The startup behind Elon Musk's wild AI image tool isn’t his own xAI
Elon Musk’s Grok delivered another man-made intelligence picture age include on Tuesday night that, very much like the man-made intelligence chatbot, has not very many protections. That implies you can produce counterfeit pictures of Donald Trump partaking in pot on the Joe Rogan show, for instance, and transfer it directly to the X stage. In any case, it’s not actually Elon Musk’s artificial intelligence organization controlling the franticness; rather, another startup — Dark Timberland Labs — is the outfit behind the disputable component.
The cooperation between the two was uncovered when xAI reported it is working with Dark Woodland Labs to control Grok’s picture generator utilizing its FLUX.1 model. An artificial intelligence picture and video startup that sent off on August 1, Dark Timberland Labs seems to feel for Musk’s vision for Grok as an “hostile to woke chatbot,” without the severe guardrails found in OpenAI’s Dall-E or Google’s Imagen. The web-based entertainment webpage is now overflowed with incredible pictures from the new component.
It looks like Grok’s AI image generator is here. And yes, as expected, very few safeguards pic.twitter.com/AB3IO7A3Dc
— Max Zeff (@ZeffMax) August 14, 2024
Dark Woodland Labs is situated in Germany and as of late emerged from covertness with $31 million in seed financing, drove by Andreessen Horowitz, as per a public statement. Other prominent financial backers incorporate Y Combinator President Garry Tan and previous Oculus Chief Brendan Iribe. The startup’s fellow benefactors, Robin Rombach, Patrick Esser, and Andreas Blattmann, were previously analysts who made Dependability computer based intelligence’s Steady Dispersion models.
As per Counterfeit Examination, Dark Backwoods Lab’s FLUX.1 models outperform Midjourney’s and OpenAI’s simulated intelligence picture generators with regards to quality, as positioned by clients in their picture field.
The startup says it is “making our models accessible to a wide crowd,” with open source man-made intelligence picture age models on Embracing Face and GitHub. The organization says it intends to make a text-to-video model soon, too.
Dark Woodland Labs didn’t quickly answer TechCrunch’s solicitation for input.
Oh my god. Grok has absolutely no filters for its image generation. This is one of the most reckless and irresponsible AI implementations I’ve ever seen. pic.twitter.com/oiyRhW5jpF
