- The AI Crunch
- Posts
- š„£ Google Halts AI-Generated Images
š„£ Google Halts AI-Generated Images
PLUS: Pika Adds Lip Sync Feature
Subscribe || Advertise || AI Tools || The AI Tech Stack || Top AI Repos
Read time: 3.0 minutes
Greetings Tech Enthusiasts, welcome to the latest edition of The AI Crunch. Prepare for a deep dive into the core of today's tech revelations. We have a collection of insightful updates and noteworthy developments to share with you. Stay tuned as we explore the unfolding advancements in the world of artificial intelligence. Let's delve into the details together!
Top Stories
Google Halts AI-Generated Images Amid Racism Concerns
Well, brace yourself for the latest chapter in the AI artistry saga. Google's Gemini AI, attempting to shatter historical stereotypes, ventured into generating diverse depictions of the US Founding Fathers and even threw in 1940s German soldiers for good measure. However, this bold move took an unexpected turn, with users stumbling upon inaccuracies that fueled conspiracy theories. The plot thickened as Gemini seemingly erased historical nuances, portraying US senators from the 1800s as a surprisingly diverse group of native americans.
In a swift response, Google decided to hit the pause button on Gemini's artistic escapades, particularly its knack for crafting people-centric images. An official statement assures us that they're diligently working to address the recent hiccups, promising a more polished version in the near future. So, for now, Gemini's canvas remains blank on the portrait front, awaiting a grand return with improved historical accuracy.
The incident sheds light on the delicate dance between AI creativity and historical precision. As Gemini takes a brief hiatus, one can't help but wonder about the intricacies of crafting an AI that accurately captures the nuances of the past without inadvertently rewriting history. Will Gemini rise from the temporary setback as a more seasoned digital portraitist, or will it continue to raise eyebrows with its avant-garde interpretations? Only time will tell in this AI-driven artistry odyssey.
Pika Adds Lip Sync Feature Powered by ElevenLabs
Grab your popcorn, as Pika, the scrappy contender in the AI video space, has just unleashed a game-changing move. While OpenAI tantalizes with Sora demos, Pika refuses to play second fiddle, unveiling the Lip Sync feature. For a monthly fee of $58, early access is granted to Pika Pro users and the select few in the "Super Collaborators" club. This groundbreaking tool empowers users to infuse AI-generated voices into their videos, perfectly synchronizing animated lips with spoken dialogues, thanks to ElevenLabs' prowess.
The introduction of the Lip Sync feature propels it ahead in disrupting traditional filmmaking software. This move specifically addresses one of the lingering challenges in AI's utility for crafting extended narrative films. Unlike most other AI video generators, Pika now offers a native solution to seamlessly integrate spoken dialogues, eliminating the need for cumbersome third-party tools and post-production hacks that often result in a budget-friendly, Monty Python-esque quality.
The AI video battleground continues to escalate, with companies like Runway making strides in features like Multi Motion Brush, adding region detection for automatic highlighting and selection of objects in videos. Meanwhile, Pika, despite lacking a similar granular "brush" tool, allows users to edit components and expand the canvas. As these AI video players vie for supremacy, concerns about training data and ethical considerations persist. Questions raised by figures like Ed Newton-Rex, CEO of Fairly Trained, echo the broader industry sentiment. Despite these uncertainties, the relentless innovation in AI video generation signals a fierce competition, leaving the filmmaking community on the edge, contemplating the implications of this technological arms race.
News
Quick bites to start your morning:
Klarnaās AI assistant handled +2.3M chats in its first month (KL)
Tumblr to sell data to OpenAI and Midjourney (TV)
Google DeepMind released a new video game AI model, āGenieā (G)
NYC wants to make "deceptive AI" / deepfakes illegal (FTC)
Apollo states that the AI ābubbleā is worse than even the dotcom era (F)
Adobe releases AI feature that allows users to chat with PDFs (A)
Tyler Perry stops $800M studio expansion due to AI fears (HR)
Elon leaks a potential partnership between X and Midjourney (X)
Sora to soon be available via Microsoft Copilot (MSP)
AI Funding
Mistral AI: OpenAI rival raised $16M from Microsoft (TC)
CellVoyant: An AI-first biotech startup raised $10M (FN)
Moonshot AI: Startup building large language models raised $1B (TC)
AI Tools
Newest AI tools (AI Tool Leaderboards):
AI GIFs: AI-Powered GIF generation tool
Typho: AI-Powered artistic portrait creation platform
Watermarkly: Watermark photos online for free
Elite-AI-Write: Empowering content creation with AI
Proseable: AI-Powered language learning assistant
Other
New Sora videos just dropped on TikTok by OpenAI team and people are going crazy.
Nothing in this video is realš¤Æ
6 wild new examples: š§µš
1. Labrador Hacker
ā Min Choi (@minchoi)
2:14 PM ā¢ Feb 21, 2024
Excited to share: Figure 01 completing real-world tasks
This is end-to-end autonomous
We have made advances in our autonomous navigation, learned perception models, manipulation robust to pose variation, & generalizable systems for future applications
ā Brett Adcock (@adcock_brett)
6:38 PM ā¢ Feb 26, 2024
If you have anything interesting to share, please reach out to us by sending us a DM on Twitter: @theaicrunch