Grok-2: Elon's Uncensored AI

Good day, humans!

AI News of the Week

• Grok-2: Elon’s Uncensored AI

• Midjourney Improves Image Editor

• Runway’s AI Video Goes Turbo

• AI Goes to Trial, Again

• Humans Replace AI

• Use This Midjourney Style

Grok-2: Elon’s Uncensored AI

This week, Elon Musk's xAI announced Grok-2, a chatbot on par with Claude 3.5 Sonnet and GPT-4-Turbo on the LMSYS leaderboard, with significant improvements in chat, coding, and reasoning, available for Premium users on the X platform.

Additionally, Grok-2 users also have access to the open-source image generator FLUX.1 from Black Forest Labs, which has become a serious competitor to Midjourney. Grok-2 can generate quite realistic images quickly and with few restrictions. Just take a look at the following example:

Why it matters

AI art-generating tools are becoming more powerful, more realistic, and available to more users. But so far, most of these companies have made efforts to control their models after public outcry.

Now, Grok-2 is generating images with minimal censorship, including violent and copyright-infringing content. Musk's response so far has been “Grok is the most fun AI in the world!”:

Midjourney Improves Image Editor

Midjourney's website image editor has received a series of significant upgrades: reframe / repaint / vary region / pan / zoom are now in a single unified interface. Area selection for modification is done with an adjustable brush for more control.

Why it matters

Gradually, AI tools are giving users more control over the results, reducing the importance of the initial prompt, and allowing for precise modifications to the image.

Runway’s AI Video Goes Turbo

Runway introduced Gen-3 Alpha Turbo. It’s 7 times faster than the original Gen-3 Alpha and 50% cheaper, supporting the generation of video clips of up to 10 seconds.

AI Goes to Trial, Again

First, it was AI music generators; now, it's AI art generators going to court: A federal judge in the United States has allowed a landmark copyright lawsuit to proceed against major AI art generators, including Stability AI, Runway, Midjourney, and DeviantArt.

At the center of the legal debate is a crucial question: Does the use of copyrighted materials to train AI systems constitute infringement, or is it considered "fair use," in the same way that humans can mimic or remix styles they've seen?

Why it matters

This lawsuit not only has the potential to redefine the boundaries of AI-created art but also to establish a new framework for the use of copyrighted materials in the training of these advanced systems. The industry is watching closely, and the outcome of this trial could shape the future of artificial intelligence and its relationship with intellectual property.

Humans Replace AI

First, AI made videos in the style of humans. Now, humans are making videos in the style of AI.

Use This Midjourney Style

To use this style, add the parameter --sref 3520822763 at the end of your prompt. Example:
Photographic portrait --sref 3520822763

Thanks for reading. See you next week!
Hi 👋 I'm Erik Knobl, Product Designer by day and explorer of Generative AI on weekends. I share my learnings in this newsletter. Consider subscribing to stay in touch.