OpenAI, the company that put generative AI on the map, is making bank. First came the very spurious (and click-baity) headlines suggesting OpenAI was burning money at a wild rate. Now come the stories that the company is going to surpass $1 billion over the next year.
The Information (paywall) says OpenAI is earning more than $80 million a month, which is a significant bump from the $28 million it made for the whole of fiscal 2022.
Of course, revenue is not profit, so it’s possible that OpenAI is not serving champagne and caviar at staff lunches quite yet. But still, it does show just how mainstream AI has become in a very short space of time.
Many have spoken about AI as the next great revolution in computing. So considering Meta (back when it was Facebook) took five years to become profitable, OpenAI’s fast rise should give all the Big Tech companies something to be twitchy about.
In a world of Balenciaga Pope and Ghandi’s selfie, spotting an AI-generated image is getting harder and harder. Fortunately Google DeepMind could have the solution.
SynthID is an AI-powered tool designed to watermark images. It does this by replicating the original photo, and then making minor adjustments to some of the pixels. The resulting pattern is purportedly imperceptible to mere humans, but SynthID can supposedly still identify the watermark - even if the image is cropped.
SynthID will initially be available via Imagen, Google’s AI generation platform. So perhaps keep this in mind when using AI for your next headshot.
We’re quickly moving from a place where AI-generated images are no longer created just for fun or bragging rights, but also for use in advertising and politics. Therefore, identifying manipulated images is only going to become more important.