Meta AI's Llama 3.1 & Kling AI now accessible to everyone
Kling AI video opens doors worldwide
Total reading time around 4 minutes.
Welcome to Visually AI!
🔮AI News This Week
Meta AI Releases Llama 3.1 Models: What You Need to Know
Meta AI released its Llama 3.1 models, including a 405B variant that stands out as the company's first open-source "frontier-level" AI model.
Llama 3.1 boasts a large context window of 128K tokens and is comparable to closed source model leaders like OpenAI's GPT-4, Google Gemini Ultra, and Anthropic’s Claude 2.
The models are open-source and free, making them accessible to developers and researchers. Recent tests show that Llama 3.1 405B is competitive with GPT-4 and Claude 3.5 Sonnet, excelling in multiturn reasoning and coding tasks.
Its open-source nature and cost-effectiveness make it a compelling choice for those looking to leverage advanced AI technology.
Learn more about Llama 3.1 here.
🎬Kling AI Video available to anyone
Finally, you don’t need a Chinese cell phone number to use Kling AI’s video generator!
Kling AI is a revolutionary text-to-video model that allows users to create stunning, high-quality videos with advanced 3D mechanisms and cinema-grade quality, offering features such as large-scale reasonable motion, up to 2 minutes of video generation, and simulation of physical world characteristics.
I used the web app and I learned after several tries, you need to go to this link: klingai.com
Then, you’ll need to create an account with your cell phone number (any country) and it will send you a verification code to enter on the app.
New users automatically get 66 free credits, which are replenished each day you log in to the app. (Video generations = 10 credits)
A quick demo:
Try the Kling AI web app here.
Kling is good for text-to-video, but excels with image-to-video, as you can see by my result below:
📸AI Snapshots
OpenAI's SearchGPT is a prototype that combines the conversational capabilities of AI models with real-time web information to provide fast and timely answers with clear and relevant sources, aiming to make finding information online faster and easier. Sign up on the waitlist here.
Mistral AI announced Mistral Large 2, a significantly enhanced AI model that excels in code generation, mathematics, and reasoning, with advanced multilingual support for dozens of languages and 80+ coding languages, and a large 128k context window. Download or try Mistral Large 2 in LeChat here.
Runway allegedly trained its Gen-3 model using thousands of pirated YouTube videos and movies without permission, sparking concerns over copyright infringement and ethics.
Udio AI's new v.15 model offers advanced tools for music creators, including improved audio quality, stem downloads, key guidance, new creation workflows, global language support, shareable lyric videos, and audio-to-audio capabilities. (link)
PixVerse V2 has enhanced video generation capabilities, including 8-second video creation, improved details and motion, and consistent style, subject, and scenes across multiple clips for seamless storytelling. (link)
🛠️ This Week’s AI Tools
Image to Prompt: Need a prompt for an image? AP’s Prompt generator Glif gives you prompts for uploaded images to use on Midjourney, Dalle, SD, Luma, and Runway’s Gen-3. (link)
GummySearch: Reddit audience research tool searches keywords, for topics, pain points, and sentiment. (link)
Exploding Topics: Tracks rapidly growing trends, before they take off. (link)
Thunderbit: Comprehensive video hosting platform offering advanced AI-powered search, customizable streaming, and analytics for creators, businesses, and educators. (link)
Bytez: Explore trending AI research papers, get summaries and ask questions to learn. (link)
Watch how it works:
💻Visually AI on YouTube
I highlight the range of Luma’s Dream Machine features with several results in this video:
Luma Labs AI Dream Machine | What can it produce? Results only (No Edits)
🖼️ Image Prompts
Prompt: Sunset on a deserted town street, vintage car parked by the roadside, silhouettes of abandoned carnival with ferris wheel in the background, warm golden hour lighting, cinematic atmosphere, ultra-detailed, hyper-realistic, wide aspect ratio
Prompt: An old shipwreck on a foggy coastline at dusk, with the remains of the ship silhouetted against the setting sun. The sea mist creates an eerie, atmospheric effect. The shot is taken from a distance, through the twisted branches of a nearby tree, adding depth and a sense of foreboding.
🎞️ Video Prompt
Prompt: Static low-angle tracking shot captures a man walking down an urban street inParis. the man is dressed in a black hoodie and black pants, his head covered by a hood. The background of the image is intentionally blurred to create a sense of depth. The street is lined with buildings on both sides, each building displaying various signs and advertisements.
Thanks for reading, and have a creative week!
This is crazy, I hadn't heard of Kling or the Dream Machine. Was not prepared for such quality videos. Kind of worried YouTube will be flooded with videos that look really good, but the creator doesn't actually care to make the content great, if that makes sense