AI image generation service Midjourney is getting into hardware, whatever that means

It’s been a minute since anyone said anything new about Midjourney. You remember them, right? The service that popularised creating custom images using little more than a sentence prompt before OpenAI and ChatGPT came along to suck all the oxygen out of the room?

In a bid for everyone’s attention, the company has announced a mysterious new hardware product. It’s not saying what it is or even giving proper hints as to what it has in mind, it’s just looking for interested souls to join the team.

A Midjourney to the future

There are a few clues, probably intentionally scattered around by whoever runs Midjourney’s social media. The new device is “not gonna be a pendant” and there might be several different types of gadgets in the offing (“we have multiple efforts in flight“). What it’ll do? That’s entirely mysterious.

You’d expect a device coming from Midjourney would incorporate what it has done in the past. A screen, along the lines of Samsung’s Frame TV, that will generate custom artwork based on verbal prompts? That makes a sort of sense, though why you’d create custom hardware for that instead of launching a smart TV app is a legitimate question.

It’s far more likely that the company is simply throwing things at the wall to see what sticks. The market for AI-generated images is either saturated, suffering from intentional poisoning, or has become a victim of recursive learning as the models used to create pictures run out of human-generated data. In that case, it’s time for a change that lets companies use existing functional models in a new way. The easiest answer is hardware. The difficult one is convincing folks to use this new hardware.

That, at least, doesn’t have to be answered for a while. There’s no timeline on when or where (or if) we’ll ever see this product. But hey, at least investors can see that we’re talking about Midjourney again. That was the whole point, right?