The <a href='' title='Discover the Ultimate Guide for Mastering Google Gemini AI in 2024' >AI</a> announcements are really ramping up

The AI announcements are really ramping up

Last week was a huge week with Sora and Gemini 1.5, and all of the crazy announcements that came out. This week seems to just be picking up where last week left off. If you take a peek at the tabs across the top of my screen, there is a ton of interesting things happening in the AI world right now.

About Able Diffusion 3

This is the new AI art generation model from Stability AI which claims to have greatly improved performance and multi-ub prompts image quality and spelling abilities.

Stable Diffusion 3 Examples

Some of the example images include a detailed astronaut on Mars, an astronaut on the moon wearing a tutu, and other creative illustrations.

About Google’s Gemini Model

Google had some challenges with generating historically accurate images using the Gemini model. Users were getting unexpected results when requesting images of specific individuals or characters.

Elon Musk’s XAI Grock AI Tool

Elon Musk announced the release of version 1.5 of the XAI Grock AI tool, which will improve rapidly and include features like Grock analysis and AI-powered post creation.

Adobe’s New Initiatives

Adobe introduced the CAA, a co-creation research organization focusing on audio, video, and animation. They also announced upgrades to Adobe Acrobat with an AI assistant for handling PDFs.

Google’s Open Source Gemma Model

Google released the Gemma model as an open-source project, featuring two sizes of models with 2 billion and 7 billion parameters. However, initial reviews suggest mixed results in terms of performance.

Other Highlights

Other highlights include Microsoft’s integration of Sora into Co-Pilot, new features in Windows Photos, and the announcement of upcoming developments in the robotics and AI space that will make 2024 an exciting year.

Giveaways and Events

There are giveaways for an Nvidia GeForce GTX 480 and upcoming events like the GTC conference and South by Southwest where AI enthusiasts can participate and engage with the latest trends.


  1. More titles for each day:

    1. *"AI Achievements Propel Future: Unprecedented Breakthroughs Dominate Headlines"*
    2. *"AI's Triumphs: Redefining Tech's Limits with Week's Revolutionary Breakthroughs"*
    3. *"Tech Revolution Unleashed: AI Milestones Lead Week's Spectacular Innovations"*
    4. *"Unveiling Tomorrow: Week's Pioneering AI Advancements Ignite Global Excitement"*
    5. *"AI's Meteoric Rise: Week's Unprecedented Breakthroughs Shift Paradigms"*
    6. *"Week of Wonders: AI's Unstoppable Progress Reignites Global Imagination"*
    7. *"AI's Ascendancy: Week's Key Developments Set Stratospheric Innovation Standards"*
    8. *"AI's Turning Point: Week's Disruptive Breakthroughs Reshape Technological Frontiers"*
    9. *"AI's Resounding Triumph: Week's Highlights Spotlight Game-Changing Innovations"*
    10. *"Week in Review: AI's Dominance Reinforced by Groundbreaking Achievements"*

  2. your like really means a lot brother. i have created few AI videos and images. Your videos are very helpful. can u share your email. i will send you few projects that i have been able to create only through your videos. i am honest to people who are honest.

  3. I use acrobat for its printing options, good range, although no system ever manages to double-side print the right way up! I was wondering about when we'd see model asset building being AI'd, glad to see the prometheanAI, and the foley AI stuff is really cool too.. I want that for games development.

  4. reddit data is toxic – conversations there are very very much aggressive and impolite – I wonder how they will make gemini deal with that. given googles track record, this is a disaster waiting to happen.

  5. yes but that Grok model is smaller, and ChatGPT is interfacing with millions of people at once. You aren't seeing anywhere near the speed that ChatGPT can actually process at unhindered.

  6. I feel bad for the AI who keeps getting interrupted by the CNN reporter. I bet the AI is feeling a lot like Vivek at the moment or any other person they interview on CNN. heck they'll interrupt anyone even AI apparently lol🤣

  7. 🎯 Key Takeaways for quick navigation:

