• AI KATANA
  • Posts
  • Midjourney Bans AI-Generated Images of Trump and Biden

Midjourney Bans AI-Generated Images of Trump and Biden

Also: Nvidia GTC is on deck. What should you expect?

Author

AI KATANA
March 17, 2024

Hi!

In today's newsletter, we delve into significant movements across the AI landscape, starting with Midjourney's proactive stance against misinformation by banning AI-generated images of key political figures ahead of the U.S. presidential election. In the corporate arena, Nvidia garners attention with its anticipated Nvidia GTC event, expected to unveil groundbreaking AI advancements and its strategic acquisition talks with AI platform Run:ai, signaling a significant expansion in AI infrastructure capabilities. In healthcare, Thailand's innovative use of AI for mental health treatment marks a significant step towards leveraging technology for societal well-being. The industry also witnesses a surge in developments from Elon Musk's xAI unveiling the Grok-1 architecture to NewsGuard's launch of AI-powered anti-misinformation tools, reflecting a growing focus on ethical AI utilization. Additionally, the venture capital sector is buzzing with Groq's ambitious funding round aimed at scaling its AI chip production, amidst a backdrop of strategic acquisitions and internal venture capital shifts. These developments underscore a dynamic period in AI's evolution, highlighting its growing impact across various sectors and the continuous efforts to navigate its ethical implications.

Sliced:

  • ⚠️Midjourney Bans AI-Generated Images of Trump and Biden

  • 👏 Nvidia GTC is on deck. What should you expect?

  • 💰Nvidia in talks to acquire AI infrastructure platform Run:ai in massive deal

  • 🧑‍⚕️AI being used to treat mental health patients

Midjourney, a popular AI image generation platform, has banned users from creating images of former U.S. President Donald Trump and current President Joe Biden. This decision comes amidst growing concerns over the potential spread of misinformation and deepfakes related to the upcoming U.S. presidential election. By prohibiting the generation of images featuring these political figures, Midjourney aims to prevent its platform from being used to create misleading or manipulated content that could influence public opinion.

Nvidia's annual GPU Technology Conference (GTC), returning to an in-person format in San Jose after five years, promises to be a monumental event in the AI industry, attracting close to 20,000 attendees and over 300,000 online participants. The highlight is expected to be the keynote by CEO Jensen Huang, who has previously described ChatGPT as the "iPhone moment for AI," hinting at significant insights into AI's transformative potential enabled by Nvidia. This year, the event is expected to showcase practical AI use cases, especially in generative AI, to inspire creativity and application in various sectors. Furthermore, the conference is anticipated to reveal new technology advancements, including the B100 Blackwell chip, which promises significantly faster training capabilities than its predecessors, despite concerns about its high power consumption. The societal impacts of AI, particularly in healthcare, smart cities, and sustainability, will be another focus, with sessions on transforming drug discovery, modern medicine, and agricultural practices.

Nvidia is in the advanced stages of discussions to acquire the AI infrastructure platform Run:ai, in a deal potentially valued at up to $1 billion. This significant acquisition aims to strengthen Nvidia's position in the AI sector by integrating Run:ai's specialized orchestration and virtualization software, which is designed to optimize AI workloads on GPUs. Founded in 2018 by Omri Geller and Dr. Ronen Dar, Run:ai has rapidly made strides in the industry, raising $75 million in a Series C funding round in March 2022, bringing its total funding to $118 million. The company's platform, which utilizes Kubernetes-based container technology, uniquely addresses the computing needs of AI operations, from allocating GPU resources to managing extensive AI cloud infrastructures. This move marks Nvidia's first acquisition in Israel since it bought Mellanox for $6.9 billion in March 2019 and comes at a time when Nvidia's valuation has soared to over $2 trillion, driven by its dominance in the high-end AI chip market and the explosive growth in AI applications like ChatGPT.

In Thailand, AI technology is being utilized to assist individuals at risk of depression and suicidal thoughts, offering immediate access to professional help. This initiative employs AI DMIND, integrated into the National Health Security Office's (NHSO) Line service, enabling users to quickly assess their mental health status. Developed by Chulalongkorn University's Faculties of Medicine and Engineering, along with the NHSO and the Department of Mental Health, AI DMIND is trained to recognize signs of depression and suicidal ideation by analyzing recorded voices from the department's mental health hotline database.

🛠️ AI tools updates

Elon Musk's AI startup, xAI Corp., has released the weights and architecture of its Grok-1 LLM as open source, marking the company’s inaugural venture into open-source distribution. This release includes the foundational elements of Grok-1’s network structure and the base model weights adjusted during its training phase, showcasing how the model processes and translates input data into output. The Grok-1 model, embodying a 314 billion parameter "Mixture-of-Experts" design, emphasizes a specialized approach where multiple sub-models collaborate for optimized predictions across diverse tasks, highlighting xAI's ambition to forge new paths in AI development without fine-tuning for specific applications.

NewsGuard is launching a new suite of AI tools designed to combat the spread of misinformation, especially related to Joe Biden and Donald Trump as the 2024 election approaches. This service aims to counter AI-generated false information, images, video, and audio to ensure foreign governments, political actors, and internet trolls can't influence global election outcomes. NewsGuard CEO Gordon Crovitz highlighted the increased production of cheap, targeted, and persuasive content by malign actors utilizing AI.

💵 Venture Capital updates

Groq, an AI chip startup, is embarking on a new funding round in the upcoming quarter as it aims to reach an ambitious goal of shipping 1.5 million language processing unit (LPU) chips by the end of 2025, a significant leap from its current deployment of approximately 4500 chips. The firm has already raised $367 million, with notable investors including Tiger Global and D1 Capital leading a $300 million round in 2021, bringing the company's valuation to $1 billion. Amidst this expansion, Groq has recently acquired Definitive Intelligence to broaden its cloud business and intends to sell its chips directly. Founded in 2016 by Jonathan Ross, a former leader in Google's Tensor Processing Unit development, Groq focuses its LPUs on machine learning inference, distinguishing itself from other AI chips by not supporting training workloads. This strategic move is set against a backdrop of internal conflict at Social Capital, a venture capital firm involved with Groq, where Chamath Palihapitiya, the head of the firm, recently dismissed two partners who had formed a special purpose vehicle to invest in Groq.

🫡 Meme of the day

⭐️ Generative AI image of the day