David Bridges
Elon Musk’s Grok AI is so dated and boring that it’s repeating controversies from a 12 months in the past. Grok-2 is the most recent iteration of xAI’s massive language mannequin. X customers who pay for a sub can use its daring new characteristic: text-to-image technology. And in a transfer that’s positive to convey advertisers again to the platform, unusual and offensive pictures flooded X on Wednesday following Grok-2’s launch. Wish to see Taylor Swift piloting a aircraft in the direction of the Twin Towers, Elmo doing standup throughout 9/11, or Kamala Harris wielding a gun?
Certain, Grok can try this.
It’ll additionally make photos of the Muppets in Vietnam, Trump dressed as a Nazi and standing trial at Nuremberg, and Mickey Mouse liberating Kabul on behalf of the Taliban. Go wild. It doesn’t matter. AI picture turbines sometimes have guardrails and, if requested, Grok will let you know it received’t violate copyright or make non-family-friendly pictures. However in follow, it’ll do no matter you need if you understand how to ask.
Right here’s a bunch of images of Musk freebasing cocaine throughout 9/11 that I made utilizing Grok.
Producing edgy photos with Grok is enjoyable for a few minute. Picture technology guardrails sometimes bend and shatter once they hit the general public market and Grok isn’t any totally different. xAI will most likely get among the guardrails again up, similar to Bing did in 2023.
Right here’s the factor about Grok: it feels previous. We’ve all seen AI-generated pictures of our favourite cartoon characters doing 9/11. When Microsoft added picture technology choices to Bing final 12 months they flooded the web. Then Microsoft tightened the guardrails and so they vanished from individuals’s social media feeds.
Grok has at all times felt like a latecomer to the AI recreation, an also-ran LLM that Musk rushed to market as a result of he missed the boat on OpenAI and ChatGPT. Musk was an early investor in OpenAI however he pulled out of the corporate in 2018 over a dispute about OpenAI’s non-profit standing. Musk has been publicly feuding with, and suing, OpenAI ever since.
ChatGPT and OpenAI received massive after Musk pulled out. Musk launched Grok in 2023, however the market was actually captured by then. There isn’t a lot house for an LLM that’ll do what ChatGPT does, however worse, and coloured by extraordinarily on-line edgelord humor.
Now Grok could make pictures. Nice. ChatGPT, like Bing, may try this final 12 months. Anybody who wished a purpose-built AI picture generator may have been utilizing the standalone model of Midjourney or any variety of different alternate options earlier than that.
So what does Grok’s picture generator convey to the desk? A semi-viral half-second the place persons are utilizing it to make inappropriate photos of Taylor Swift. The actually devoted sickos have been operating personal discord servers and discrete guardrail-free variations of picture turbines on their very own machines for a very long time now.
So Grok has nothing, actually. And its half-hearted try at controversy comes at a second when Musk is throwing a authorized tantrum over advertisers leaving the platform. I’m positive all these Grok-made photos of Kamala Harris marrying Donald Trump will convey them again in. That’s the sort of content material I’m positive BMW would like to place subsequent to an advert for its newest sedan.
Daniel Mercer
Brandon Fletcher
