Recap of What Has Been Happening in AI World

Artificial Intelligence has been taking the world by storm in the recent days from developments across various areas. Seemingly, new AI tools keep flooding the web day after day. Some more advanced like Sora and some just a replica of others.

Just the other day Google paused Gemini’s AI image generation of people after major backlash on the internet. Google later responded stating that they were aware of the inaccuracies in some historic image generation depictions.

Google’s rough road

Google seem to be having a rough month. Not too long after Gemini 1.5 breakthrough they were overshadowed by OpenAI’s Sora. And now, their release of Gemma, a new open-source family of AI models, was dwarfed by the backlash.

While they are still holding on the last straw, the tech giant announced a yet new AI tool branded ‘Help Me Write’ too. The AI is powered by Gemini and allows users to generate AI writing or revisions with a single click.

Google DeepMind also unveiled Genie, an AI system that can generate interactive video games from just a text or image prompt.

Despite not being trained on action or text annotations, the foundation model can determine who the main character is and enable a user to control that character in the generated world.

Stability AI

Stability AI unveiled Stable Diffusion 3, a new iteration of the company’s image generator. The model is a text-to-image model, utilizing a diffusion transformer architecture for greatly improved performance in multi-subject prompts, image quality, and spelling abilities. It is backed by a new transformer architecture similar to OpenAI’s Sora.

An Image generated by Stable Fusion

Magic doing some magic

Magic, a startup building AI coding “coworkers”, claims to have made a massive AI breakthrough. The startup claims its model now can ingest up to 3.5M words (nearly 5x Gemini 1.5) and made an ‘active reasoning’ breakthrough reminiscent of OpenAI’s Q*

On February 15, 2024, Magic announced a $117M raise led by former Github CEO Nat Friedman to build a frontier-scale AI code models, seeking to structure an ‘AI software engineer’ on the path towards Artificial General Intelligence (AGI).

Mistral coming after ChatGPT

French startup Mistral just launched ‘Mistral Large’, a new open-source model that beats ALL LLMs other than GPT-4 on key benchmarks. The startup also launched a beta version of its open-source ChatGPT competitor ‘Le Chat’.

Mistral AI competitors

Reddit partners with Google

When Reddit reported that they were selling its data for AI training, its buyer remained a mystery to many. But not for too long. Google was revealed as the partnering company for this project. The $60M deal will grant the search giant access to its user content for AI model training.

Social media platforms are goldmines of data for training AI models. While Meta and X have their own AI ambitions, others (like Reddit) could easily sign huge deals like this to reap immediate value from the data.

The deal also comes prior to Reddit’s initial public offering, so there’s a reason for them to boost value quickly

Reddit and Google partnership.

Moonshot

Chinese startup Moonshot AI raised over $1B at a $2.5B valuation from backers including Alibaba. The startup focuses on the development of large language models  (LLM) handling long-form text and has a ChatGPT alternative called ‘Kimi Chat’.

China’s AI has been accelerating behind the curtains and very little is known on what’s in the kitchen.

GPT updates

OpenAI released new updates to GPTs for both creators and users. Expanded profile options are coming to the platform alongside new ratings and feedback forms to provide to builders.

GPTs ‘About’ section can now include features such as builder social profiles, ratings, categories, conversation starters and other GPTs by the builder.

Nvidia

Nvidia blew past expectations with its Q4 earnings report, with a 265% year-over-year increase and massive data center revenue. The result? Up over a whopping 16%, pushing market cap passed Google and claiming position as the third largest stock in the world.

Nvidia seem to be dominating AI, blockchain and gaming spaces.

While still digesting all of that, the tech company’s CEO, Jensen Huang said he believes humans will no longer need to know how to code in the future. In a video that went viral online like wildfire, Huang argued that kids should stop learning how to code.

Sora shaking film industry

OpenAI’s Sora still remains to be an out-of-this world text-to-video model. Sora can generate videos up to a minute long while maintaining visual quality and adherence to the user’s prompt.

Sora is able to generate complex scenes with multiple characters, specific types of motion, and accurate details of the subject and background. The model understands not only what the user has asked for in the prompt, but also how those things exist in the physical world.

In an interview with The Hollywood Reporter Tyler Perry stated that he has put a halt to his planned $800 million expansion of his Atlanta studio, a decision influenced by the advancements in artificial intelligence, particularly OpenAI’s Sora.

Specifically, Perry worries that AI systems that can generate scripts, scores and even finished video content with minimal human input will put writers, composers, directors and other creative professionals out of work.

Over the past four years, Perry had been planning an $800 million expansion of his studio in Atlanta, which would have added 12 soundstages to the 330-acre property. Now, however, those ambitions are on hold — thanks to the rapid developments he’s seeing in the realm of artificial intelligence.

[videopress rbhy5oWx]