniques that make it possible to obtain high-quality results by training neural networks with fewer layers of artificial neurons are changing the picture. The computational resources used to train stable diffusion would have cost only $600,000 at market prices, corresponding Mostaque.

The third change was the availability of the technology. Google and OpenAI have been wary of making their technology widely available, in part out of concerns about potential abuse. In contrast, Midjourney’s image system is available to all users via a freemium pricing model. Stable Diffusion has gone a step further, disclosing its software and releasing details on how it trained its system. This allows other organizations to train an image model on their own datasets.

Much attention has been paid to the risks posed by such generative systems. They produce new images or text based on millions of examples they have learned from without understanding the underlying material. This can lead to both nonsensical results and intentional misinformation.

But in a business environment, at least some of these shortcomings could be controlled. The trick will be finding ways to embed the technology into existing work processes, develop tools that can suggest new ideas or speed up creative production, with human workers filtering the output. The idea is already being used to generate computer code.

The big question now, says one investor, will the existing giants of industries like marketing, media and entertainment be the first to embrace these powerful new creative tools? Or will they be disrupted by a new generation of upstarts with their roots in AI?

[email protected]