🍓 AI makes its way into our games and our brains

Hey AI team! We’re back with a new edition of Five Minute AI this Tuesday ⏰

On today’s menu:

OpenAI is preparing “Project Strawberry” to revolutionize AI reasoning
An AI reinvents the game DOOM in real time, opening up new possibilities for the video game industry
Researchers are using AI to analyze 1.6 million brain scans to better detect dementia
A new AI tool pushes the limits of early cancer detection at the nanoscale

Ready? Let’s go...

THE MAIN STORY

🍓 OpenAI is preparing “Project Strawberry”

Midjourney, prompt in alt of the image.

Artificial intelligence keeps advancing, and OpenAI looks ready to take another big step. The AI giant is currently developing a model called “Strawberry” that could very well revolutionize the reasoning abilities of AI systems.

Details:

Strawberry is expected to be integrated into ChatGPT as early as fall 2024
The model excels at solving complex problems and puzzles
It will be used to generate high-quality training data for a future model called “Orion”
This breakthrough could significantly improve OpenAI’s AI agents

Why I’m telling you about this

If Strawberry lives up to its promise, we could see a major leap forward in AI reasoning. This shift could mark the start of a new wave of innovation in artificial intelligence, with potential applications across many industries. It will be worth keeping a close eye on how these new capabilities roll out in ChatGPT and how they affect the way we use AI every day.

🎮 An AI reinvents DOOM in real time

A screenshot of the first Doom game (1993) with a big modern title saying "QUAND L'IA IMAGINE DOOM". The background is a futuristic city with destroyed buildings. There are demons with various abilities. The game has a dark and gritty atmosphere.

Ideogram 2.0, prompt in the alt of the image.

Artificial intelligence keeps pushing the boundaries of digital creation. Researchers from Google and Tel Aviv University have just developed GameNGen, an AI system that can simulate the cult game DOOM in real time, opening up fascinating new possibilities for the video game industry.

Details:

GameNGen generates gameplay at more than 20 frames per second
The AI was trained on 900 million frames of gameplay
The system runs on a single chip, with no traditional game engine
Human testers struggle to tell the generated clips apart from the original game

Why I’m telling you about this

This breakthrough marks an important milestone in AI-generated game content. GameNGen shows that it’s possible to create a complex, dynamic gaming experience without a traditional game engine. Down the road, this technology could make it possible to build fully personalized games, generated on the fly for each player. We may be witnessing the early days of a revolution in how video games are designed and experienced.

LEARN TO MASTER AI

🖼️ Generate ultra-realistic images of yourself with a custom LoRA

A LoRA is an AI model that can be trained to add specific elements, like a person or an object, to generated images in a realistic way.

Artificial intelligence is opening fascinating doors in image creation. Today, let’s explore Fal AI’s Flux LoRA training tool, which lets you build your own custom image generation model. Picture yourself in any scenario or style, all from just a few selfies!

Here’s how to do it, step by step:

Head to Fal AI’s Flux LoRA training page and create an account. Note that you’ll need about $10 in credits to get started.
Pick 6 to 12 high-quality photos of yourself with clean backgrounds. The quality of your starting images will have a big impact on the final result.
Set up the training parameters: set the number of steps to 1000 and choose a unique trigger word (for example, “YourName”). This word will be the key to generating images of you.
Start the training process. Expect to wait about 20 minutes while the AI learns to recognize you.
Once training is done, have fun! Generate images with prompts like “Portrait of [YourName] as a superhero” and let your imagination run wild.

To go further: import your creations into Runway’s Gen-3 image-to-video feature to turn your images into short animated videos. A super realistic way to bring your creations to life!

WHAT THE RESEARCH SAYS

🧠 AI for the early detection of dementia

Artificial intelligence is about to play a crucial role in the fight against dementia. A team of Scottish researchers is launching an ambitious project to analyze more than 1.6 million brain scans using AI. The goal? Develop powerful tools to predict and diagnose dementia early.

Details:

The NEURii project, led by the Universities of Edinburgh and Dundee, will use machine learning to examine CT and MRI scans collected over a decade.
The researchers plan to build digital tools that let radiologists assess dementia risk during routine exams.
The study will combine imaging data with medical records to identify early warning signs of the disease.

Why I’m telling you about this

With dementia cases expected to reach 153 million worldwide by 2050, this research could have a significant impact on early detection and the development of new treatments. It’s part of a broader trend of using AI in healthcare, paving the way for more personalized and preventive medicine.

🔬 AI pushes the limits of early cancer detection

A new breakthrough in medical AI has just been achieved with the development of AINU, a tool that can analyze cell images at an unprecedented scale. This technological feat opens up new possibilities for the early detection of cancer and viral infections.

Details:

AINU uses a convolutional neural network to analyze images of cell nuclei at nanoscale resolution.
The AI can detect structural changes in cells as small as 20 nanometers, or 5,000 times finer than a human hair.
AINU was able to spot a viral infection (herpes simplex type 1) just one hour after infection by picking up subtle changes in how the DNA was organized.
The tool can also accurately identify stem cells, which could speed up research in that field without resorting to animal testing.

Why I’m telling you about this

Although clinical applications may still be a few years away, this breakthrough is a giant step forward in the early detection of disease. It shows how AI keeps pushing the frontiers of medical research, paving the way for faster, more accurate diagnoses and new personalized therapeutic approaches.

THE LATEST TRENDING TOOLS

🗣️ Theysaid turns your simple questions into AI-powered conversational surveys, offering a fresh, engaging way to collect customer feedback.
📚 Unriddle helps researchers read and write scientific papers, using AI to speed up the research and analysis process.
🎨 SuperCraft offers an innovative collaborative workspace that lets teams design physical products more intuitively and efficiently.
⏰ Clockwise optimizes time management with smart calendars, helping professionals better organize their days and boost their productivity.
😃 Discopixel Face Animator instantly generates facial expressions on any photo, opening up new creative possibilities for designers and photographers. VERY, VERY funny, I’ve tested it 😂.
👗 Kolors Virtual Try-On offers a personalized virtual try-on experience, letting shoppers see clothes on themselves before buying.
🎤 SpeechGeneratorAI instantly generates custom speeches for any occasion, making it easier to prepare oral presentations.

IN THE NEWS

Apple has announced its keynote for September 9, where the iPhone 16, packed with new generative AI features, is expected to debut.
Google Meet is adding an AI note-taking feature that automatically generates meeting summaries.
Cruise robotaxis will be integrated into the Uber app in 2025, following a multiyear partnership between General Motors’ self-driving subsidiary and the ride-hailing platform.
The Zed text editor is announcing new AI features powered by Anthropic’s Claude, leveraging its Rust-based design to deliver speed and security.
Klarna has cut the number of employees needed to handle customer requests thanks to AI, reducing resolution time from 11 minutes to 2.
Lightricks’ LTX Studio, an AI-powered visual storyboarding platform, is now available to everyone, offering real-time generation and editing tools for creative professionals.
Nvidia and Apple are reportedly considering joining Microsoft in a new OpenAI funding round that could value the AI startup at more than $100 billion.