AI For Humans: Weekly AI News, Tools & Trends

Google Wins AI Video, ChatGPT Gets a 1-800 Number & More Massive AI News

whatshot 156 snips

Dec 19, 2024

Exciting advancements in AI take center stage with Google’s breakthrough VO2 video model and Gemini updates that redefine quality. OpenAI's introduction of a 1-800 number for ChatGPT humorously explores its role in crisis management. We also discuss the implications of creators' rights in the AI landscape and witness whimsical experiments with Google Whisk. Plus, the transformative potential of new AI tools like video watermarking and 3D avatar creation is highlighted, proving that innovation keeps getting more fun and practical!

44:15

forum

Ask episode

web_stories

AI Snips

view_agenda

Chapters

menu_book

Books

auto_awesome

Transcript

info_circle

Episode notes

insights

INSIGHT

AI Rights Deal

CAA and YouTube partnered on an AI rights deal.
This aims to protect artists' likenesses and potentially enable profit from their AI-generated use.

volunteer_activism

ADVICE

Broader Likeness Protection

Advocate for broader likeness protection, not just for celebrities.
Everyone should have the option to control and monetize their AI-generated likenesses.

question_answer

ANECDOTE

MKBHD's Plant

A Sora video of a tech reviewer included MKBHD's iconic plant.
This highlights the unexpected ways AI can incorporate details and the potential for likeness issues.

Get the Snipd Podcast app to discover more snips from this episode

Calling for Help: ChatGPT and the 1-800 Revolution

01:27 • 2min

chevron_right

AI Advancements Unveiled: Google's Gemini and VO2

03:39 • 13min

chevron_right

The AI Landscape and Its Evolving Challenges

16:12 • 19min

chevron_right

Innovative AI Tools: Watermarking and 3D Avatar Creation

35:30 • 2min

chevron_right

Exploring the Latest AI Tools and Creative Experiments

37:52 • 3min

chevron_right

Exploring AI Creativity through Iconic Characters and Products

41:20 • 3min

chevron_right

#83412

CAP4D: Creating Animatable 4D Portrait Avatars with Morphable Multi-View Diffusion Models

Ruihang Zhang

Felix Taubner

Mathieu Tuli

This paper presents CAP4D, an approach that uses a morphable multi-view diffusion model to reconstruct photoreal 4D (dynamic 3D) portrait avatars from any number of reference images. It demonstrates state-of-the-art performance for single-, few-, and multi-image 4D portrait avatar reconstruction and bridges the gap in visual fidelity between single-image and multi-view reconstruction techniques. The method involves generating novel views from reference images and using 3D Gaussian splatting for real-time animation and rendering.

#51899

Barry Trotter and the Unauthorized Parody

Rodger Roundy

Michael Gerber

This book is the first in the Barry Trotter series and parodies the Harry Potter books. It follows Barry Trotter, a spoilt brat who has been at Hogwash (Hogwarts) for an extended period, and his friends Lon and Ermine as they embark on a mission to stop the production of a film based on Barry's life. The story is filled with comedic elements, including parodies of England's Sun newspaper, pop culture, and the business of publishing and moviemaking. The book is known for its crude humor and breaking of the fourth wall, which adds to its comedic effect.

#80495

Pika 2.0

AI Video Creation Platform

Pika Labs

Pika 2. 0 is an advanced AI video model that empowers creators to craft immersive, personalized stories with ease. It features Scene Ingredients for intricate customization, improved text alignment, and advanced cinematic effects such as 'Bullet Time' and 'Dolly Left' shots. The platform also includes lifelike character animations, dynamic character movements, and effortless editing capabilities. Pika 2. 0 is designed to democratize video production, making high-quality video creation accessible to everyone, from beginners to seasoned professionals.

#57898

Gemini 2.0 Flash Experimental Advanced

Experimental AI Model

Google AI Team

Gemini 2. 0 Flash Experimental Advanced is an experimental AI model that introduces new features such as a Multimodal Live API for real-time vision and audio streaming, improved performance on complex tasks like coding and math, and enhanced multimodal understanding. It also includes native image generation and controllable text-to-speech capabilities. This model is available for testing in Google AI Studio and Vertex AI, with general availability expected in the future.

#95003

Pika 2.0

AI Video Generation Model

Pika Labs

Pika 2. 0 is a revolutionary AI video creation platform that allows users to generate high-quality videos from text descriptions and images. It includes features such as Scene Ingredients for detailed scene customization, advanced cinematic effects like 'Bullet Time' and 'Dolly Left' shots, and lifelike character animations. The platform is designed to be user-friendly, making video production accessible to everyone from beginners to professionals. It is particularly useful for social media marketers, content creators, and filmmakers who need to produce content quickly and efficiently.