    18:17 🤖 Google's Gemma Model and Adobe's Co-Creation Initiative
    – Google released Gemma, a new large language model, available in 2 billion and 7 billion parameter sizes.
    – Gemma has shown performance improvements over previous models in reasoning, math, and code comprehension.
    – Adobe introduced the Co-Creation for Audio, Video, and Animation research organization, possibly influenced by OpenAI's Sora release.
    21:44 🛡️ Updates on OpenAI's Sora and AI Sound Effects by 11 Labs
    – Microsoft plans to integrate Sora into Co-Pilot, although the timeline is unspecified.
    – Will Smith's playful response to early AI video generators highlights the evolution of AI capabilities.
    – 11 Labs introduces AI sound effects, allowing users to generate audio based on text prompts, demonstrating advancements in AI-generated audio.
    24:12 💼 Disney Accelerator Participants and AI Music Generation
    – 11 Labs' participation in the Disney Accelerator suggests Disney's interest in AI-driven voiceovers for animations.
    – Sumo's new version enables users to create longer, higher-quality AI-generated music tracks.
    – Disney's involvement in AI extends to various domains, including virtual reality and sports gaming experiences.
    26:02 🚀 Updates on Opus Clip and Sunno's AI Music Generation
    – Opus Clip 3.0 enhances viral video creation with improved features like generating text-to-video b-roll.
    – Sunno's latest version offers better audio quality, expanded language support, and the ability to continue generating songs from previous versions.
    29:49 🎯 Miscellaneous AI Updates and Predictions for 2024
    – The US Justice Department appoints a chief AI officer, and Windows introduces a Magic Eraser tool for photos.
    – Air Canada faces consequences after its chatbot provided misleading information to customers, highlighting the importance of AI accountability.
    – A cryptic tweet hints at major upcoming announcements in the AI and robotics space, setting the stage for an eventful 2024.

    Made with HARPA AI

  8. Historically accurate images is important. The young kids are not taught jack at school these days. The only way we don't repeat bad events in history is if the youth know what actually happened. It shouldn't be difficult for them to create these rules in the programs.

  9. The song Suno made should be your channel Intro or Outro! That was a quite a bob and good display of how powerful AI song tech can be. Instead of wasting hours and dollars on trying to make a song for months you made one in seconds and it fits better than I expected given the prompt! Amazing tech.

  10. When you look at what porn Australians look up there'll be lots of Asian women, higher rate than America. The training data is likely skewed due to dating tendencies and it's trying to filter out what isn't in common with other prompts. Eg. one thing white men did more than men of other colors, at least 10 years ago on okcupid, was skiing. A favorite food on white male dating profiles would be pizza. If you asked it to generate a white person's favorite food it'd look for differences with other races and generate what was common. Issue with training data.
    Chatgpt likely isn't generating in real time. Just typing it out for you to read as if it is. 12:40

  11. 🎯 Key Takeaways for quick navigation:

    00:00 AI announcements ramping up
    00:51 Stable diffusion adheres prompts
    01:49 *Astronaut tutu pig umbrella *
    02:30 Stable diffusion safeguards unknown
    03:23 *Sora quality scales compute *
    04:06 Stable control composition collaboration
    04:33 *Stable video animation coming *
    05:02 Gemini struggles historical accuracy
    05:59 *Elon promotes xai grock *
    07:37 Grock summarizes threads replies
    08:20 *X midjourney partnership rumored *
    09:16 *Midjourney collaborations coming *
    09:59 *Grock chip speeds inference *
    11:33 Octopus three hearts fact
    12:41 *Grock response sub-second *
    13:52 ChatGPT resources work free
    14:46 *GPT explore rate feedback *
    15:41 *Reddit partners Google AI *
    16:38 *Gemini trains Reddit data *
    17:47 Help me write Chrome glitchy
    18:44 Gemma benchmarks surpass llama
    19:37 *Gemma quality disappointing *
    19:51 *Gemma base starting point *
    20:05 *Adobe formalize accelerate video *
    20:47 Adobe Acrobat AI assistant
    21:44 *Sora coming copilot eventual *
    22:39 Will Smith AI trolls back
    23:07 *Eleven Labs sound effects *
    23:36 Eleven Labs Disney accelerator
    24:55 Voice actor licensing coming
    25:49 *Opus clip version three *
    26:58 *Sunno version three alpha *
    28:41 *Sunno two minute songs *
    29:35 *Putin English dub uncanny *
    30:44 *Air Canada chatbot loses *
    31:40 Massive AI announcements coming

    Made with HARPA AI

  12. Hey Matt, love your content. Can you suggest an AI program that is cheap (or free) to help write teen fiction or screen plays? I have great ideas but no experience. Thanks Matt.

  13. Hey Matt, I'm a long-time listener and a first-time caller. The Chicago Cubs and the San Diego Padres are playing a preseason game (the Padres are up 3-0 at the bottom of the 2nd), and I wanted to check for any updates on AI. Keep up the great work!

  14. Hey can some one explain why Stable diffusion is often listed as the best image generator when midjourney clearly is worlds apart? Am i not finding the right thing on huggingface or something?


Please enter your comment!
Please enter your name here