#45545

Gemini 2.0 Flash

Google's New AI Model for the Agentic Era

Demis Hassabis and Koray Kavukcuoglu

Gemini 2. 0 Flash is a significant upgrade to Google's AI models, offering twice the speed and enhanced performance compared to its predecessor, Gemini 1. 5 Flash. It supports multimodal inputs and outputs, including images, video, audio, and text. The model introduces new features such as native image generation, controllable text-to-speech, and the ability to call external tools and execute code. It is designed to enable more immersive and interactive applications and is currently available as an experimental version for developers and trusted testers.

#31649

• Mentioned in 2 episodes

OpenAI's Advanced Reasoning Model

OpenAI Research Team

The o1 model, part of OpenAI's 'Strawberry' AI reasoning project, is designed to break down complex tasks into component parts and handle them step-by-step. It excels in math, science, coding, and logic tasks, and has shown significant improvements over previous models in benchmarks such as GPQA and MMLU. The model is available in iterations like o1-preview and o1-mini, with o1 Pro offering enhanced capabilities for users.

#8378

• Mentioned in 6 episodes

The adventures of Pinocchio

Carlo Collodi

Set in 19th-century Tuscany, Italy, 'The Adventures of Pinocchio' follows the misadventures of Pinocchio, a wooden marionette created by the poor woodcarver Geppetto. Pinocchio's journey is marked by his propensity for lying, which causes his nose to grow, and his encounters with various characters, including the benevolent Blue Fairy. Through his experiences, Pinocchio learns valuable lessons about honesty, hard work, and the importance of making the right choices. The novel is a timeless classic that continues to resonate with readers of all ages.

#45494

Meta's MovieGen: The AI Revolution Transforming Video Creation

Why Filmmakers and Creators Can't Overlook the Next Generation of Artificial Intelligence-Enhanced Content

Not specified

This book explores the revolutionary AI technology of Meta's MovieGen, which enables the generation, editing, and personalization of videos using simple text inputs. It delves into how MovieGen is transforming the video creation process, its features such as personalized video generation and precise video editing, and its potential impact on the film and creative industries. The book also touches on the ethical considerations and the future of creativity in the context of generative AI.

#15806

• Mentioned in 3 episodes

Minimax

Vlastimil Tresnak

#58748

VO2 Max: Athlete's Journal

Richard B. Kent

VO2 Max: Athlete's Journal is designed to help athletes of all ages enhance their mental game and overall performance. The book includes 72 multi-themed journal prompts, pre-, mid-, and postseason reflection pages, and a season's worth of Competition Analyses. It also provides a short introduction on the theory behind using writing as a learning tool in athletics.

#3276

• Mentioned in 14 episodes

QualityLand

Marc-Uwe Kling

In 'QualityLand', Marc-Uwe Kling presents a satirical vision of a future where technology and algorithms govern every aspect of life. The story is set in QualityLand, a society where a universal ranking system determines social advantages, career opportunities, and even romantic partners. Peter Jobless, a machine scrapper, feels out of place in this highly automated world and becomes the leader of a group of robotic misfits. When Peter receives an unwanted product from TheShop, the dominant company in QualityLand, he decides to return it, which challenges the infallible algorithms and the very foundations of the society. The book is a sharp critique of consumer culture, technology, and the impact of algorithms on human relationships and identity, balanced with humor and absurd scenarios.

Google's VEO 2 and Gemini updates are… freaking good?! Plus, OpenAI lets you call ChatGPT, o1 is in the API now & Nvidia's Jensen Huang is giving us a new robot brain. Plus, new Pika 2.0 AI video tools, Ilya Sutskever returns to tell us pre-training is over, YouTube and CAA comes together on a deal for protecting celebrity likenesses, Google's Whisk tool is a fun AI toy and we meet an angry old man who calls us wanting help for his furnace whom we then get to drink Monster Milk. IT'S A NEW SHOW, A PRESENT JUST FOR YOU

Join the discord: https://discord.gg/muD2TYgC8f

Join our Patreon: https://www.patreon.com/AIForHumansShow

AI For Humans Newsletter: https://aiforhumans.beehiiv.com/

Join our TikTok @aiforhumansshow

To book us for speaking, please visit our website: https://www.aiforhumans.show/

// SHOW LINKS //

Call ChatGPT aka 1-800-ChatGPT

https://www.youtube.com/live/LWa6OHeNK3s?si=0vogE9s_qVOmp81-

VEO 2 is actually pretty insane

https://deepmind.google/technologies/veo/veo-2/

Knight on a Zebra

https://x.com/emollick/status/1868897308529787248